diff --git a/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/render.log b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/render.log new file mode 100644 index 0000000000000000000000000000000000000000..cdc6cb4424539bb3c410ffeed49f85ecedc0f062 --- /dev/null +++ b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/render.log @@ -0,0 +1,137 @@ + 7 none fig 12.29x5.70 in Experimental overview + 8 none fig 12.30x5.70 in Pretraining data: 33.14 million training cells + 9 none fig 12.45x5.67 in Training corpus: complete counts and minor categories +10 none fig 12.45x5.67 in Full source corpus: training, validation and test +11 none fig 12.45x5.67 in From 41.40M source cells to 33.14M training cells +12 none fig 12.45x5.67 in Recipe K and training exposure are different counts +13 below fig 12.38x5.33 in Corpus scale and generation workflow +14 none fig 12.45x4.07 in From Gaussian noise to generated cell states +17 none fig 12.45x5.67 in Where context width and trunk width change +18 none fig 12.45x5.67 in Predicted and oracle hints: one trained denoiser +19 none fig 12.45x5.67 in Architecture: full and late validation progression +20 below fig 9.57x4.63 in Width & context ablation: benchmark +21 none fig 12.45x5.67 in Architecture arms: dimensions and endpoints +22 below fig 9.57x4.63 in Gene context: 2k / 3k / 6k / 15k +23 none fig 12.45x5.67 in Gene panels: progression relative to each endpoint +24 none fig 12.45x5.67 in Gene-panel arms: exact sizes and training provenance +25 none fig 12.45x5.67 in All ablation arms: parameters and training time +26 none fig 12.15x5.70 in Real-cell repeat: a sampling reference +27 none fig 12.45x5.48 in Native-pseudobulk model comparisons +30 none fig 11.76x5.70 in Bulk and pooled cells: different expression values +31 none fig 12.45x5.19 in Paired calibration reduces same-donor expression error +32 below fig 10.47x4.63 in Consequence: cells off-manifold +33 none fig 12.33x5.70 in Adaptation recovers the observed cell mixture +34 none fig 12.45x5.35 in Adaptation: mixture error and expression agreement +35 none fig 12.15x5.70 in Healthy paired data: full FT and LoRA +36 none fig 12.45x5.40 in Full FT and LoRA ranks: quality and parameters +37 none fig 11.45x5.70 in What does “all pairs” actually mean? +38 none fig 12.19x5.70 in Healthy pair budget: correlation in all cell types +39 none fig 12.19x5.70 in Healthy pair budget: error in all cell types +40 right fig 9.20x5.17 in Five pairs, no extra validation: LoRA vs scSemi +41 none fig 12.17x5.70 in Composition: error and donor-level wins +44 right fig 9.20x5.18 in Disease signals with five training donors +45 right fig 9.20x4.80 in 5-fold CV: disease signal +46 below fig 11.93x4.63 in Disease genes at every list length (CAT) +47 below fig 10.54x4.63 in Cell-type deconvolution, same pairs: disease signal +48 none fig 12.45x5.70 in Disease gene rankings: complete CV comparison +49 none fig 11.99x5.70 in Disease integrity depends on cohort and endpoint +50 none fig 12.19x5.70 in Five-pair Sepsis: metrics across all support draws +53 none fig 12.33x5.70 in Figure 5 targets begin with the source paper +54 none fig 12.18x5.70 in Ye sepsis: published markers in paired scRNA +55 none fig 12.33x5.70 in Transcriptional programs and rare-population identity +61 none fig 12.45x2.62 in Corpus reference: assay and source composition +62 below fig 12.45x4.61 in Backbone: SDD → PBISC-Diffusion +63 below fig 12.45x4.61 in Bulk conditioning +64 below fig 11.68x5.33 in Denoiser: joint value + support diffusion +65 below fig 12.45x4.71 in Fine-tuning with paired bulk + sc +66 below fig 12.35x4.63 in Evaluation design +67 below fig 11.19x4.63 in How we score generated cells +68 none fig 12.45x5.31 in Does the model follow its bulk conditioning? +69 below fig 10.30x4.63 in SLE benchmark: PBISC vs scSemiProfiler +70 below fig 10.79x4.63 in SLE benchmark: per cell type +71 none fig 12.45x5.38 in Disjoint-cell controls test the bulk-broadcast baseline +72 none fig 12.45x5.60 in SLE: what is added beyond the bulk broadcast? +73 below fig 12.45x5.24 in Same shift in three paired cohorts +74 none fig 12.45x5.64 in Input corrections on one real-reference UMAP +75 none fig 12.45x4.86 in Where generated-cell density differs from real cells +76 none fig 12.45x4.89 in Input calibration changes cell-type composition +77 none fig 12.45x5.63 in Same noise, different input: matched-cell shifts +78 none fig 12.45x4.77 in Matched noise separates cell state from cell identity +79 none fig 12.45x5.63 in Input changes can switch the generated cell label +80 none fig 12.45x5.58 in Nearest real cells: distance and label agreement +81 none fig 12.34x5.70 in Composition agreement does not ensure state agreement +82 none fig 12.45x5.47 in Gene-length effects in the bulk input +83 none fig 12.45x5.47 in Gene-length effects within label-stable cells +84 none fig 12.45x5.41 in Raw and TPM shifts persist within cell labels +85 none fig 12.40x5.70 in Reference calibration leaves specific gene-state shifts +86 none fig 12.45x5.45 in Donor trait axes and donor identification differ +87 none fig 12.45x4.61 in Reference choice changes different quality endpoints +89 none fig 12.45x4.72 in Technical repeats and follow-up are different controls +90 none fig 12.45x5.32 in Annotation agreement and difficult categories +91 none fig 12.45x5.40 in Depth matching changes precision and coverage +92 none fig 12.45x4.86 in Detection and library size are separate diagnostics +93 right fig 9.20x5.18 in Healthy cohort: LoRA cells cover the real ones +94 right fig 9.20x4.79 in Healthy cohort: realism beyond precision/coverage +95 right fig 9.20x5.14 in Healthy cohort: full vs LoRA vs condition encoder +96 below fig 9.37x4.63 in Representative-dependent composition +97 below fig 11.39x4.63 in Full vs LoRA vs cond.-encoder-only +98 none fig 9.32x5.70 in Pair-budget summary: composition, coverage and steps +99 right fig 9.20x5.09 in Healthy cohort: LoRA rank and K +100 right fig 9.20x5.09 in Healthy paired-data learning curves +101 right fig 9.20x4.66 in Time: days for scSemiProfiler, minutes for LoRA +102 below fig 10.77x4.63 in 5 donors: time to apply +103 below fig 9.99x4.63 in Pairs needed: composition, coverage, training steps +104 below fig 10.08x4.63 in LoRA rank: r = 4, 16, 64 +105 right fig 8.31x5.70 in K diversification: no gain +106 below fig 12.45x3.82 in Five paired cohorts +107 none fig 12.45x4.70 in Sepsis: exact-five reconstruction across five draws +108 none fig 12.45x5.18 in Sepsis: realism beyond precision and coverage +109 none fig 12.45x4.70 in Imperial: first completed exact-five comparison +110 below fig 10.01x4.63 in 5 donors: LoRA vs scSemiProfiler — cells +111 below fig 10.04x4.63 in 5 donors: realism beyond precision/coverage +112 below fig 10.02x4.63 in 5 donors: precision/coverage at k = 3, 5, 10 +113 below fig 10.09x4.63 in 5-fold: realism beyond precision/coverage +114 below fig 11.74x4.63 in Same training pairs: LoRA vs scSemiProfiler +115 below fig 11.85x4.63 in scSemiProfiler and the representative-copy baseline +116 none fig 10.46x5.70 in Sepsis: patient-aware disease-gene agreement +117 none fig 10.46x5.70 in Imperial: patient-aware disease-gene agreement +118 below fig 10.60x4.63 in Sepsis: fine-tuning variants +119 right fig 8.44x5.70 in Pairs needed: donor and disease signal +120 below fig 10.53x4.63 in Cell-type deconvolution, same pairs: composition +121 none fig 12.45x4.95 in Deconvolution: expression coverage limits the comparison +122 none fig 12.45x5.08 in K=4,000 recipe: paired reconstruction quality +123 none fig 12.45x5.08 in K=4,000 recipe: disease signal does not improve uniformly +124 none fig 12.45x5.22 in K=4,000 recipe: calibration gains are selective +125 none fig 12.45x4.66 in Influenza: response order versus response genes +126 none fig 10.30x5.70 in COVID: calibration on a shared real-cell map +127 none fig 10.82x5.70 in COVID monocytes: direction after depth matching +128 none fig 11.44x5.70 in COVID monocytes: direction, magnitude and null +129 none fig 11.44x5.70 in COVID CD14 monocytes: direction, magnitude and null +130 none fig 12.45x5.02 in COVID: biological endpoints give a mixed verdict +131 none fig 11.83x5.70 in Rare-state frequencies in observed cells +132 none fig 11.35x5.70 in Rare-state misses: rules 1–8 +133 none fig 11.72x5.70 in Rare-state misses: rules 9–15 +134 none fig 12.06x5.70 in Rare-state fidelity: zero-aware reanalysis +135 none fig 10.76x5.70 in Sepsis rare states across paired-data budgets +136 none fig 12.45x5.62 in Sepsis rare-state direction across support budgets +137 none fig 11.35x5.70 in Rare-state screen in DRESS, Kawasaki and SLE +138 right fig 7.24x5.70 in Consequence: disease direction flips +139 none fig 12.45x5.18 in C1Q: single-gene detection versus co-detection +140 none fig 12.45x4.54 in Behçet: K=4,000 changes direction more than identity +141 none fig 12.45x5.59 in Behcet: one-gene and multi-gene C1Q rules +142 below fig 9.46x4.63 in Pairs from other cohorts do not transfer +143 below fig 9.44x4.63 in Longer training on other cohorts: gap stays +144 below fig 10.61x4.63 in No pretraining: same cohort, from scratch +145 below fig 10.53x4.63 in No pretraining: other cohorts, from scratch +146 below fig 10.20x4.63 in Rare states: from scratch vs pretrained +147 below fig 10.12x4.63 in Appendix: scSemiProfiler without output threshold +148 none fig 12.45x5.28 in Bulk-only exploration: calibration populations differ +150 right fig 9.20x5.19 in Rare-state fractions: original analysis +151 right fig 9.20x5.01 in Rare-state ranking: original analysis +152 below fig 9.23x4.98 in Rare states: original five-fold analysis +153 right fig 8.01x5.70 in Bulk-only cohort plan: original snapshot +154 none fig 12.45x4.73 in Historical bulk-only search: no confirmed new finding +156 below fig 12.45x4.62 in Next: other tissues from a few dozen pairs +{"slides": 156, "overview": 7, "training_corpus_main": 8, "training_detail": 9, "full_source_support": 10, "source_bindings": 155} +convert /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_expanded_master_20261015_figure1_focused.pptx as a Impress document -> /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_expanded_master_20261015_figure1_focused.pdf using filter : impress_pdf_Export:{"UseTaggedPDF":{"type":"boolean","value":"false"},"PDFUACompliance":{"type":"boolean","value":"false"}} +{"pass": true, "counts": {"spec": 156, "pptx": 156, "pdf": 156}, "failures": []} diff --git a/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/render.py b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/render.py new file mode 100644 index 0000000000000000000000000000000000000000..ee5e5dfe978b4d47036c91a23abd3dc247b6e3c3 --- /dev/null +++ b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/render.py @@ -0,0 +1,20 @@ +"""CPU-only source build, PDF export and verification.""" +from pathlib import Path +import os +import subprocess +import sys +if os.environ.get('CUDA_VISIBLE_DEVICES') != '': + raise SystemExit('CPU-only bounded unit required') +A=Path(__file__).resolve().parent +D=A.parents[1] +commands=[ + [sys.executable,str(D/'tools/build_figure1_focused_master_20261015.py')], + ['libreoffice','-env:UserInstallation=file:///tmp/pbisc-figure1-focused-lo-20261015','--headless','--convert-to', + 'pdf:impress_pdf_Export:{"UseTaggedPDF":{"type":"boolean","value":"false"},"PDFUACompliance":{"type":"boolean","value":"false"}}', + '--outdir',str(D),str(D/'PBISC-Diffusion_expanded_master_20261015_figure1_focused.pptx')], + [sys.executable,str(D/'tools/verify_figure1_focused_master_20261015.py')], +] +with (A/'render.log').open('a') as log: + for command in commands: + subprocess.run(command,cwd=D,stdout=log,stderr=subprocess.STDOUT,check=True,timeout=600) +print('Build, PDF and automated verification finished',flush=True) diff --git a/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/spec.json b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/spec.json new file mode 100644 index 0000000000000000000000000000000000000000..b6a9174fc350be0a13a75b14c50dc5be8aace2c3 --- /dev/null +++ b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure1_focus_20261015/spec.json @@ -0,0 +1,9641 @@ +{ + "deck": { + "output_pptx": "PBISC-Diffusion_expanded_master_20261015_figure1_focused.pptx", + "output_pdf": "PBISC-Diffusion_expanded_master_20261015_figure1_focused.pdf" + }, + "slides": [ + { + "kind": "title", + "title": "PBISC-Diffusion: manuscript evidence", + "subtitle": "156 slides | Diffusion-explicit overview | Readable training-corpus panel", + "original_slide": 1, + "original_title": "PBISC-Diffusion: bulk-to-single-cell generation", + "original_footnote": "", + "source_page_131": 1, + "deck_group": "Introduction", + "section": "Introduction", + "previous_section_label": null, + "section_label": "Introduction", + "parent_page": 1, + "parent_enriched_page": 1, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 1, + "parent_154_page": 1, + "parent_155_page": 1 + }, + { + "kind": "agenda", + "title": "Figures 1-6: the expanded evidence map", + "items": [ + "Figure 1 Experimental overview and corpus", + "Figure 2 SLE pseudobulk: architecture and gene context", + "Figure 3 Real-bulk mismatch and healthy paired adaptation", + "Figure 4 Disease integrity on measured paired cohorts", + "Figure 5 Paired rare-population qualitative evidence", + "Figure 6 Published-finding reproduction from bulk-only data", + "Supplementary evidence and controls", + "Historical scoring / plans and discussion" + ], + "original_slide": 2, + "original_title": "Outline", + "original_footnote": "", + "source_page_131": 2, + "deck_group": "Introduction", + "section": "Introduction", + "previous_section_label": null, + "section_label": "Introduction", + "parent_page": 2, + "parent_enriched_page": 2, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 2, + "parent_154_page": 2, + "parent_155_page": 2 + }, + { + "kind": "evidence_note", + "title": "How this expanded deck is organized", + "section_label": "Introduction", + "rows": [ + [ + "Figure 1: noise to cells", + "A recorded four-stage PCA view follows the experiment overview and corpus statistics." + ], + [ + "Figure 2: model design and ablations", + "Context/trunk and hint diagrams, full validation curves, SLE benchmarks, exact parameters and observed training times." + ], + [ + "Figure 3: all cell types", + "T, B, Monocyte, NK, DC and Other are compared using both donor correlation and absolute fraction error." + ] + ], + "after_original": 2, + "addition_key": "guide:scope", + "footnote": "Method palette: PBISC blue, scSemiProfiler mint, external comparators yellow. Biological category and quantitative colour scales retain their meaning.", + "footnote_full": "Method palette: PBISC blue, scSemiProfiler mint, external comparators yellow. Biological category and quantitative colour scales retain their meaning.", + "source_page_131": 3, + "deck_group": "Introduction", + "section": "Introduction", + "previous_section_label": "Evidence scope", + "parent_page": 3, + "parent_enriched_page": 3, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 3, + "parent_154_page": 3, + "parent_155_page": 3 + }, + { + "kind": "plan_card", + "section": "Introduction", + "section_label": "Current scope", + "title": "Manuscript map: the six figures", + "footnote": "Six author-defined scopes. Training validation and SLE biological benchmarks are labelled separately within Figure 2.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "Six author-defined scopes. Training validation and SLE biological benchmarks are labelled separately within Figure 2.", + "original_slide": 3, + "original_title": "Manuscript plan: six main figures", + "original_footnote": "Nature style: short abstract and introduction, descriptive results, short discussion, detailed methods; up to 17 supplementary figures.", + "source_page_131": 4, + "deck_group": "Introduction", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_overview.png" + ], + "rows": [ + [ + "Figure 1", + "Experimental overview, corpus and recorded denoising", + "Overview / illustration" + ], + [ + "Figure 2", + "Model components, validation, SLE ablations and cost", + "Pseudobulk development" + ], + [ + "Figure 3", + "Expression mismatch > full FT > rank > pair budget", + "Healthy measured pairs" + ], + [ + "Figure 4", + "Disease logFC, DEG and CAT benchmarks", + "Disease measured pairs" + ], + [ + "Figure 5", + "Rare-population identity in paired samples", + "Qualitative assembly" + ], + [ + "Figure 6", + "Published-finding reproduction from bulk-only samples", + "Planned" + ] + ], + "parent_page": 4, + "parent_enriched_page": 4, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 4, + "parent_154_page": 4, + "parent_155_page": 4 + }, + { + "kind": "divider", + "title": "Figure 1. Overview and corpus", + "original_slide": 4, + "original_title": "Figure 1. Design and performance of PBISC", + "original_footnote": "", + "source_page_131": 6, + "deck_group": "Figure 1", + "section": "Figure 1", + "previous_section_label": null, + "section_label": "Figure 1", + "parent_page": 5, + "parent_enriched_page": 5, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 5, + "parent_154_page": 5, + "parent_155_page": 5 + }, + { + "kind": "plan_card", + "section": "Figure 1", + "section_label": "Figure 1 plan", + "title": "Figure 1: overview, corpus and denoising", + "footnote": "Main proportions use 33,141,380 unique training cells. All grouped categories and exact counts remain in support.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "Main proportions use 33,141,380 unique training cells. All grouped categories and exact counts remain in support.", + "original_slide": 5, + "original_title": "Figure 1 plan: Design and performance of PBISC", + "original_footnote": "Diffusion on each gene's value and detection, conditioned on bulk, writes realistic blood cells. 5 of 6 panels ready.", + "source_page_131": 7, + "deck_group": "Figure 1", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_fig1.png" + ], + "rows": [ + [ + "1a", + "Pretrain > adapt > iterative denoising > sparse expression", + "p. 7" + ], + [ + "1b", + "Training corpus: major proportions and clear denominators", + "p. 8" + ], + [ + "Support", + "All training categories, full source corpus and split", + "pp. 9, 10, 11" + ], + [ + "1c", + "K recipes and repeated training exposure", + "p. 12" + ], + [ + "1d", + "Recorded noise-to-cell PCA stages", + "p. 14" + ] + ], + "parent_page": 6, + "parent_enriched_page": 6, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 6, + "parent_154_page": 6, + "parent_155_page": 6 + }, + { + "kind": "content", + "section": "Figure 1", + "section_label": "Fig. 1a", + "title": "Experimental overview", + "layout": "full", + "figures": [ + "figures/figure1_focus_20261015/overview/experimental_overview.png" + ], + "keywords": [], + "footnote": "Schematic gene profiles. Each independent noise draw yields one cell; N_gen is separate from recipe K. Held-out scRNA is evaluation only.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/figure1_overview_20261015.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_assignment_20261013/AUTHOR_SCOPE.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/diffusion.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/figure1_overview_20261014.py; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/build_figure1_architecture_nature.py", + "footnote_full": "Diffusion-explicit schematic with editable organic vector circles, muted fills, thin dark arrows and open white space, inspired by the supplied reference panel d. No scientific objects or numeric results were copied from that paper. 33,141,380 is the verified unique training-cell universe. A same-sample K-cell recipe supplies pseudobulk conditioning to each member target cell. Pretrained weights initialize LoRA adaptation on measured bulk\u2013scRNA pairs. New measured bulk and independent Gaussian-noise draws generate cells, without supplied cell-type quotas. Generic network nodes are replaced by explicit PBISC denoiser labels. The inference path shows an initial noisy gene-space state, iterative denoising, and a value/detection readout producing sparse expression with exact zeros. Plotted profile values are fixed illustrative glyphs, not recorded states, data or quantitative predictions. Intermediate noisy state is not nonnegative expression; the displayed final zeros are schematic measured-gene readouts, not an encoding of unmeasured genes. Cell colours are schematic, not cell-type quotas. The main model uses no cell-type/disease input labels; Figure2 oracle/prediction-hint ablations are separate. Generated annotations are post hoc. Conditioning recipe K and generated-cell number N_gen are distinct. Held-out scRNA is evaluation truth, never shown as an inference input. Rare-population endpoints are evaluation objectives, not a claim of completed source-cluster recovery; Figure6 bulk-only published-finding reproduction remains TODO.", + "numbers": [], + "original_slide": 7, + "original_title": "Overview draft (Nature style)", + "original_footnote": "Draft for Fig. 1b; schematic, no data. Plan: manuscript editorial/fig1_nature_redesign_plan.", + "figures_before_palette": [ + "figures/fig1b_overview_nature.png" + ], + "source_page_131": 9, + "deck_group": "Figure 1", + "previous_section_label": "Fig. 1b", + "parent_page": 7, + "parent_enriched_page": 7, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig1b_overview_nature.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 7, + "parent_154_page": 7, + "superseded_overview_figure": [ + "figures/figure1_revision_20261014/overview/experimental_overview.png" + ], + "revision_key": "figure1a:explicit_diffusion", + "parent_155_page": 7 + }, + { + "key": "main_figure1_training_corpus_focused", + "kind": "content", + "title": "Pretraining data: 33.14 million training cells", + "subtitle": "Main proportions; full category counts are retained in Figure 1 support", + "figure": "figures/figure1_focus_20261015/corpus/training_corpus_main.png", + "figures": [ + "figures/figure1_focus_20261015/corpus/training_corpus_main.png" + ], + "section": "Figure 1", + "section_label": "Fig. 1b", + "keywords": [], + "footnote": "Disease / at-risk includes ACPA+ risk. Proportions count each training-source cell once.", + "note": "Exact 33,141,380-cell training universe: 22 datasets, 2,572 corrected biological donor identities, 5,232 samples. Counts unchanged from the verified 2026-10-14 train-corpus bundle. Healthy 29,304,952 (88.4241%); disease/at-risk 3,836,428. Disease-detail percentages use only the disease/at-risk subset as denominator. ACPA+ at risk of RA is kept separate from RA. Other conditions/history comprises Cytopenia, DRESS, Influenza, JDM, MDS, MDS/MPN, Post-COVID-19. Other cell types comprises Progenitor cell, Erythrocyte, ILC, Platelet, Other. Unlabelled remains a separate category including all 99,769 training overlay cells and other unlabelled cells. Original unmerged counts and full category labels are retained on the Figure 1 support panel. Healthy includes stimulated, CMV-seropositive healthy and post-vaccination samples. Post-COVID-19 is historical/post-infection, and cytopenia is not equated with confirmed malignancy. Unique source-cell composition is not recipe-weighted exposure. The original 27-donor internal corpus cross-split caveat remains; this chart makes no new donor-disjoint validation claim.", + "footnote_full": "Exact 33,141,380-cell training universe: 22 datasets, 2,572 corrected biological donor identities, 5,232 samples. Counts unchanged from the verified 2026-10-14 train-corpus bundle. Healthy 29,304,952 (88.4241%); disease/at-risk 3,836,428. Disease-detail percentages use only the disease/at-risk subset as denominator. ACPA+ at risk of RA is kept separate from RA. Other conditions/history comprises Cytopenia, DRESS, Influenza, JDM, MDS, MDS/MPN, Post-COVID-19. Other cell types comprises Progenitor cell, Erythrocyte, ILC, Platelet, Other. Unlabelled remains a separate category including all 99,769 training overlay cells and other unlabelled cells. Original unmerged counts and full category labels are retained on the Figure 1 support panel. Healthy includes stimulated, CMV-seropositive healthy and post-vaccination samples. Post-COVID-19 is historical/post-infection, and cytopenia is not equated with confirmed malignancy. Unique source-cell composition is not recipe-weighted exposure. The original 27-donor internal corpus cross-split caveat remains; this chart makes no new donor-disjoint validation claim.", + "palette": "biological", + "deck_group": "Figure 1", + "layout": "full", + "addition_key": "figure1b:focused_training_corpus", + "sources": "data/figure1_focus_20261015/corpus/result.json; data/figure1_focus_20261015/corpus/display_aggregations.csv; tools/figure1_corpus_main_20261015.py" + }, + { + "key": "main_figure1_training_corpus", + "kind": "content", + "title": "Training corpus: complete counts and minor categories", + "subtitle": "Actual training-split composition, with each source cell counted once", + "figure": "figures/figure1_revision_20261014/train_corpus/training_corpus_composition.png", + "figures": [ + "figures/figure1_revision_20261014/train_corpus/training_corpus_composition.png" + ], + "section": "Figure 1", + "section_label": "Figure 1 support", + "keywords": [], + "footnote": "Exact training split; disease / at-risk includes ACPA+ risk. Counts describe unique source cells, not recipe-weighted exposure.", + "note": "Exact training-source composition from the same 33,141,380-cell universe covered by the final f01_6k_1x_seed20260627 run. 22 datasets, 5,232 samples, 2,572 biological donor identities after the original corpus GEO corrections. Healthy 29,304,952 (88.4241%); disease/at-risk 3,836,428. Condition counts use saved cell-level groups filtered to train, never sample-majority labels. Cell types are recomputed from matching corpus and source annotation metadata, with row identity checked for every base shard and exact sample membership joined to the registered train split. All 99,769 train overlay cells remain included as Unlabelled. Source-cell replay across all splits matches the independently saved 41,395,184-cell L1 summary exactly. ACPA+ is risk, not confirmed RA; Post-COVID is historical/post-infection; Cytopenia is not equated with confirmed malignancy. Healthy includes stimulation, post-vaccination and CMV-seropositive donors. Cell/condition labels are descriptive, not final base-model inputs. Unique source-cell proportions do not describe recipe-weighted target exposure. This panel does not assert biological donor-disjoint internal validation; the original 27-donor cross-split caveat remains.", + "footnote_full": "Exact training-source composition from the same 33,141,380-cell universe covered by the final f01_6k_1x_seed20260627 run. 22 datasets, 5,232 samples, 2,572 biological donor identities after the original corpus GEO corrections. Healthy 29,304,952 (88.4241%); disease/at-risk 3,836,428. Condition counts use saved cell-level groups filtered to train, never sample-majority labels. Cell types are recomputed from matching corpus and source annotation metadata, with row identity checked for every base shard and exact sample membership joined to the registered train split. All 99,769 train overlay cells remain included as Unlabelled. Source-cell replay across all splits matches the independently saved 41,395,184-cell L1 summary exactly. ACPA+ is risk, not confirmed RA; Post-COVID is historical/post-infection; Cytopenia is not equated with confirmed malignancy. Healthy includes stimulation, post-vaccination and CMV-seropositive donors. Cell/condition labels are descriptive, not final base-model inputs. Unique source-cell proportions do not describe recipe-weighted target exposure. This panel does not assert biological donor-disjoint internal validation; the original 27-donor cross-split caveat remains.", + "palette": "biological", + "deck_group": "Figure 1", + "layout": "full", + "addition_key": "figure1b:training_corpus", + "sources": "data/figure1_revision_20261014/train_corpus/result.json; data/figure1_revision_20261014/train_corpus/source_hashes.json; tools/figure1_train_corpus_20261014.py", + "parent_155_page": 8 + }, + { + "key": "corpus_conditions", + "kind": "content", + "title": "Full source corpus: training, validation and test", + "subtitle": "Two-level condition view with the existing cell-type statistics", + "figure": "figures/strengthening_20261013/corpus/healthy_disease_overview.png", + "figures": [ + "figures/strengthening_20261013/corpus/healthy_disease_overview.png" + ], + "section": "Figure 1", + "section_label": "Figure 1 support", + "keywords": [], + "footnote": "Source-cell shares. Disease / at-risk includes ACPA+ at-risk, not confirmed RA. Healthy includes stimulation, vaccination and CMV contexts.", + "note": "Saved corpus metadata and final f01_6k_1x_seed20260627 E16 training contract; no raw expression arrays or new model runs. Biological condition/split colors retain their existing meanings, independent of the method palette. Left denominator41,395,184; Healthy36,639,276(88.51%); disease/at-risk4,755,908(11.49%). Detail contains11 source conditions, including ACPA+ at-risk1,792,129; disease/history subtotal excluding at-risk2,963,779. Post-COVID remains a history/post-infection label and Cytopenia is retained verbatim, not called a confirmed malignancy. Counts derive from cell-level groups, not per-sample majority condition. Healthy includes9,762,726 in-vitro cells: Parse9,697,974 (90 cytokines or PBS) and IL-36/PBS healthy-donor cells64,752. CMV-seropositive healthy cells9,215,690; post-vaccination source draws remain Healthy. These are donor/handling context, not inferred active disease. Right donut preserves original harmonized L1 source cell-type counts and colors: NKT/Granulocyte/Neutrophil/Mast\u2192Other, unknown/overlay labels\u2192Unlabelled, matching the source figure merge rules. Total source-cell denominator is checked again. The model receives no disease/cell-type labels in this final base run.", + "footnote_full": "Saved corpus metadata and final f01_6k_1x_seed20260627 E16 training contract; no raw expression arrays or new model runs. Biological condition/split colors retain their existing meanings, independent of the method palette. Left denominator41,395,184; Healthy36,639,276(88.51%); disease/at-risk4,755,908(11.49%). Detail contains11 source conditions, including ACPA+ at-risk1,792,129; disease/history subtotal excluding at-risk2,963,779. Post-COVID remains a history/post-infection label and Cytopenia is retained verbatim, not called a confirmed malignancy. Counts derive from cell-level groups, not per-sample majority condition. Healthy includes9,762,726 in-vitro cells: Parse9,697,974 (90 cytokines or PBS) and IL-36/PBS healthy-donor cells64,752. CMV-seropositive healthy cells9,215,690; post-vaccination source draws remain Healthy. These are donor/handling context, not inferred active disease. Right donut preserves original harmonized L1 source cell-type counts and colors: NKT/Granulocyte/Neutrophil/Mast\u2192Other, unknown/overlay labels\u2192Unlabelled, matching the source figure merge rules. Total source-cell denominator is checked again. The model receives no disease/cell-type labels in this final base run.", + "palette": "biological", + "source_refs": [ + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py" + ], + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json", + "host": "server_1", + "sha256": "cbccff88a6872bc67dd80c0dfde7d4741733274194f5fd10e002fe0a633a6fa3", + "bytes": 3298, + "role": "saved corpus condition, split and overlap totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv", + "host": "server_1", + "sha256": "cad720e9c8ab5759b9ec73e88a443c9733ebd04c22cbf5a28ccf3d56104c5d71", + "bytes": 599627, + "role": "cell-level condition counts; not sample-majority labels" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv", + "host": "server_1", + "sha256": "f4314d4e43d0ee3fa2d8c6fcabb616a7fb5cc3e06dd6dbe38a00e389190d0cea", + "bytes": 4711, + "role": "independent dataset-level cell and split totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv", + "host": "server_1", + "sha256": "7f729ace1d5d3ee2b8f58d99dbc6609db173d0414aa54057bf9cf3f6a69d5ccf", + "bytes": 1166063, + "role": "registered sample and split totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv", + "host": "server_1", + "sha256": "d3adc5349083175d0a9e0f196cdf03b2eb28001b345e6e5b8f8b7149c90586a6", + "bytes": 652, + "role": "independent assay-condition aggregation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv", + "host": "server_1", + "sha256": "5b29e8373ad2597f02a2a7535405106f1a2ed0c9a966d828e89d43c1c59e89d1", + "bytes": 14066, + "role": "existing harmonized source L1/L2 cell-type counts" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md", + "host": "server_1", + "sha256": "ae5ec49b2c2d2c16d065a95165142ba7809a6cb70ed399a884d094bb8f81154a", + "bytes": 3207, + "role": "condition definitions and internal donor overlap caveat" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv", + "host": "server_1", + "sha256": "a8150b6cff42aabd20d1f5778efa2543ff0680d91afef63bacba0732ade44b4b", + "bytes": 2130, + "role": "healthy stimulation/CMV and disease-label interpretation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py", + "host": "server_1", + "sha256": "42348c3fd16c4deaba40f3fda2ba46a288298d9779693b5cbada03294d15b702", + "bytes": 16832, + "role": "source condition harmonization implementation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py", + "host": "server_1", + "sha256": "629283828350064b0d71c841a321f2fcce61a7739b82ffc57b2fa219ebd2b721", + "bytes": 22102, + "role": "existing biological condition and split colors" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json", + "host": "server_1", + "sha256": "4f5f16f272db090c256322eac8e3dc8ceca2097589f673fe3492f0107a7f32f4", + "bytes": 10046, + "role": "registered source corpus, split unit, K and platform recipe quotas" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json", + "host": "server_1", + "sha256": "461b46107f75984000747877a11823ec15042b5c2104b4b979729dcee10ee30f", + "bytes": 6968, + "role": "final Model8 all-cell training coverage and emitted targets" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "host": "server_1", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "bytes": 1423, + "role": "final Model8 completion and corpus fingerprint" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "host": "server_1", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "bytes": 3319, + "role": "final Model8 gene panel and all-K target cap" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json", + "host": "server_1", + "sha256": "3904c5b4026f3a5358d6553948f0d4456c80494eab941ab2f7c2975f459d24c5", + "bytes": 4772, + "role": "registered recipe unique coverage and sampling semantics" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py", + "host": "server_1", + "sha256": "6f6db69002ae01857a0f1882b2e6e98552f3ee00aa1116e4fd4a308cbf761b71", + "bytes": 58198, + "role": "K distinct cells sampled within a sample; count-sum pseudobulk" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "host": "server_1", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "bytes": 51067, + "role": "all-K plus top-up training contract implementation" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv", + "host": "server_1", + "sha256": "ad94944067f1697e6a6f7d33b82a0fecbc63e271af25448d712cd5de69a7d700", + "bytes": 966455, + "role": "saved explicit top-up recipe sizes and split" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md", + "host": "server_1", + "sha256": "feff43bfd27e54b39f94f5170b1b75a76cd378985d869b612075f592ae429958", + "bytes": 2058, + "role": "method palette; biological category colors exempt" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py", + "host": "server_1", + "sha256": "da454c6d35eaad145091a7761cb23900d752c256bf894589bdd1071d02623b20", + "bytes": 27130, + "role": "reproducible metadata-only figure builder" + } + ], + "deck_group": "Figure 1", + "addition_key": "strengthening:corpus:0", + "layout": "full", + "sources": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py", + "parent_154_page": 8, + "parent_155_page": 9 + }, + { + "key": "corpus_splits", + "kind": "content", + "title": "From 41.40M source cells to 33.14M training cells", + "subtitle": "The difference is a split boundary, not K-based exclusion", + "figure": "figures/strengthening_20261013/corpus/source_split_flow.png", + "figures": [ + "figures/strengthening_20261013/corpus/source_split_flow.png" + ], + "section": "Figure 1", + "section_label": "Figure 1 support", + "keywords": [], + "footnote": "Source train/validation/test counts, not validation recipe exposure. No sample overlap; 27 biological donors span internal corpus splits.", + "note": "Saved corpus metadata and final f01_6k_1x_seed20260627 E16 training contract; no raw expression arrays or new model runs. Biological condition/split colors retain their existing meanings, independent of the method palette. Exact counts: training33,141,380, validation4,361,032, test3,892,772; sum41,395,184. Source samples5232/648/627. Split rule dataset_id+donor_id with sample_id fallback; nominal0.8/0.1/0.1 and realized cell proportions differ. Independent source metadata found27 biological donors with multi-split samples (26GSE274680,1GSE132802),38validation/test samples and659,112cells (331,377val,327,735test) sharing a biological donor with training. These are INTERNAL corpus overlaps, not evidence that external SLE/paired benchmark patients occur in training. Actual validation recipes consume selected cells, not the entire validation source split. Source-count identity and final-run corpus fingerprint are checked.", + "footnote_full": "Saved corpus metadata and final f01_6k_1x_seed20260627 E16 training contract; no raw expression arrays or new model runs. Biological condition/split colors retain their existing meanings, independent of the method palette. Exact counts: training33,141,380, validation4,361,032, test3,892,772; sum41,395,184. Source samples5232/648/627. Split rule dataset_id+donor_id with sample_id fallback; nominal0.8/0.1/0.1 and realized cell proportions differ. Independent source metadata found27 biological donors with multi-split samples (26GSE274680,1GSE132802),38validation/test samples and659,112cells (331,377val,327,735test) sharing a biological donor with training. These are INTERNAL corpus overlaps, not evidence that external SLE/paired benchmark patients occur in training. Actual validation recipes consume selected cells, not the entire validation source split. Source-count identity and final-run corpus fingerprint are checked.", + "palette": "biological", + "source_refs": [ + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py" + ], + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json", + "host": "server_1", + "sha256": "cbccff88a6872bc67dd80c0dfde7d4741733274194f5fd10e002fe0a633a6fa3", + "bytes": 3298, + "role": "saved corpus condition, split and overlap totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv", + "host": "server_1", + "sha256": "cad720e9c8ab5759b9ec73e88a443c9733ebd04c22cbf5a28ccf3d56104c5d71", + "bytes": 599627, + "role": "cell-level condition counts; not sample-majority labels" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv", + "host": "server_1", + "sha256": "f4314d4e43d0ee3fa2d8c6fcabb616a7fb5cc3e06dd6dbe38a00e389190d0cea", + "bytes": 4711, + "role": "independent dataset-level cell and split totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv", + "host": "server_1", + "sha256": "7f729ace1d5d3ee2b8f58d99dbc6609db173d0414aa54057bf9cf3f6a69d5ccf", + "bytes": 1166063, + "role": "registered sample and split totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv", + "host": "server_1", + "sha256": "d3adc5349083175d0a9e0f196cdf03b2eb28001b345e6e5b8f8b7149c90586a6", + "bytes": 652, + "role": "independent assay-condition aggregation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv", + "host": "server_1", + "sha256": "5b29e8373ad2597f02a2a7535405106f1a2ed0c9a966d828e89d43c1c59e89d1", + "bytes": 14066, + "role": "existing harmonized source L1/L2 cell-type counts" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md", + "host": "server_1", + "sha256": "ae5ec49b2c2d2c16d065a95165142ba7809a6cb70ed399a884d094bb8f81154a", + "bytes": 3207, + "role": "condition definitions and internal donor overlap caveat" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv", + "host": "server_1", + "sha256": "a8150b6cff42aabd20d1f5778efa2543ff0680d91afef63bacba0732ade44b4b", + "bytes": 2130, + "role": "healthy stimulation/CMV and disease-label interpretation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py", + "host": "server_1", + "sha256": "42348c3fd16c4deaba40f3fda2ba46a288298d9779693b5cbada03294d15b702", + "bytes": 16832, + "role": "source condition harmonization implementation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py", + "host": "server_1", + "sha256": "629283828350064b0d71c841a321f2fcce61a7739b82ffc57b2fa219ebd2b721", + "bytes": 22102, + "role": "existing biological condition and split colors" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json", + "host": "server_1", + "sha256": "4f5f16f272db090c256322eac8e3dc8ceca2097589f673fe3492f0107a7f32f4", + "bytes": 10046, + "role": "registered source corpus, split unit, K and platform recipe quotas" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json", + "host": "server_1", + "sha256": "461b46107f75984000747877a11823ec15042b5c2104b4b979729dcee10ee30f", + "bytes": 6968, + "role": "final Model8 all-cell training coverage and emitted targets" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "host": "server_1", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "bytes": 1423, + "role": "final Model8 completion and corpus fingerprint" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "host": "server_1", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "bytes": 3319, + "role": "final Model8 gene panel and all-K target cap" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json", + "host": "server_1", + "sha256": "3904c5b4026f3a5358d6553948f0d4456c80494eab941ab2f7c2975f459d24c5", + "bytes": 4772, + "role": "registered recipe unique coverage and sampling semantics" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py", + "host": "server_1", + "sha256": "6f6db69002ae01857a0f1882b2e6e98552f3ee00aa1116e4fd4a308cbf761b71", + "bytes": 58198, + "role": "K distinct cells sampled within a sample; count-sum pseudobulk" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "host": "server_1", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "bytes": 51067, + "role": "all-K plus top-up training contract implementation" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv", + "host": "server_1", + "sha256": "ad94944067f1697e6a6f7d33b82a0fecbc63e271af25448d712cd5de69a7d700", + "bytes": 966455, + "role": "saved explicit top-up recipe sizes and split" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md", + "host": "server_1", + "sha256": "feff43bfd27e54b39f94f5170b1b75a76cd378985d869b612075f592ae429958", + "bytes": 2058, + "role": "method palette; biological category colors exempt" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py", + "host": "server_1", + "sha256": "da454c6d35eaad145091a7761cb23900d752c256bf894589bdd1071d02623b20", + "bytes": 27130, + "role": "reproducible metadata-only figure builder" + } + ], + "deck_group": "Figure 1", + "addition_key": "strengthening:corpus:1", + "layout": "full", + "sources": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py", + "parent_154_page": 9, + "parent_155_page": 10 + }, + { + "key": "corpus_recipes", + "kind": "content", + "title": "Recipe K and training exposure are different counts", + "subtitle": "All training cells are covered after explicit top-up recipes", + "figure": "figures/strengthening_20261013/corpus/recipe_coverage_and_exposure.png", + "figures": [ + "figures/strengthening_20261013/corpus/recipe_coverage_and_exposure.png" + ], + "section": "Figure 1", + "section_label": "Figure 1 support", + "keywords": [], + "footnote": "Saved all-cell audit: 33,141,380 unique training cells covered. K defines recipe membership; 195,639,930 target occurrences include reuse.", + "note": "Saved corpus metadata and final f01_6k_1x_seed20260627 E16 training contract; no raw expression arrays or new model runs. Biological condition/split colors retain their existing meanings, independent of the method palette. Registered K250/500/1000/2000/4000 has recipe quotas745/1489/2234/2979/745 per part (weights1/2/3/4/1);8192recipes/part\u00d716parts=131072. All K members are targets because cells_per_block4000 is at least maxK. Registered recipes alone cover31,145,450 unique train cells (93.9775%); top-up3658recipes adds1,995,930 previously unseen train cells once, totaling33,141,380(100%). Top-upK1\u20134000 preserves small tails without padding; total134730recipes. Total195,639,930target occurrences across16parts include repeated source cells and do not imply195Munique cells. Platform quotas10x4691,Parse2625,ScaleBio698,Seq-Well178 per part derive from square-root source-cell counts, not natural corpus prevalence. Disease-specific exposure proportions are not available in these figures. K is distinct from N_gen, the number of cells sampled at inference.", + "footnote_full": "Saved corpus metadata and final f01_6k_1x_seed20260627 E16 training contract; no raw expression arrays or new model runs. Biological condition/split colors retain their existing meanings, independent of the method palette. Registered K250/500/1000/2000/4000 has recipe quotas745/1489/2234/2979/745 per part (weights1/2/3/4/1);8192recipes/part\u00d716parts=131072. All K members are targets because cells_per_block4000 is at least maxK. Registered recipes alone cover31,145,450 unique train cells (93.9775%); top-up3658recipes adds1,995,930 previously unseen train cells once, totaling33,141,380(100%). Top-upK1\u20134000 preserves small tails without padding; total134730recipes. Total195,639,930target occurrences across16parts include repeated source cells and do not imply195Munique cells. Platform quotas10x4691,Parse2625,ScaleBio698,Seq-Well178 per part derive from square-root source-cell counts, not natural corpus prevalence. Disease-specific exposure proportions are not available in these figures. K is distinct from N_gen, the number of cells sampled at inference.", + "palette": "biological", + "source_refs": [ + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py", + "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md", + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py" + ], + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json", + "host": "server_1", + "sha256": "cbccff88a6872bc67dd80c0dfde7d4741733274194f5fd10e002fe0a633a6fa3", + "bytes": 3298, + "role": "saved corpus condition, split and overlap totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv", + "host": "server_1", + "sha256": "cad720e9c8ab5759b9ec73e88a443c9733ebd04c22cbf5a28ccf3d56104c5d71", + "bytes": 599627, + "role": "cell-level condition counts; not sample-majority labels" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv", + "host": "server_1", + "sha256": "f4314d4e43d0ee3fa2d8c6fcabb616a7fb5cc3e06dd6dbe38a00e389190d0cea", + "bytes": 4711, + "role": "independent dataset-level cell and split totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv", + "host": "server_1", + "sha256": "7f729ace1d5d3ee2b8f58d99dbc6609db173d0414aa54057bf9cf3f6a69d5ccf", + "bytes": 1166063, + "role": "registered sample and split totals" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv", + "host": "server_1", + "sha256": "d3adc5349083175d0a9e0f196cdf03b2eb28001b345e6e5b8f8b7149c90586a6", + "bytes": 652, + "role": "independent assay-condition aggregation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv", + "host": "server_1", + "sha256": "5b29e8373ad2597f02a2a7535405106f1a2ed0c9a966d828e89d43c1c59e89d1", + "bytes": 14066, + "role": "existing harmonized source L1/L2 cell-type counts" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md", + "host": "server_1", + "sha256": "ae5ec49b2c2d2c16d065a95165142ba7809a6cb70ed399a884d094bb8f81154a", + "bytes": 3207, + "role": "condition definitions and internal donor overlap caveat" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv", + "host": "server_1", + "sha256": "a8150b6cff42aabd20d1f5778efa2543ff0680d91afef63bacba0732ade44b4b", + "bytes": 2130, + "role": "healthy stimulation/CMV and disease-label interpretation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py", + "host": "server_1", + "sha256": "42348c3fd16c4deaba40f3fda2ba46a288298d9779693b5cbada03294d15b702", + "bytes": 16832, + "role": "source condition harmonization implementation" + }, + { + "path": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py", + "host": "server_1", + "sha256": "629283828350064b0d71c841a321f2fcce61a7739b82ffc57b2fa219ebd2b721", + "bytes": 22102, + "role": "existing biological condition and split colors" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json", + "host": "server_1", + "sha256": "4f5f16f272db090c256322eac8e3dc8ceca2097589f673fe3492f0107a7f32f4", + "bytes": 10046, + "role": "registered source corpus, split unit, K and platform recipe quotas" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json", + "host": "server_1", + "sha256": "461b46107f75984000747877a11823ec15042b5c2104b4b979729dcee10ee30f", + "bytes": 6968, + "role": "final Model8 all-cell training coverage and emitted targets" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "host": "server_1", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "bytes": 1423, + "role": "final Model8 completion and corpus fingerprint" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "host": "server_1", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "bytes": 3319, + "role": "final Model8 gene panel and all-K target cap" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json", + "host": "server_1", + "sha256": "3904c5b4026f3a5358d6553948f0d4456c80494eab941ab2f7c2975f459d24c5", + "bytes": 4772, + "role": "registered recipe unique coverage and sampling semantics" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py", + "host": "server_1", + "sha256": "6f6db69002ae01857a0f1882b2e6e98552f3ee00aa1116e4fd4a308cbf761b71", + "bytes": 58198, + "role": "K distinct cells sampled within a sample; count-sum pseudobulk" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "host": "server_1", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "bytes": 51067, + "role": "all-K plus top-up training contract implementation" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv", + "host": "server_1", + "sha256": "ad94944067f1697e6a6f7d33b82a0fecbc63e271af25448d712cd5de69a7d700", + "bytes": 966455, + "role": "saved explicit top-up recipe sizes and split" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md", + "host": "server_1", + "sha256": "feff43bfd27e54b39f94f5170b1b75a76cd378985d869b612075f592ae429958", + "bytes": 2058, + "role": "method palette; biological category colors exempt" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py", + "host": "server_1", + "sha256": "da454c6d35eaad145091a7761cb23900d752c256bf894589bdd1071d02623b20", + "bytes": 27130, + "role": "reproducible metadata-only figure builder" + } + ], + "deck_group": "Figure 1", + "addition_key": "strengthening:corpus:2", + "layout": "full", + "sources": "/home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_summary.json; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_cellgroups.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_datasets.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_samples.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_assay_condition.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/corpus_celltypes.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/README.md; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/source_data/corpus/dataset_info.csv; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_corpus_tables.py; /home/sj_server_1/PBISC/paper-release-20260928/manuscript/scripts/corpus/build_figure1_corpus.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/data/e2_full_index/dataset_manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/ALLCELL_DATA_CONTRACT.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/coverage_v1/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_7_no_annotation_sh_hwang/src/model7_no_annotation/data_e2.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/full_recipe_allcells_6k_broad_20260909/topup_v1/topup_recipes.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_corpus_20261013.py", + "parent_154_page": 10, + "parent_155_page": 11 + }, + { + "kind": "content", + "section": "Figure 1", + "section_label": "Fig. 1c", + "title": "Corpus scale and generation workflow", + "layout": "full", + "figures": [ + "figures/sec01_slide_a.png" + ], + "keywords": [ + "Train pair: K-cell pseudobulk \u2192 each member cell" + ], + "footnote": "Fig. 1b redraw. Final model corpus: 33.1M train cells; K = 1\u20134,000 cells per pseudobulk.", + "footnote_full": "Original slide 13: Top: deck redraw of manuscript Fig. 1b (larger labels; cells differ by x_T and the Bernoulli support draw, Methods 'Support read-out'). Bottom: corpus of final run f01_6k_1x_seed20260627 (index fingerprint matches). Cells per dataset = base + overlay rows (Model 7 e2_full_source/e2_full_index manifests); 33.1M cells in the donor-level 80/10/10 train split. Registered K grid = 8,192 recipes/epoch. Plus 3,658 top-up recipes over 16 epochs (196-283 per epoch, K 1-4,000), so every train cell is a target at least once (ALLCELL_DATA_CONTRACT.json coverage_after_topup).", + "sources": "tools/sec01_overview.py; data/sec01_model_numbers.csv, data/sec01_corpus_datasets.csv, data/sec01_recipe_k.csv", + "refs": [ + "Ostheimer 2025, arXiv", + "Song 2021, ICLR" + ], + "refs_full": [ + "Ostheimer et al. Sparse Data Diffusion for scientific simulations in biology and physics. arXiv:2502.02448 (2025)", + "Song, Meng & Ermon. Denoising Diffusion Implicit Models. ICLR 2021" + ], + "original_slide": 13, + "original_title": "PBISC-Diffusion: bulk in, cells out", + "original_footnote": "Fig. 1b redraw. Final model corpus: 33.1M train cells; K = 1\u20134,000 cells per pseudobulk.", + "figures_before_palette": [ + "figures/sec01_slide_a.png" + ], + "source_page_131": 16, + "deck_group": "Figure 1", + "previous_section_label": "Fig. 1 details", + "parent_page": 9, + "parent_enriched_page": 9, + "figures_before_final_palette": [ + "figures/sec01_slide_a.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 9, + "parent_154_page": 11, + "parent_155_page": 12 + }, + { + "title": "From Gaussian noise to generated cell states", + "figures": [ + "figures/figure_enrichment_20261013/trajectory_four_pca_frames.png" + ], + "footnote": "Recorded ctx2048 E12 EMA; same-donor pseudobulk illustration. Colour reveal shows step progress, not identity confidence.", + "footnote_full": "Four recorded noisy-state PCA frames (steps 0, 17, 33 and 50) from an existing ctx2048 E12 EMA run. All 4,096 tracked cells retain their row identities; axes are fitted once on 4,362 observed cells from the conditioning donor. Colours use frozen final predicted PBMC5+Other labels, revealed according to step/50; early states were not classified. This is computational denoising, not biological time, and the same-donor reference is not independent validation.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/trajectory_provenance.json", + "kind": "content", + "deck_group": "Figure 1", + "section": "Figure 1", + "section_label": "Fig. 1d", + "layout": "full", + "keywords": [], + "addition_key": "enrichment:trajectory_four_pca_frames", + "parent_enriched_page": 10, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/trajectory_four_pca_frames.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 10, + "parent_154_page": 12, + "parent_155_page": 13 + }, + { + "kind": "divider", + "title": "Figure 2. SLE model ablations", + "original_slide": 15, + "original_title": "Figure 2. Fine-tuning PBISC with LoRA", + "original_footnote": "", + "source_page_131": 19, + "deck_group": "Figure 2", + "section": "Figure 2", + "previous_section_label": null, + "section_label": "Figure 2", + "parent_page": 10, + "parent_enriched_page": 11, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 11, + "parent_154_page": 13, + "parent_155_page": 14 + }, + { + "kind": "plan_card", + "section": "Figure 2", + "section_label": "Figure 2 plan", + "title": "Figure 2: architecture and gene context", + "footnote": "Validation curves use corpus recipes; biological benchmarks use SLE pseudobulk. No measured bulk-sc paired performance in Figure 2.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "Validation curves use corpus recipes; biological benchmarks use SLE pseudobulk. No measured bulk-sc paired performance in Figure 2.", + "original_slide": 16, + "original_title": "Figure 2 plan: Fine-tuning PBISC with LoRA", + "original_footnote": "On a healthy cohort, a few pairs and a LoRA of minutes give cells close to the real ones. No disease metrics. 9 of 9 panels ready.", + "source_page_131": 20, + "deck_group": "Figure 2", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_fig2.png" + ], + "rows": [ + [ + "2a-b", + "Context / trunk widths and predicted / oracle hint input", + "pp. 17, 18" + ], + [ + "2c-d", + "Architecture validation progression and SLE benchmark", + "pp. 19, 20" + ], + [ + "2e-f", + "Gene-context benchmark and validation progression", + "pp. 22, 23" + ], + [ + "2g", + "All arms: parameters and observed training time", + "p. 25" + ], + [ + "2h-i", + "Real-cell repeat explained; common-axis model comparison", + "pp. 26, 27" + ], + [ + "Arm details", + "Exact widths, genes, endpoints, run completion and timing", + "pp. 21, 24" + ] + ], + "parent_page": 11, + "parent_enriched_page": 12, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 12, + "parent_154_page": 14, + "parent_155_page": 15 + }, + { + "key": "arch_context_vs_trunk", + "kind": "content", + "type": "content", + "title": "Where context width and trunk width change", + "section": "Figure 2", + "section_label": "Fig. 2a", + "figures": [ + "figures/final_palette_20261013/arch_context_vs_trunk.png" + ], + "keywords": [], + "footnote": "Code-derived diagram. Width multipliers are not parameter multipliers; all architecture arms use 6,000 genes.", + "footnote_full": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Seven conditioned blocks: down widths [1024,1024,512,256], up widths [512,1024,1024], three additive skips. Context = mixer(time encoder + condition encoder); each block adds its own projected context. No latent-space diffusion. Visual organization follows the public website's condition strip / input / denoiser / output cards; fresh vector implementation, not copied recordings.", + "notes": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Seven conditioned blocks: down widths [1024,1024,512,256], up widths [512,1024,1024], three additive skips. Context = mixer(time encoder + condition encoder); each block adds its own projected context. No latent-space diffusion. Visual organization follows the public website's condition strip / input / denoiser / output cards; fresh vector implementation, not copied recordings.", + "after_current": 11, + "after_current_page": 11, + "palette": "native", + "role": "diagram", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "replace_current_page": null, + "replacement_previous_131_page": null, + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py", + "sha256": "c3d725b97d003b9b2fb45d5061d5f7daffe89fd37147f9d53753b3ed2da77ffa", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py", + "sha256": "5e4f6169410fa165e5b1c39391ab731c38c32b7d3202618da326b2f82131e6de", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py", + "sha256": "0d959a35175549a263e1b23c2709bfe2699dc293141d5699f01cf70cd7a0ba73", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py", + "sha256": "7a352d73728e78c9d262e29c35628d2437af0bc7743b88c077ba81d911dd2a52", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py", + "sha256": "09631a39fb388b56c4be88814559d82fac100660571a4b2a8e74e9972c400758", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md", + "sha256": "7fcae090a1e4417eb52c7a51556f0c82bacdc3bcf47193f41be2c30d350c357b", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md", + "sha256": "8c24781e40dd4284f06a07b9d7eac35dae2cfe013e1b46bd5a1805fd9afe83db", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py", + "sha256": "f023dc2fa788c8bac628fa922e5a34ccf05d355edec2a57807ca9bb24f3c241e", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py", + "sha256": "044cf1efdb17761914b1707861aa606095ca4a8ede109bb337782395cff01c47", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py", + "sha256": "271834f39e5afb30c7c9e377b0b0f9aba06a85f8fe566a76ff263649f9a09ef0", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py", + "sha256": "95ace096cdc58a9ca92a18af649c60fe6ac0a55e97f977e7ab9bcb04efa968cb", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json", + "sha256": "8377fa821211086a56b3450af3a464f55a5799c6df1b2486019dc96d46331587", + "role": "saved separate hint predictor metrics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json", + "sha256": "af11b163a08666849621e7e66d89b2234314231bbecf976afa6f4b0110e83fa3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json", + "sha256": "24742955efffffdc425477dca23489862ed24b175479ce5f6f177b8b65586216", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json", + "sha256": "e5e2e77e662c42901d7e591506394b55781dcb5d352843e16e6625ecaebcc47e", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json", + "sha256": "5e4c82793244d7b38961e000c035328a281aa4b283ac15eb066b665634131bd4", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json", + "sha256": "2fcf0161aebe500bcc4120d76db9e65431899c52d00e8db04372211236a1b8e3", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json", + "sha256": "57f8bbe6f9b6bfeabfb0ef7b4c87d415dbee6bd95e8746e2a810b81b13d391d1", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json", + "sha256": "b34dd4ac1b2b02cbeb2636327c352fd6a707a5bab284e9de3865e631c7665549", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json", + "sha256": "2f248ed36e4385ebd9fcb47bedba1dac7496ee282557bb1a55597fa507ed015b", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json", + "sha256": "d3c64355bc9d6a447db8cd9703dd32f3df69654a464f04d394dc26a104201aa6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json", + "sha256": "da539ccf3266f283b7354eb13631a14e5efe3fcc4727fb5ec11b31a4a61ec798", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json", + "sha256": "ab817caa1f25cfcf5d4e6d69ff4ccb7d6dafcfd654427a8964c4a5f8f56fdce5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json", + "sha256": "4003cc7e63db9fdd56b483f7e37377216821895bf214f1bb57a47d2d0327e3f6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json", + "sha256": "253b0145370ccdb6916f85d5288ad3c192ab63e78dd3be9f6925b1ab65b2d2e0", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json", + "sha256": "5a4e3efedb25e25731d0122b437eabdb6c8078ee4facd46c3a9f551a7f99732a", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json", + "sha256": "2ac8b1a1f00e76241ff9623f76405374a30f3a0804593040576b705a2aa97f56", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json", + "sha256": "ff2f34ac82e86d6901244f87962244340bc6e40b177e6d2b010da03829569426", + "role": "16x declared parameter count", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json", + "sha256": "0b676b2bd9baea8e9276279efe925d836c593607a3a46dd5196ca7eac7670bb3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json", + "sha256": "421e3232a98c6a15ce4781567f91cf94e6278d930a5a590ca94a9b38573247cc", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json", + "sha256": "51f05a0525cf71010d9f6364621ca518a6db583e45ff4b6639140b3fa39dfdd0", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json", + "sha256": "4126f2d70c0fd8300cd1083c4b1e3b0736fcb6cd71bb4b5c07f076174b440ae7", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json", + "sha256": "42b3f10470db8faafb7390f2fb1bf76ce3e33feb65576a6acf336985e7e7c7c8", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json", + "sha256": "b4e7f8d2a8ad984cbfe16b12709d49465824b8120d23c30e27e96b3f47fddf3b", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json", + "sha256": "a657f884b950151baeea56ec3eade6615160da829aa41bc1a9045576613850cf", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json", + "sha256": "45e228d2e95b65a59c6e7a75c15c110e9aa9aaba8b2ca0273872189c29f790f0", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json", + "sha256": "89d9bd8b9bab76d5c59f16a44b81e772ae3342c526b6b51eafe0a90d6ecc03f7", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json", + "sha256": "2296899ba51f26a553205ef325d726e5ca7608cd21815b031d97675208273a97", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json", + "sha256": "3e020261d15f330bdf312b3a7737b9d73412fbb60da240497af89c80c35799f9", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json", + "sha256": "b3986d40398420f2f5d3d77cbe4101f5e0e7acf85fcd8f5815f0ff4cd9a55d66", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json", + "sha256": "6ed461a488da9c64e58367f82e329c8544869f48e0cbe158bc3422d9f661576f", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv", + "sha256": "7c8805749d4484d1eea3503a8410c09572d799b219401f48b17c9bf56533e444", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv", + "sha256": "4c9af91a9497c5d2335c1ca79b90e217b90fb0ded9bc330d0f4970554ed5d7ec", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv", + "sha256": "0e2deb3606b58e45cdbdae9077fb3a34f3c67fb09eda1d9589843c7e4cfd09da", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv", + "sha256": "841bd71e2768ff64157fb765691dd3d45546a0ed7bab5476da0910e314d5aea7", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv", + "sha256": "e6a124c41a4c31931de181a99d9b77b820dc00f8ca975e83e370fadbabad4ab5", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "sha256": "770df0d5e3d6dbdba0ccedb9317a807167a66b02a82005255b9c5351ca1e5f4c", + "role": "reproducible CPU-only enrichment renderer", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py" + } + ], + "deck_group": "Figure 2", + "layout": "full", + "addition_key": "enrichment:arch_context_vs_trunk", + "parent_enriched_page": 13, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/arch_context_vs_trunk.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 13, + "parent_154_page": 15, + "parent_155_page": 16 + }, + { + "key": "arch_hint_injection", + "kind": "content", + "type": "content", + "title": "Predicted and oracle hints: one trained denoiser", + "section": "Figure 2", + "section_label": "Fig. 2b", + "figures": [ + "figures/final_palette_20261013/arch_hint_injection.png" + ], + "keywords": [], + "footnote": "Pred/oracle share hint-trained weights. Oracle reads target scANVI fractions; the separate predictor's training time is unavailable.", + "footnote_full": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Training sample fractions are joined by recipe block_id\u2192sample_key. Pred is corpus-trained MLP (train5232/val648/test627 samples), chosen by validation L1 against ridge; log1p(CPM), train-standardized, two hidden GELU/dropout layers and softmax. Oracle aggregates target donor observed AIFI-L1 labels over nine categories. Hint is appended after base G+6, before 2G measurement context; first condition layer 18015\u2192512. Baseline denoiser and hint denoiser are separately trained; pred and oracle share hint denoiser. Predictor params3206153 calculated from recorded architecture, excluded from denoiser count.", + "notes": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Training sample fractions are joined by recipe block_id\u2192sample_key. Pred is corpus-trained MLP (train5232/val648/test627 samples), chosen by validation L1 against ridge; log1p(CPM), train-standardized, two hidden GELU/dropout layers and softmax. Oracle aggregates target donor observed AIFI-L1 labels over nine categories. Hint is appended after base G+6, before 2G measurement context; first condition layer 18015\u2192512. Baseline denoiser and hint denoiser are separately trained; pred and oracle share hint denoiser. Predictor params3206153 calculated from recorded architecture, excluded from denoiser count.", + "after_current": 11, + "after_current_page": 11, + "palette": "native", + "role": "hint_diagram", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "replace_current_page": null, + "replacement_previous_131_page": null, + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py", + "sha256": "c3d725b97d003b9b2fb45d5061d5f7daffe89fd37147f9d53753b3ed2da77ffa", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py", + "sha256": "5e4f6169410fa165e5b1c39391ab731c38c32b7d3202618da326b2f82131e6de", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py", + "sha256": "0d959a35175549a263e1b23c2709bfe2699dc293141d5699f01cf70cd7a0ba73", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py", + "sha256": "7a352d73728e78c9d262e29c35628d2437af0bc7743b88c077ba81d911dd2a52", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py", + "sha256": "09631a39fb388b56c4be88814559d82fac100660571a4b2a8e74e9972c400758", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md", + "sha256": "7fcae090a1e4417eb52c7a51556f0c82bacdc3bcf47193f41be2c30d350c357b", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md", + "sha256": "8c24781e40dd4284f06a07b9d7eac35dae2cfe013e1b46bd5a1805fd9afe83db", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py", + "sha256": "f023dc2fa788c8bac628fa922e5a34ccf05d355edec2a57807ca9bb24f3c241e", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py", + "sha256": "044cf1efdb17761914b1707861aa606095ca4a8ede109bb337782395cff01c47", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py", + "sha256": "271834f39e5afb30c7c9e377b0b0f9aba06a85f8fe566a76ff263649f9a09ef0", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py", + "sha256": "95ace096cdc58a9ca92a18af649c60fe6ac0a55e97f977e7ab9bcb04efa968cb", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json", + "sha256": "8377fa821211086a56b3450af3a464f55a5799c6df1b2486019dc96d46331587", + "role": "saved separate hint predictor metrics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json", + "sha256": "af11b163a08666849621e7e66d89b2234314231bbecf976afa6f4b0110e83fa3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json", + "sha256": "24742955efffffdc425477dca23489862ed24b175479ce5f6f177b8b65586216", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json", + "sha256": "e5e2e77e662c42901d7e591506394b55781dcb5d352843e16e6625ecaebcc47e", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json", + "sha256": "5e4c82793244d7b38961e000c035328a281aa4b283ac15eb066b665634131bd4", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json", + "sha256": "2fcf0161aebe500bcc4120d76db9e65431899c52d00e8db04372211236a1b8e3", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json", + "sha256": "57f8bbe6f9b6bfeabfb0ef7b4c87d415dbee6bd95e8746e2a810b81b13d391d1", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json", + "sha256": "b34dd4ac1b2b02cbeb2636327c352fd6a707a5bab284e9de3865e631c7665549", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json", + "sha256": "2f248ed36e4385ebd9fcb47bedba1dac7496ee282557bb1a55597fa507ed015b", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json", + "sha256": "d3c64355bc9d6a447db8cd9703dd32f3df69654a464f04d394dc26a104201aa6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json", + "sha256": "da539ccf3266f283b7354eb13631a14e5efe3fcc4727fb5ec11b31a4a61ec798", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json", + "sha256": "ab817caa1f25cfcf5d4e6d69ff4ccb7d6dafcfd654427a8964c4a5f8f56fdce5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json", + "sha256": "4003cc7e63db9fdd56b483f7e37377216821895bf214f1bb57a47d2d0327e3f6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json", + "sha256": "253b0145370ccdb6916f85d5288ad3c192ab63e78dd3be9f6925b1ab65b2d2e0", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json", + "sha256": "5a4e3efedb25e25731d0122b437eabdb6c8078ee4facd46c3a9f551a7f99732a", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json", + "sha256": "2ac8b1a1f00e76241ff9623f76405374a30f3a0804593040576b705a2aa97f56", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json", + "sha256": "ff2f34ac82e86d6901244f87962244340bc6e40b177e6d2b010da03829569426", + "role": "16x declared parameter count", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json", + "sha256": "0b676b2bd9baea8e9276279efe925d836c593607a3a46dd5196ca7eac7670bb3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json", + "sha256": "421e3232a98c6a15ce4781567f91cf94e6278d930a5a590ca94a9b38573247cc", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json", + "sha256": "51f05a0525cf71010d9f6364621ca518a6db583e45ff4b6639140b3fa39dfdd0", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json", + "sha256": "4126f2d70c0fd8300cd1083c4b1e3b0736fcb6cd71bb4b5c07f076174b440ae7", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json", + "sha256": "42b3f10470db8faafb7390f2fb1bf76ce3e33feb65576a6acf336985e7e7c7c8", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json", + "sha256": "b4e7f8d2a8ad984cbfe16b12709d49465824b8120d23c30e27e96b3f47fddf3b", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json", + "sha256": "a657f884b950151baeea56ec3eade6615160da829aa41bc1a9045576613850cf", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json", + "sha256": "45e228d2e95b65a59c6e7a75c15c110e9aa9aaba8b2ca0273872189c29f790f0", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json", + "sha256": "89d9bd8b9bab76d5c59f16a44b81e772ae3342c526b6b51eafe0a90d6ecc03f7", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json", + "sha256": "2296899ba51f26a553205ef325d726e5ca7608cd21815b031d97675208273a97", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json", + "sha256": "3e020261d15f330bdf312b3a7737b9d73412fbb60da240497af89c80c35799f9", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json", + "sha256": "b3986d40398420f2f5d3d77cbe4101f5e0e7acf85fcd8f5815f0ff4cd9a55d66", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json", + "sha256": "6ed461a488da9c64e58367f82e329c8544869f48e0cbe158bc3422d9f661576f", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv", + "sha256": "7c8805749d4484d1eea3503a8410c09572d799b219401f48b17c9bf56533e444", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv", + "sha256": "4c9af91a9497c5d2335c1ca79b90e217b90fb0ded9bc330d0f4970554ed5d7ec", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv", + "sha256": "0e2deb3606b58e45cdbdae9077fb3a34f3c67fb09eda1d9589843c7e4cfd09da", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv", + "sha256": "841bd71e2768ff64157fb765691dd3d45546a0ed7bab5476da0910e314d5aea7", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv", + "sha256": "e6a124c41a4c31931de181a99d9b77b820dc00f8ca975e83e370fadbabad4ab5", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "sha256": "770df0d5e3d6dbdba0ccedb9317a807167a66b02a82005255b9c5351ca1e5f4c", + "role": "reproducible CPU-only enrichment renderer", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py" + } + ], + "deck_group": "Figure 2", + "layout": "full", + "addition_key": "enrichment:arch_hint_injection", + "parent_enriched_page": 14, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/arch_hint_injection.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 14, + "parent_154_page": 16, + "parent_155_page": 17 + }, + { + "kind": "content", + "section": "Figure 2", + "section_label": "Fig. 2c", + "title": "Architecture: full and late validation progression", + "figures": [ + "figures/final_palette_20261013/arch_width_loss_progression.png" + ], + "footnote": "Corpus validation; same 6k genes. Hollow ring: within 1% of own E16 thereafter. 16\u00d7 stops at E2; saturation remains NA.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "original_slide": 10, + "original_title": "Width & context ablation: loss saturation", + "original_footnote": "Old 6k panel (1% floor); one validation set. 16x: memory-floor stop after part 2.", + "footnote_full": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Full saved E1\u2013E16 curves with a late-region view. No smoothing, curve fitting or extrapolation. One training seed. Endpoint proximity is a retrospective summary, not established convergence; tolerance sensitivity is in arm register.", + "source_page_131": 12, + "deck_group": "Figure 2", + "previous_section_label": "Fig. 1e", + "parent_page": 43, + "key": "arch_width_loss_progression", + "type": "content", + "keywords": [], + "notes": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Full saved E1\u2013E16 curves with a late-region view. No smoothing, curve fitting or extrapolation. One training seed. Endpoint proximity is a retrospective summary, not established convergence; tolerance sensitivity is in arm register.", + "after_current": 12, + "after_current_page": 12, + "palette": "native", + "role": "width_validation", + "replace_current_page": 43, + "replacement_previous_131_page": 12, + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py", + "sha256": "c3d725b97d003b9b2fb45d5061d5f7daffe89fd37147f9d53753b3ed2da77ffa", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py", + "sha256": "5e4f6169410fa165e5b1c39391ab731c38c32b7d3202618da326b2f82131e6de", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py", + "sha256": "0d959a35175549a263e1b23c2709bfe2699dc293141d5699f01cf70cd7a0ba73", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py", + "sha256": "7a352d73728e78c9d262e29c35628d2437af0bc7743b88c077ba81d911dd2a52", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py", + "sha256": "09631a39fb388b56c4be88814559d82fac100660571a4b2a8e74e9972c400758", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md", + "sha256": "7fcae090a1e4417eb52c7a51556f0c82bacdc3bcf47193f41be2c30d350c357b", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md", + "sha256": "8c24781e40dd4284f06a07b9d7eac35dae2cfe013e1b46bd5a1805fd9afe83db", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py", + "sha256": "f023dc2fa788c8bac628fa922e5a34ccf05d355edec2a57807ca9bb24f3c241e", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py", + "sha256": "044cf1efdb17761914b1707861aa606095ca4a8ede109bb337782395cff01c47", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py", + "sha256": "271834f39e5afb30c7c9e377b0b0f9aba06a85f8fe566a76ff263649f9a09ef0", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py", + "sha256": "95ace096cdc58a9ca92a18af649c60fe6ac0a55e97f977e7ab9bcb04efa968cb", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json", + "sha256": "8377fa821211086a56b3450af3a464f55a5799c6df1b2486019dc96d46331587", + "role": "saved separate hint predictor metrics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json", + "sha256": "af11b163a08666849621e7e66d89b2234314231bbecf976afa6f4b0110e83fa3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json", + "sha256": "24742955efffffdc425477dca23489862ed24b175479ce5f6f177b8b65586216", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json", + "sha256": "e5e2e77e662c42901d7e591506394b55781dcb5d352843e16e6625ecaebcc47e", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json", + "sha256": "5e4c82793244d7b38961e000c035328a281aa4b283ac15eb066b665634131bd4", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json", + "sha256": "2fcf0161aebe500bcc4120d76db9e65431899c52d00e8db04372211236a1b8e3", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json", + "sha256": "57f8bbe6f9b6bfeabfb0ef7b4c87d415dbee6bd95e8746e2a810b81b13d391d1", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json", + "sha256": "b34dd4ac1b2b02cbeb2636327c352fd6a707a5bab284e9de3865e631c7665549", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json", + "sha256": "2f248ed36e4385ebd9fcb47bedba1dac7496ee282557bb1a55597fa507ed015b", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json", + "sha256": "d3c64355bc9d6a447db8cd9703dd32f3df69654a464f04d394dc26a104201aa6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json", + "sha256": "da539ccf3266f283b7354eb13631a14e5efe3fcc4727fb5ec11b31a4a61ec798", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json", + "sha256": "ab817caa1f25cfcf5d4e6d69ff4ccb7d6dafcfd654427a8964c4a5f8f56fdce5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json", + "sha256": "4003cc7e63db9fdd56b483f7e37377216821895bf214f1bb57a47d2d0327e3f6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json", + "sha256": "253b0145370ccdb6916f85d5288ad3c192ab63e78dd3be9f6925b1ab65b2d2e0", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json", + "sha256": "5a4e3efedb25e25731d0122b437eabdb6c8078ee4facd46c3a9f551a7f99732a", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json", + "sha256": "2ac8b1a1f00e76241ff9623f76405374a30f3a0804593040576b705a2aa97f56", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json", + "sha256": "ff2f34ac82e86d6901244f87962244340bc6e40b177e6d2b010da03829569426", + "role": "16x declared parameter count", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json", + "sha256": "0b676b2bd9baea8e9276279efe925d836c593607a3a46dd5196ca7eac7670bb3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json", + "sha256": "421e3232a98c6a15ce4781567f91cf94e6278d930a5a590ca94a9b38573247cc", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json", + "sha256": "51f05a0525cf71010d9f6364621ca518a6db583e45ff4b6639140b3fa39dfdd0", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json", + "sha256": "4126f2d70c0fd8300cd1083c4b1e3b0736fcb6cd71bb4b5c07f076174b440ae7", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json", + "sha256": "42b3f10470db8faafb7390f2fb1bf76ce3e33feb65576a6acf336985e7e7c7c8", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json", + "sha256": "b4e7f8d2a8ad984cbfe16b12709d49465824b8120d23c30e27e96b3f47fddf3b", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json", + "sha256": "a657f884b950151baeea56ec3eade6615160da829aa41bc1a9045576613850cf", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json", + "sha256": "45e228d2e95b65a59c6e7a75c15c110e9aa9aaba8b2ca0273872189c29f790f0", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json", + "sha256": "89d9bd8b9bab76d5c59f16a44b81e772ae3342c526b6b51eafe0a90d6ecc03f7", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json", + "sha256": "2296899ba51f26a553205ef325d726e5ca7608cd21815b031d97675208273a97", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json", + "sha256": "3e020261d15f330bdf312b3a7737b9d73412fbb60da240497af89c80c35799f9", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json", + "sha256": "b3986d40398420f2f5d3d77cbe4101f5e0e7acf85fcd8f5815f0ff4cd9a55d66", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json", + "sha256": "6ed461a488da9c64e58367f82e329c8544869f48e0cbe158bc3422d9f661576f", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv", + "sha256": "7c8805749d4484d1eea3503a8410c09572d799b219401f48b17c9bf56533e444", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv", + "sha256": "4c9af91a9497c5d2335c1ca79b90e217b90fb0ded9bc330d0f4970554ed5d7ec", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv", + "sha256": "0e2deb3606b58e45cdbdae9077fb3a34f3c67fb09eda1d9589843c7e4cfd09da", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv", + "sha256": "841bd71e2768ff64157fb765691dd3d45546a0ed7bab5476da0910e314d5aea7", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv", + "sha256": "e6a124c41a4c31931de181a99d9b77b820dc00f8ca975e83e370fadbabad4ab5", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "sha256": "770df0d5e3d6dbdba0ccedb9317a807167a66b02a82005255b9c5351ca1e5f4c", + "role": "reproducible CPU-only enrichment renderer", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py" + } + ], + "layout": "full", + "previous_figures": [ + "figures/expanded_master_20261013/palette/sec03_width_loss.png" + ], + "parent_enriched_page": 15, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/arch_width_loss_progression.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 15, + "parent_154_page": 17, + "parent_155_page": 18 + }, + { + "kind": "content", + "section": "Figure 2", + "section_label": "Fig. 2d", + "title": "Width & context ablation: benchmark", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec03_width_bench.png" + ], + "keywords": [ + "trunk x4: loss \u221231%, no DEG gain", + "Prop hint: oracle = predicted, L1 worse", + "1x kept: ties ctx x4, 54M vs 102M" + ], + "footnote": "SLE, 42 donors, 2,980 genes; old 6k panel. ctx x4: better L1 (0.161 vs 0.197).", + "say": "Bars = mean of parts 14-16 with the cut p > 0.5; diamonds = Bernoulli read-out (4 width arms only). Prop hint = 9 L1 cell-type proportions; older panel, not the final model.", + "sources": "tools/sec03_width_bench.py; data/sec03_width_bench.csv, data/sec03_width_bench_summary.csv, data/sec03_width_bench_headline.csv", + "original_slide": 11, + "original_title": "Width & context ablation: benchmark", + "original_footnote": "SLE, 42 donors, 2,980 genes; old 6k panel. ctx x4: better L1 (0.161 vs 0.197).", + "footnote_full": "Original slide 11: SLE, 42 donors, 2,980 genes; old 6k panel. ctx x4: better L1 (0.161 vs 0.197).", + "figures_before_palette": [ + "figures/sec03_width_bench.png" + ], + "source_page_131": 13, + "deck_group": "Figure 2", + "previous_section_label": "Fig. 1e", + "parent_page": 12, + "parent_enriched_page": 16, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec03_width_bench.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 16, + "parent_154_page": 18, + "parent_155_page": 19 + }, + { + "key": "arch_width_arm_register", + "kind": "content", + "type": "content", + "title": "Architecture arms: dimensions and endpoints", + "section": "Figure 2", + "section_label": "Fig. 2 details", + "figures": [ + "figures/final_palette_20261013/arch_width_arm_register.png" + ], + "keywords": [], + "footnote": "Six independent training runs; pred/oracle reuse the hint run. 16\u00d7 incomplete; retired 2\u00d7 is not a completed result.", + "footnote_full": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Counts derived arithmetically from exact Linear/LayerNorm definitions and checked against all available RUN_STATUS parameters; 16\u00d7 checked against recorded manifest. Saturation columns use 1%,0.5%,2% relative to own E16. 16\u00d7 memory-floor failure preserved; discarded2\u00d7 design not counted as trained evidence.", + "notes": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Counts derived arithmetically from exact Linear/LayerNorm definitions and checked against all available RUN_STATUS parameters; 16\u00d7 checked against recorded manifest. Saturation columns use 1%,0.5%,2% relative to own E16. 16\u00d7 memory-floor failure preserved; discarded2\u00d7 design not counted as trained evidence.", + "after_current": 12, + "after_current_page": 12, + "palette": "native", + "role": "architecture_arm_table", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "replace_current_page": null, + "replacement_previous_131_page": null, + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py", + "sha256": "c3d725b97d003b9b2fb45d5061d5f7daffe89fd37147f9d53753b3ed2da77ffa", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py", + "sha256": "5e4f6169410fa165e5b1c39391ab731c38c32b7d3202618da326b2f82131e6de", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py", + "sha256": "0d959a35175549a263e1b23c2709bfe2699dc293141d5699f01cf70cd7a0ba73", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py", + "sha256": "7a352d73728e78c9d262e29c35628d2437af0bc7743b88c077ba81d911dd2a52", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py", + "sha256": "09631a39fb388b56c4be88814559d82fac100660571a4b2a8e74e9972c400758", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md", + "sha256": "7fcae090a1e4417eb52c7a51556f0c82bacdc3bcf47193f41be2c30d350c357b", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md", + "sha256": "8c24781e40dd4284f06a07b9d7eac35dae2cfe013e1b46bd5a1805fd9afe83db", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py", + "sha256": "f023dc2fa788c8bac628fa922e5a34ccf05d355edec2a57807ca9bb24f3c241e", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py", + "sha256": "044cf1efdb17761914b1707861aa606095ca4a8ede109bb337782395cff01c47", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py", + "sha256": "271834f39e5afb30c7c9e377b0b0f9aba06a85f8fe566a76ff263649f9a09ef0", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py", + "sha256": "95ace096cdc58a9ca92a18af649c60fe6ac0a55e97f977e7ab9bcb04efa968cb", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json", + "sha256": "8377fa821211086a56b3450af3a464f55a5799c6df1b2486019dc96d46331587", + "role": "saved separate hint predictor metrics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json", + "sha256": "af11b163a08666849621e7e66d89b2234314231bbecf976afa6f4b0110e83fa3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json", + "sha256": "24742955efffffdc425477dca23489862ed24b175479ce5f6f177b8b65586216", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json", + "sha256": "e5e2e77e662c42901d7e591506394b55781dcb5d352843e16e6625ecaebcc47e", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json", + "sha256": "5e4c82793244d7b38961e000c035328a281aa4b283ac15eb066b665634131bd4", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json", + "sha256": "2fcf0161aebe500bcc4120d76db9e65431899c52d00e8db04372211236a1b8e3", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json", + "sha256": "57f8bbe6f9b6bfeabfb0ef7b4c87d415dbee6bd95e8746e2a810b81b13d391d1", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json", + "sha256": "b34dd4ac1b2b02cbeb2636327c352fd6a707a5bab284e9de3865e631c7665549", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json", + "sha256": "2f248ed36e4385ebd9fcb47bedba1dac7496ee282557bb1a55597fa507ed015b", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json", + "sha256": "d3c64355bc9d6a447db8cd9703dd32f3df69654a464f04d394dc26a104201aa6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json", + "sha256": "da539ccf3266f283b7354eb13631a14e5efe3fcc4727fb5ec11b31a4a61ec798", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json", + "sha256": "ab817caa1f25cfcf5d4e6d69ff4ccb7d6dafcfd654427a8964c4a5f8f56fdce5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json", + "sha256": "4003cc7e63db9fdd56b483f7e37377216821895bf214f1bb57a47d2d0327e3f6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json", + "sha256": "253b0145370ccdb6916f85d5288ad3c192ab63e78dd3be9f6925b1ab65b2d2e0", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json", + "sha256": "5a4e3efedb25e25731d0122b437eabdb6c8078ee4facd46c3a9f551a7f99732a", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json", + "sha256": "2ac8b1a1f00e76241ff9623f76405374a30f3a0804593040576b705a2aa97f56", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json", + "sha256": "ff2f34ac82e86d6901244f87962244340bc6e40b177e6d2b010da03829569426", + "role": "16x declared parameter count", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json", + "sha256": "0b676b2bd9baea8e9276279efe925d836c593607a3a46dd5196ca7eac7670bb3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json", + "sha256": "421e3232a98c6a15ce4781567f91cf94e6278d930a5a590ca94a9b38573247cc", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json", + "sha256": "51f05a0525cf71010d9f6364621ca518a6db583e45ff4b6639140b3fa39dfdd0", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json", + "sha256": "4126f2d70c0fd8300cd1083c4b1e3b0736fcb6cd71bb4b5c07f076174b440ae7", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json", + "sha256": "42b3f10470db8faafb7390f2fb1bf76ce3e33feb65576a6acf336985e7e7c7c8", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json", + "sha256": "b4e7f8d2a8ad984cbfe16b12709d49465824b8120d23c30e27e96b3f47fddf3b", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json", + "sha256": "a657f884b950151baeea56ec3eade6615160da829aa41bc1a9045576613850cf", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json", + "sha256": "45e228d2e95b65a59c6e7a75c15c110e9aa9aaba8b2ca0273872189c29f790f0", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json", + "sha256": "89d9bd8b9bab76d5c59f16a44b81e772ae3342c526b6b51eafe0a90d6ecc03f7", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json", + "sha256": "2296899ba51f26a553205ef325d726e5ca7608cd21815b031d97675208273a97", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json", + "sha256": "3e020261d15f330bdf312b3a7737b9d73412fbb60da240497af89c80c35799f9", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json", + "sha256": "b3986d40398420f2f5d3d77cbe4101f5e0e7acf85fcd8f5815f0ff4cd9a55d66", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json", + "sha256": "6ed461a488da9c64e58367f82e329c8544869f48e0cbe158bc3422d9f661576f", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv", + "sha256": "7c8805749d4484d1eea3503a8410c09572d799b219401f48b17c9bf56533e444", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv", + "sha256": "4c9af91a9497c5d2335c1ca79b90e217b90fb0ded9bc330d0f4970554ed5d7ec", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv", + "sha256": "0e2deb3606b58e45cdbdae9077fb3a34f3c67fb09eda1d9589843c7e4cfd09da", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv", + "sha256": "841bd71e2768ff64157fb765691dd3d45546a0ed7bab5476da0910e314d5aea7", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv", + "sha256": "e6a124c41a4c31931de181a99d9b77b820dc00f8ca975e83e370fadbabad4ab5", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "sha256": "770df0d5e3d6dbdba0ccedb9317a807167a66b02a82005255b9c5351ca1e5f4c", + "role": "reproducible CPU-only enrichment renderer", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py" + } + ], + "deck_group": "Figure 2", + "layout": "full", + "addition_key": "enrichment:arch_width_arm_register", + "parent_enriched_page": 17, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/arch_width_arm_register.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 17, + "parent_154_page": 19, + "parent_155_page": 20 + }, + { + "kind": "content", + "section": "Figure 2", + "section_label": "Fig. 2e", + "title": "Gene context: 2k / 3k / 6k / 15k", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec03_panel.png" + ], + "keywords": [ + "6k \u2248 3k: 3k r, DEG inside 6k seed range", + "2k best L1; 15k worst r, L1", + "6k kept: first panel with CLEC9A (cDC1)" + ], + "footnote": "Existing panel benchmark. 6k: five sampling seeds; other panels: one seed. Different axis from the architecture sweep.", + "say": "Runs f01_{2k,3k,6k,p15k}_1x, part 16, Bernoulli, EMA, DDIM 50. Different gene axis from the previous slide, so the two slides cannot be compared. Loss curves are divided by each panel's own part-16 value.", + "sources": "tools/sec03_panel.py; data/sec03_panel_bench.csv, data/sec03_panel_seeds.csv, data/sec03_panel_loss.csv, data/sec03_panel_headline.csv", + "original_slide": 12, + "original_title": "Gene context: 2k / 3k / 6k / 15k", + "original_footnote": "SLE, 42 donors, 1,006 genes in every panel. 6k: 5 seeds; base seed best of 5 on r and DEG.", + "footnote_full": "Current scope: Existing panel benchmark. 6k: five sampling seeds; other panels: one seed. Different axis from the architecture sweep.\n\nOriginal slide 12: SLE, 42 donors, 1,006 genes in every panel. 6k: 5 seeds; base seed best of 5 on r and DEG.", + "figures_before_palette": [ + "figures/sec03_panel.png" + ], + "source_page_131": 14, + "deck_group": "Figure 2", + "previous_section_label": "Fig. 1f", + "parent_page": 13, + "parent_enriched_page": 18, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec03_panel.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 18, + "parent_154_page": 20, + "parent_155_page": 21 + }, + { + "key": "arch_panel_loss_progression", + "kind": "content", + "type": "content", + "title": "Gene panels: progression relative to each endpoint", + "section": "Figure 2", + "section_label": "Fig. 2f", + "figures": [ + "figures/final_palette_20261013/arch_panel_loss_progression.png" + ], + "keywords": [], + "footnote": "Corpus validation, normalized to each panel's E16. Different gene sets prevent comparing absolute losses across panels.", + "footnote_full": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. All four panels have 16 saved parts. Curves use each panel's own denominator; this compares progression shape, not identical-objective performance. One training seed for every panel; five 6k generation seeds are not five training repetitions.", + "notes": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. All four panels have 16 saved parts. Curves use each panel's own denominator; this compares progression shape, not identical-objective performance. One training seed for every panel; five 6k generation seeds are not five training repetitions.", + "after_current": 13, + "after_current_page": 13, + "palette": "native", + "role": "gene_validation", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "replace_current_page": null, + "replacement_previous_131_page": null, + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py", + "sha256": "c3d725b97d003b9b2fb45d5061d5f7daffe89fd37147f9d53753b3ed2da77ffa", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py", + "sha256": "5e4f6169410fa165e5b1c39391ab731c38c32b7d3202618da326b2f82131e6de", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py", + "sha256": "0d959a35175549a263e1b23c2709bfe2699dc293141d5699f01cf70cd7a0ba73", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py", + "sha256": "7a352d73728e78c9d262e29c35628d2437af0bc7743b88c077ba81d911dd2a52", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py", + "sha256": "09631a39fb388b56c4be88814559d82fac100660571a4b2a8e74e9972c400758", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md", + "sha256": "7fcae090a1e4417eb52c7a51556f0c82bacdc3bcf47193f41be2c30d350c357b", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md", + "sha256": "8c24781e40dd4284f06a07b9d7eac35dae2cfe013e1b46bd5a1805fd9afe83db", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py", + "sha256": "f023dc2fa788c8bac628fa922e5a34ccf05d355edec2a57807ca9bb24f3c241e", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py", + "sha256": "044cf1efdb17761914b1707861aa606095ca4a8ede109bb337782395cff01c47", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py", + "sha256": "271834f39e5afb30c7c9e377b0b0f9aba06a85f8fe566a76ff263649f9a09ef0", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py", + "sha256": "95ace096cdc58a9ca92a18af649c60fe6ac0a55e97f977e7ab9bcb04efa968cb", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json", + "sha256": "8377fa821211086a56b3450af3a464f55a5799c6df1b2486019dc96d46331587", + "role": "saved separate hint predictor metrics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json", + "sha256": "af11b163a08666849621e7e66d89b2234314231bbecf976afa6f4b0110e83fa3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json", + "sha256": "24742955efffffdc425477dca23489862ed24b175479ce5f6f177b8b65586216", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json", + "sha256": "e5e2e77e662c42901d7e591506394b55781dcb5d352843e16e6625ecaebcc47e", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json", + "sha256": "5e4c82793244d7b38961e000c035328a281aa4b283ac15eb066b665634131bd4", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json", + "sha256": "2fcf0161aebe500bcc4120d76db9e65431899c52d00e8db04372211236a1b8e3", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json", + "sha256": "57f8bbe6f9b6bfeabfb0ef7b4c87d415dbee6bd95e8746e2a810b81b13d391d1", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json", + "sha256": "b34dd4ac1b2b02cbeb2636327c352fd6a707a5bab284e9de3865e631c7665549", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json", + "sha256": "2f248ed36e4385ebd9fcb47bedba1dac7496ee282557bb1a55597fa507ed015b", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json", + "sha256": "d3c64355bc9d6a447db8cd9703dd32f3df69654a464f04d394dc26a104201aa6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json", + "sha256": "da539ccf3266f283b7354eb13631a14e5efe3fcc4727fb5ec11b31a4a61ec798", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json", + "sha256": "ab817caa1f25cfcf5d4e6d69ff4ccb7d6dafcfd654427a8964c4a5f8f56fdce5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json", + "sha256": "4003cc7e63db9fdd56b483f7e37377216821895bf214f1bb57a47d2d0327e3f6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json", + "sha256": "253b0145370ccdb6916f85d5288ad3c192ab63e78dd3be9f6925b1ab65b2d2e0", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json", + "sha256": "5a4e3efedb25e25731d0122b437eabdb6c8078ee4facd46c3a9f551a7f99732a", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json", + "sha256": "2ac8b1a1f00e76241ff9623f76405374a30f3a0804593040576b705a2aa97f56", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json", + "sha256": "ff2f34ac82e86d6901244f87962244340bc6e40b177e6d2b010da03829569426", + "role": "16x declared parameter count", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json", + "sha256": "0b676b2bd9baea8e9276279efe925d836c593607a3a46dd5196ca7eac7670bb3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json", + "sha256": "421e3232a98c6a15ce4781567f91cf94e6278d930a5a590ca94a9b38573247cc", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json", + "sha256": "51f05a0525cf71010d9f6364621ca518a6db583e45ff4b6639140b3fa39dfdd0", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json", + "sha256": "4126f2d70c0fd8300cd1083c4b1e3b0736fcb6cd71bb4b5c07f076174b440ae7", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json", + "sha256": "42b3f10470db8faafb7390f2fb1bf76ce3e33feb65576a6acf336985e7e7c7c8", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json", + "sha256": "b4e7f8d2a8ad984cbfe16b12709d49465824b8120d23c30e27e96b3f47fddf3b", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json", + "sha256": "a657f884b950151baeea56ec3eade6615160da829aa41bc1a9045576613850cf", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json", + "sha256": "45e228d2e95b65a59c6e7a75c15c110e9aa9aaba8b2ca0273872189c29f790f0", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json", + "sha256": "89d9bd8b9bab76d5c59f16a44b81e772ae3342c526b6b51eafe0a90d6ecc03f7", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json", + "sha256": "2296899ba51f26a553205ef325d726e5ca7608cd21815b031d97675208273a97", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json", + "sha256": "3e020261d15f330bdf312b3a7737b9d73412fbb60da240497af89c80c35799f9", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json", + "sha256": "b3986d40398420f2f5d3d77cbe4101f5e0e7acf85fcd8f5815f0ff4cd9a55d66", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json", + "sha256": "6ed461a488da9c64e58367f82e329c8544869f48e0cbe158bc3422d9f661576f", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv", + "sha256": "7c8805749d4484d1eea3503a8410c09572d799b219401f48b17c9bf56533e444", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv", + "sha256": "4c9af91a9497c5d2335c1ca79b90e217b90fb0ded9bc330d0f4970554ed5d7ec", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv", + "sha256": "0e2deb3606b58e45cdbdae9077fb3a34f3c67fb09eda1d9589843c7e4cfd09da", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv", + "sha256": "841bd71e2768ff64157fb765691dd3d45546a0ed7bab5476da0910e314d5aea7", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv", + "sha256": "e6a124c41a4c31931de181a99d9b77b820dc00f8ca975e83e370fadbabad4ab5", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "sha256": "770df0d5e3d6dbdba0ccedb9317a807167a66b02a82005255b9c5351ca1e5f4c", + "role": "reproducible CPU-only enrichment renderer", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py" + } + ], + "deck_group": "Figure 2", + "layout": "full", + "addition_key": "enrichment:arch_panel_loss_progression", + "parent_enriched_page": 19, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/arch_panel_loss_progression.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 19, + "parent_154_page": 21, + "parent_155_page": 22 + }, + { + "key": "arch_panel_arm_register", + "kind": "content", + "type": "content", + "title": "Gene-panel arms: exact sizes and training provenance", + "section": "Figure 2", + "section_label": "Fig. 2 details", + "figures": [ + "figures/final_palette_20261013/arch_panel_arm_register.png" + ], + "keywords": [], + "footnote": "Four separately trained panels; 6k sampling seeds reuse one checkpoint. Saturation is endpoint-relative, not proof of convergence.", + "footnote_full": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Nominal15k is15114 genes. Fixed trunk/context widths; parameters change because input/output and condition projections depend on G. Architecture6k 1% and final6k0.1% are distinct panels/checkpoints despite equal parameter counts.", + "notes": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. Nominal15k is15114 genes. Fixed trunk/context widths; parameters change because input/output and condition projections depend on G. Architecture6k 1% and final6k0.1% are distinct panels/checkpoints despite equal parameter counts.", + "after_current": 13, + "after_current_page": 13, + "palette": "native", + "role": "gene_arm_table", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "replace_current_page": null, + "replacement_previous_131_page": null, + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py", + "sha256": "c3d725b97d003b9b2fb45d5061d5f7daffe89fd37147f9d53753b3ed2da77ffa", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py", + "sha256": "5e4f6169410fa165e5b1c39391ab731c38c32b7d3202618da326b2f82131e6de", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py", + "sha256": "0d959a35175549a263e1b23c2709bfe2699dc293141d5699f01cf70cd7a0ba73", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py", + "sha256": "7a352d73728e78c9d262e29c35628d2437af0bc7743b88c077ba81d911dd2a52", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py", + "sha256": "09631a39fb388b56c4be88814559d82fac100660571a4b2a8e74e9972c400758", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md", + "sha256": "7fcae090a1e4417eb52c7a51556f0c82bacdc3bcf47193f41be2c30d350c357b", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md", + "sha256": "8c24781e40dd4284f06a07b9d7eac35dae2cfe013e1b46bd5a1805fd9afe83db", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py", + "sha256": "f023dc2fa788c8bac628fa922e5a34ccf05d355edec2a57807ca9bb24f3c241e", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py", + "sha256": "044cf1efdb17761914b1707861aa606095ca4a8ede109bb337782395cff01c47", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py", + "sha256": "271834f39e5afb30c7c9e377b0b0f9aba06a85f8fe566a76ff263649f9a09ef0", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py", + "sha256": "95ace096cdc58a9ca92a18af649c60fe6ac0a55e97f977e7ab9bcb04efa968cb", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json", + "sha256": "8377fa821211086a56b3450af3a464f55a5799c6df1b2486019dc96d46331587", + "role": "saved separate hint predictor metrics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json", + "sha256": "af11b163a08666849621e7e66d89b2234314231bbecf976afa6f4b0110e83fa3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json", + "sha256": "24742955efffffdc425477dca23489862ed24b175479ce5f6f177b8b65586216", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json", + "sha256": "e5e2e77e662c42901d7e591506394b55781dcb5d352843e16e6625ecaebcc47e", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json", + "sha256": "5e4c82793244d7b38961e000c035328a281aa4b283ac15eb066b665634131bd4", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json", + "sha256": "2fcf0161aebe500bcc4120d76db9e65431899c52d00e8db04372211236a1b8e3", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json", + "sha256": "57f8bbe6f9b6bfeabfb0ef7b4c87d415dbee6bd95e8746e2a810b81b13d391d1", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json", + "sha256": "b34dd4ac1b2b02cbeb2636327c352fd6a707a5bab284e9de3865e631c7665549", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json", + "sha256": "2f248ed36e4385ebd9fcb47bedba1dac7496ee282557bb1a55597fa507ed015b", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json", + "sha256": "d3c64355bc9d6a447db8cd9703dd32f3df69654a464f04d394dc26a104201aa6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json", + "sha256": "da539ccf3266f283b7354eb13631a14e5efe3fcc4727fb5ec11b31a4a61ec798", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json", + "sha256": "ab817caa1f25cfcf5d4e6d69ff4ccb7d6dafcfd654427a8964c4a5f8f56fdce5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json", + "sha256": "4003cc7e63db9fdd56b483f7e37377216821895bf214f1bb57a47d2d0327e3f6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json", + "sha256": "253b0145370ccdb6916f85d5288ad3c192ab63e78dd3be9f6925b1ab65b2d2e0", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json", + "sha256": "5a4e3efedb25e25731d0122b437eabdb6c8078ee4facd46c3a9f551a7f99732a", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json", + "sha256": "2ac8b1a1f00e76241ff9623f76405374a30f3a0804593040576b705a2aa97f56", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json", + "sha256": "ff2f34ac82e86d6901244f87962244340bc6e40b177e6d2b010da03829569426", + "role": "16x declared parameter count", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json", + "sha256": "0b676b2bd9baea8e9276279efe925d836c593607a3a46dd5196ca7eac7670bb3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json", + "sha256": "421e3232a98c6a15ce4781567f91cf94e6278d930a5a590ca94a9b38573247cc", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json", + "sha256": "51f05a0525cf71010d9f6364621ca518a6db583e45ff4b6639140b3fa39dfdd0", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json", + "sha256": "4126f2d70c0fd8300cd1083c4b1e3b0736fcb6cd71bb4b5c07f076174b440ae7", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json", + "sha256": "42b3f10470db8faafb7390f2fb1bf76ce3e33feb65576a6acf336985e7e7c7c8", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json", + "sha256": "b4e7f8d2a8ad984cbfe16b12709d49465824b8120d23c30e27e96b3f47fddf3b", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json", + "sha256": "a657f884b950151baeea56ec3eade6615160da829aa41bc1a9045576613850cf", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json", + "sha256": "45e228d2e95b65a59c6e7a75c15c110e9aa9aaba8b2ca0273872189c29f790f0", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json", + "sha256": "89d9bd8b9bab76d5c59f16a44b81e772ae3342c526b6b51eafe0a90d6ecc03f7", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json", + "sha256": "2296899ba51f26a553205ef325d726e5ca7608cd21815b031d97675208273a97", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json", + "sha256": "3e020261d15f330bdf312b3a7737b9d73412fbb60da240497af89c80c35799f9", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json", + "sha256": "b3986d40398420f2f5d3d77cbe4101f5e0e7acf85fcd8f5815f0ff4cd9a55d66", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json", + "sha256": "6ed461a488da9c64e58367f82e329c8544869f48e0cbe158bc3422d9f661576f", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv", + "sha256": "7c8805749d4484d1eea3503a8410c09572d799b219401f48b17c9bf56533e444", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv", + "sha256": "4c9af91a9497c5d2335c1ca79b90e217b90fb0ded9bc330d0f4970554ed5d7ec", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv", + "sha256": "0e2deb3606b58e45cdbdae9077fb3a34f3c67fb09eda1d9589843c7e4cfd09da", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv", + "sha256": "841bd71e2768ff64157fb765691dd3d45546a0ed7bab5476da0910e314d5aea7", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv", + "sha256": "e6a124c41a4c31931de181a99d9b77b820dc00f8ca975e83e370fadbabad4ab5", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "sha256": "770df0d5e3d6dbdba0ccedb9317a807167a66b02a82005255b9c5351ca1e5f4c", + "role": "reproducible CPU-only enrichment renderer", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py" + } + ], + "deck_group": "Figure 2", + "layout": "full", + "addition_key": "enrichment:arch_panel_arm_register", + "parent_enriched_page": 20, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/arch_panel_arm_register.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 20, + "parent_154_page": 22, + "parent_155_page": 23 + }, + { + "key": "arch_all_arm_compute", + "kind": "content", + "type": "content", + "title": "All ablation arms: parameters and training time", + "section": "Figure 2", + "section_label": "Fig. 2g", + "figures": [ + "figures/final_palette_20261013/arch_all_arm_compute.png" + ], + "keywords": [], + "footnote": "Observed train-loop time includes loading; excludes validation/restarts/checkpoint writes. 16\u00d7 has E1\u2013E2 only; no controlled speed claim.", + "footnote_full": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. History train_seconds times train_one_epoch including data loading and heartbeat overhead. Divide its sum by recorded optimizer_steps for weighted seconds/update. Sum only completed parts; validation, checkpoint writes, resume gaps and failed partial parts excluded. RUN_STATUS.wall_seconds resets on resumed invocations and is not a total. Hosts, loader settings and competing load changed; do not interpret timing ratios as a controlled benchmark. Server_2 histories are existing local snapshots, not a live-status claim.", + "notes": "Source-only enrichment: saved configs/history + implementation; no new model training, inference or statistical scoring. Validation curves describe internal pretraining-corpus recipes (192 \u00d7 64), not SLE held-out donor scores. Figure 2 benchmark scores remain SLE pseudobulk, not measured bulk\u2013sc paired performance. Architecture and gene-panel families have different detection-floor gene selections and must be compared within family. History train_seconds times train_one_epoch including data loading and heartbeat overhead. Divide its sum by recorded optimizer_steps for weighted seconds/update. Sum only completed parts; validation, checkpoint writes, resume gaps and failed partial parts excluded. RUN_STATUS.wall_seconds resets on resumed invocations and is not a total. Hosts, loader settings and competing load changed; do not interpret timing ratios as a controlled benchmark. Server_2 histories are existing local snapshots, not a live-status claim.", + "after_current": 13, + "after_current_page": 13, + "palette": "native", + "role": "architecture_and_gene_cost", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "replace_current_page": null, + "replacement_previous_131_page": null, + "source_bindings": [ + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py", + "sha256": "c3d725b97d003b9b2fb45d5061d5f7daffe89fd37147f9d53753b3ed2da77ffa", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/model.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py", + "sha256": "5e4f6169410fa165e5b1c39391ab731c38c32b7d3202618da326b2f82131e6de", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py", + "sha256": "0d959a35175549a263e1b23c2709bfe2699dc293141d5699f01cf70cd7a0ba73", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/data.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py", + "sha256": "872bf059e14f0042eec64d290cb84685dafd9705ae364125462d1d8e06e14314", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/ladder_experiment.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py", + "sha256": "7a352d73728e78c9d262e29c35628d2437af0bc7743b88c077ba81d911dd2a52", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/src/pbisc_sdd/proportion_hint.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py", + "sha256": "09631a39fb388b56c4be88814559d82fac100660571a4b2a8e74e9972c400758", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/train_proportion_predictor.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md", + "sha256": "7fcae090a1e4417eb52c7a51556f0c82bacdc3bcf47193f41be2c30d350c357b", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md", + "sha256": "8c24781e40dd4284f06a07b9d7eac35dae2cfe013e1b46bd5a1805fd9afe83db", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/README.md" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py", + "sha256": "f023dc2fa788c8bac628fa922e5a34ccf05d355edec2a57807ca9bb24f3c241e", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_common.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py", + "sha256": "044cf1efdb17761914b1707861aa606095ca4a8ede109bb337782395cff01c47", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_panel.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py", + "sha256": "271834f39e5afb30c7c9e377b0b0f9aba06a85f8fe566a76ff263649f9a09ef0", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/sec03_width_bench.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py", + "sha256": "95ace096cdc58a9ca92a18af649c60fe6ac0a55e97f977e7ab9bcb04efa968cb", + "role": "implementation and historical protocol semantics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/scripts/sample_sdd_standalone.py" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json", + "sha256": "8377fa821211086a56b3450af3a464f55a5799c6df1b2486019dc96d46331587", + "role": "saved separate hint predictor metrics", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/prop_condition_6k_20260916/predictor_v1/predictor_report.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json", + "sha256": "af11b163a08666849621e7e66d89b2234314231bbecf976afa6f4b0110e83fa3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json", + "sha256": "24742955efffffdc425477dca23489862ed24b175479ce5f6f177b8b65586216", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json", + "sha256": "e5e2e77e662c42901d7e591506394b55781dcb5d352843e16e6625ecaebcc47e", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_bundle_v2/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json", + "sha256": "5e4c82793244d7b38961e000c035328a281aa4b283ac15eb066b665634131bd4", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json", + "sha256": "2fcf0161aebe500bcc4120d76db9e65431899c52d00e8db04372211236a1b8e3", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json", + "sha256": "57f8bbe6f9b6bfeabfb0ef7b4c87d415dbee6bd95e8746e2a810b81b13d391d1", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_archctx2048_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json", + "sha256": "b34dd4ac1b2b02cbeb2636327c352fd6a707a5bab284e9de3865e631c7665549", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json", + "sha256": "2f248ed36e4385ebd9fcb47bedba1dac7496ee282557bb1a55597fa507ed015b", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json", + "sha256": "d3c64355bc9d6a447db8cd9703dd32f3df69654a464f04d394dc26a104201aa6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_ctx512_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json", + "sha256": "da539ccf3266f283b7354eb13631a14e5efe3fcc4727fb5ec11b31a4a61ec798", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json", + "sha256": "ab817caa1f25cfcf5d4e6d69ff4ccb7d6dafcfd654427a8964c4a5f8f56fdce5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json", + "sha256": "4003cc7e63db9fdd56b483f7e37377216821895bf214f1bb57a47d2d0327e3f6", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch4x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json", + "sha256": "253b0145370ccdb6916f85d5288ad3c192ab63e78dd3be9f6925b1ab65b2d2e0", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json", + "sha256": "5a4e3efedb25e25731d0122b437eabdb6c8078ee4facd46c3a9f551a7f99732a", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json", + "sha256": "2ac8b1a1f00e76241ff9623f76405374a30f3a0804593040576b705a2aa97f56", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_arch16x_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json", + "sha256": "ff2f34ac82e86d6901244f87962244340bc6e40b177e6d2b010da03829569426", + "role": "16x declared parameter count", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/arch_ablation_6k_20260913/MANIFEST.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json", + "sha256": "0b676b2bd9baea8e9276279efe925d836c593607a3a46dd5196ca7eac7670bb3", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json", + "sha256": "421e3232a98c6a15ce4781567f91cf94e6278d930a5a590ca94a9b38573247cc", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json", + "sha256": "51f05a0525cf71010d9f6364621ca518a6db583e45ff4b6639140b3fa39dfdd0", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/s0_6k_allK_topup16_seed20260627_prop_v1/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json", + "sha256": "4126f2d70c0fd8300cd1083c4b1e3b0736fcb6cd71bb4b5c07f076174b440ae7", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json", + "sha256": "42b3f10470db8faafb7390f2fb1bf76ce3e33feb65576a6acf336985e7e7c7c8", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json", + "sha256": "b4e7f8d2a8ad984cbfe16b12709d49465824b8120d23c30e27e96b3f47fddf3b", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_2k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/effective_config.json", + "sha256": "a657f884b950151baeea56ec3eade6615160da829aa41bc1a9045576613850cf", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/RUN_STATUS.json", + "sha256": "45e228d2e95b65a59c6e7a75c15c110e9aa9aaba8b2ca0273872189c29f790f0", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_3k_1x_seed20260627/history.json", + "sha256": "89d9bd8b9bab76d5c59f16a44b81e772ae3342c526b6b51eafe0a90d6ecc03f7", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_3k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json", + "sha256": "0a8e9a0ce0b82d136d3e439cbca6d6d61ed7f6b8a8f935cedd69d8c65fb15abd", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json", + "sha256": "d70a2c283d7123428bc65cffa633e2d38e37dd70881a25b0377dffded191e9d5", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json", + "sha256": "2296899ba51f26a553205ef325d726e5ca7608cd21815b031d97675208273a97", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/effective_config.json", + "sha256": "3e020261d15f330bdf312b3a7737b9d73412fbb60da240497af89c80c35799f9", + "role": "saved training configuration", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/effective_config.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/RUN_STATUS.json", + "sha256": "b3986d40398420f2f5d3d77cbe4101f5e0e7acf85fcd8f5815f0ff4cd9a55d66", + "role": "saved run completion/failure metadata", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/RUN_STATUS.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_server2_snapshots/f01_p15k_1x_seed20260627/history.json", + "sha256": "6ed461a488da9c64e58367f82e329c8544869f48e0cbe158bc3422d9f661576f", + "role": "complete saved epoch history", + "host": "server_1", + "original_host": "server_2", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_p15k_1x_seed20260627/history.json" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv", + "sha256": "7c8805749d4484d1eea3503a8410c09572d799b219401f48b17c9bf56533e444", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv", + "sha256": "4c9af91a9497c5d2335c1ca79b90e217b90fb0ded9bc330d0f4970554ed5d7ec", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_loss.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv", + "sha256": "0e2deb3606b58e45cdbdae9077fb3a34f3c67fb09eda1d9589843c7e4cfd09da", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_width_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv", + "sha256": "841bd71e2768ff64157fb765691dd3d45546a0ed7bab5476da0910e314d5aea7", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_bench.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv", + "sha256": "e6a124c41a4c31931de181a99d9b77b820dc00f8ca975e83e370fadbabad4ab5", + "role": "existing figure source table; no rescoring", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec03_panel_seeds.csv" + }, + { + "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py", + "sha256": "770df0d5e3d6dbdba0ccedb9317a807167a66b02a82005255b9c5351ca1e5f4c", + "role": "reproducible CPU-only enrichment renderer", + "host": "server_1", + "original_host": "server_1", + "original_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_architecture_20261013.py" + } + ], + "deck_group": "Figure 2", + "layout": "full", + "addition_key": "enrichment:arch_all_arm_compute", + "parent_enriched_page": 21, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/arch_all_arm_compute.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 21, + "parent_154_page": 23, + "parent_155_page": 24 + }, + { + "title": "Real-cell repeat: a sampling reference", + "figures": [ + "figures/figure_enrichment_20261013/labels_real_cell_repeat.png" + ], + "footnote": "Compare disjoint real-cell samples from the same donors. This measures sampling variability, not a theoretical ceiling.", + "footnote_full": "Previously called Real-cell repeat. The reference uses one 100-cell draw per donor; the control uses a disjoint 100-cell draw from the same 42 donors. Biological n remains 42.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/labels_real_cell_repeat.json", + "kind": "content", + "deck_group": "Figure 2", + "section": "Figure 2", + "section_label": "Fig. 2h", + "layout": "full", + "keywords": [], + "addition_key": "enrichment:real_cell_repeat", + "parent_enriched_page": 22, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/labels_real_cell_repeat.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 22, + "parent_154_page": 24, + "parent_155_page": 25 + }, + { + "kind": "content", + "title": "Native-pseudobulk model comparisons", + "section_label": "Fig. 2i", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/labels_ablation.png" + ], + "keywords": [], + "footnote": "SLE pseudobulk; common 1,006-gene evaluation. Real-cell repeat compares disjoint observed-cell draws from the same donors.", + "footnote_full": "Additional view from the audited results review. Completed held-out rescoring. Architecture: 1% detection floor; gene-panel comparison: 0.1% floor.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/labels_ablation.json", + "addition_key": "review:6", + "after_original": 12, + "figures_before_palette": [ + "figures/review_style_20261013_ablation.png" + ], + "source_page_131": 15, + "deck_group": "Figure 2", + "section": "Figure 2", + "previous_section_label": "Fig. 2a", + "parent_page": 14, + "previous_figures": [ + "figures/expanded_master_20261013/palette/review_style_20261013_ablation.png" + ], + "parent_enriched_page": 23, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/labels_ablation.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 23, + "parent_154_page": 25, + "parent_155_page": 26 + }, + { + "kind": "divider", + "title": "Figure 3. Real-bulk adaptation", + "original_slide": 28, + "original_title": "Figure 3. PBISC vs scSemiProfiler on healthy donors", + "original_footnote": "", + "source_page_131": 52, + "deck_group": "Figure 3", + "section": "Figure 3", + "previous_section_label": null, + "section_label": "Figure 3", + "parent_page": 15, + "parent_enriched_page": 24, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 24, + "parent_154_page": 26, + "parent_155_page": 27 + }, + { + "kind": "plan_card", + "section": "Figure 3", + "section_label": "Figure 3 plan", + "title": "Figure 3: the paired-adaptation story", + "footnote": "All performance panels use healthy measured pairs. Pair-budget curves use n training + 7 validation; genuine total-five results are separate.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "All performance panels use healthy measured pairs. Pair-budget curves use n training + 7 validation; genuine total-five results are separate.", + "original_slide": 29, + "original_title": "Figure 3 plan: PBISC vs scSemiProfiler on healthy donors", + "original_footnote": "On healthy donors, LoRA cells are closer to the real ones than scSemiProfiler's, in minutes instead of days. 4 of 4 panels ready.", + "source_page_131": 53, + "deck_group": "Figure 3", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_fig3.png" + ], + "rows": [ + [ + "3a", + "Expression mismatch; same-donor output recovery", + "pp. 30, 31, 33, 34" + ], + [ + "3b", + "Full FT and LoRA ranks", + "pp. 35, 36" + ], + [ + "3c", + "Fixed test, random nested support and all six types", + "pp. 37, 38, 39" + ], + [ + "3d", + "Genuine total-five comparison", + "pp. 40, 41" + ] + ], + "parent_page": 16, + "parent_enriched_page": 25, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 25, + "parent_154_page": 27, + "parent_155_page": 28 + }, + { + "kind": "content", + "section": "Figure 3", + "section_label": "Fig. 3a", + "title": "Bulk and pooled cells: different expression values", + "figures": [ + "figures/figure_enrichment_20261013/mismatch_expression_values.png" + ], + "footnote": "107 paired donors; all 5,508 shared final-model genes. log1p CPM on the same gene axis. Saved calibration uses 10 other-pool pairs.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/MANIFEST_ImMicro.json; /home/sj_server_1/gblee/datalake/paired/analyses/offset_sample_size_20260914/features.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/reduced/reduced_Arunachalam_paired12.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/reduced/reduced_ImMicro_native.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/reduced/reduced_Arunachalam_paired12.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/conditions_ImMicro_raw.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/conditions_ImMicro_cal.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/conditions_ImMicro_pseudo.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/ImMicro/metrics_ImMicro.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/ImMicro/per_sample_distance_ImMicro_native.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_expression_mismatch_20261013.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/style.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/mismatch_gene_axis.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/mismatch_per_donor_errors.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/mismatch_density_bins.csv", + "footnote_full": "ImMicro, 107 unique paired healthy donors. All 5508 final-model panel genes measured by both assays in every donor are included; the other 492 model genes are excluded by availability, never filled as measured zeros. Each assay row is separately normalized to 1e6 over the same 5508 genes and transformed by natural log(1+CPM). Axes are gene-expression profiles, not cell-type counts, learned embeddings or PCA. Observed pseudobulk is the sum of the donor's measured single-cell UMI counts. The saved calibrated condition uses the historical gene-wise multiplier from exactly 10 other-pool paired donors, with the evaluated donor's whole sc pool excluded. It is a paired-information diagnostic, not zero-shot deployment, LoRA output or the current strict-five budget. Source condition hashes and raw/pseudobulk normalization are verified. The full-expression distances reproduce the old saved analysis within its rounded CSV precision. Assay differences can include cell capture, pool/batch, depth and assay chemistry; this plot does not attribute them to one mechanism or treat scRNA-seq as noise-free truth. No individual genes or donors were chosen for a favorable example. The plots describe 107 biological donors; donor\u2013gene points are not independent replicates.", + "original_slide": 17, + "original_title": "Real bulk \u2260 pseudobulk (model input)", + "original_footnote": "ImMicro, 107 healthy donors. Model input space (log1p CPM). D = median donor-to-donor distance.", + "source_page_131": 21, + "deck_group": "Figure 3", + "previous_section_label": "Fig. 2a", + "parent_page": 17, + "layout": "full", + "key": "mismatch_expression_values", + "role": "f3_expression_mismatch", + "keywords": [], + "notes": "ImMicro, 107 unique paired healthy donors. All 5508 final-model panel genes measured by both assays in every donor are included; the other 492 model genes are excluded by availability, never filled as measured zeros. Each assay row is separately normalized to 1e6 over the same 5508 genes and transformed by natural log(1+CPM). Axes are gene-expression profiles, not cell-type counts, learned embeddings or PCA. Observed pseudobulk is the sum of the donor's measured single-cell UMI counts. The saved calibrated condition uses the historical gene-wise multiplier from exactly 10 other-pool paired donors, with the evaluated donor's whole sc pool excluded. It is a paired-information diagnostic, not zero-shot deployment, LoRA output or the current strict-five budget. Source condition hashes and raw/pseudobulk normalization are verified. The full-expression distances reproduce the old saved analysis within its rounded CSV precision. Assay differences can include cell capture, pool/batch, depth and assay chemistry; this plot does not attribute them to one mechanism or treat scRNA-seq as noise-free truth. No individual genes or donors were chosen for a favorable example. The plots describe 107 biological donors; donor\u2013gene points are not independent replicates.", + "palette": "biological", + "after_original": 17, + "previous_figures": [ + "figures/sec05_condition_space.png" + ], + "parent_enriched_page": 26, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/mismatch_expression_values.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 26, + "parent_154_page": 28, + "parent_155_page": 29 + }, + { + "kind": "content", + "layout": "full", + "key": "mismatch_same_donor_errors", + "role": "f3_expression_discrepancy", + "title": "Paired calibration reduces same-donor expression error", + "section_label": "Fig. 3a detail", + "figures": [ + "figures/final_palette_20261013/mismatch_same_donor_errors.png" + ], + "keywords": [], + "footnote": "Each line = one donor. Error averages all 5,508 genes; no nearest-neighbor score. Ten-pair, other-pool calibration is diagnostic.", + "footnote_full": "ImMicro, 107 unique paired healthy donors. All 5508 final-model panel genes measured by both assays in every donor are included; the other 492 model genes are excluded by availability, never filled as measured zeros. Each assay row is separately normalized to 1e6 over the same 5508 genes and transformed by natural log(1+CPM). Axes are gene-expression profiles, not cell-type counts, learned embeddings or PCA. Observed pseudobulk is the sum of the donor's measured single-cell UMI counts. The saved calibrated condition uses the historical gene-wise multiplier from exactly 10 other-pool paired donors, with the evaluated donor's whole sc pool excluded. It is a paired-information diagnostic, not zero-shot deployment, LoRA output or the current strict-five budget. Source condition hashes and raw/pseudobulk normalization are verified. The full-expression distances reproduce the old saved analysis within its rounded CSV precision. Assay differences can include cell capture, pool/batch, depth and assay chemistry; this plot does not attribute them to one mechanism or treat scRNA-seq as noise-free truth. No individual genes or donors were chosen for a favorable example. The plots describe 107 biological donors; donor\u2013gene points are not independent replicates.", + "notes": "ImMicro, 107 unique paired healthy donors. All 5508 final-model panel genes measured by both assays in every donor are included; the other 492 model genes are excluded by availability, never filled as measured zeros. Each assay row is separately normalized to 1e6 over the same 5508 genes and transformed by natural log(1+CPM). Axes are gene-expression profiles, not cell-type counts, learned embeddings or PCA. Observed pseudobulk is the sum of the donor's measured single-cell UMI counts. The saved calibrated condition uses the historical gene-wise multiplier from exactly 10 other-pool paired donors, with the evaluated donor's whole sc pool excluded. It is a paired-information diagnostic, not zero-shot deployment, LoRA output or the current strict-five budget. Source condition hashes and raw/pseudobulk normalization are verified. The full-expression distances reproduce the old saved analysis within its rounded CSV precision. Assay differences can include cell capture, pool/batch, depth and assay chemistry; this plot does not attribute them to one mechanism or treat scRNA-seq as noise-free truth. No individual genes or donors were chosen for a favorable example. The plots describe 107 biological donors; donor\u2013gene points are not independent replicates.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/MANIFEST_ImMicro.json; /home/sj_server_1/gblee/datalake/paired/analyses/offset_sample_size_20260914/features.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/reduced/reduced_Arunachalam_paired12.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/reduced/reduced_ImMicro_native.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/reduced/reduced_Arunachalam_paired12.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/conditions_ImMicro_raw.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/conditions_ImMicro_cal.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/realbulk_generation_20260922/conditions/conditions_ImMicro_pseudo.npz; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/ImMicro/metrics_ImMicro.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/sample_level/ImMicro/per_sample_distance_ImMicro_native.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/enrich_expression_mismatch_20261013.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/style.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/mismatch_gene_axis.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/mismatch_per_donor_errors.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/mismatch_density_bins.csv", + "palette": "biological", + "after_original": 17, + "deck_group": "Figure 3", + "section": "Figure 3", + "addition_key": "enrichment:mismatch_same_donor_errors", + "parent_enriched_page": 27, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/mismatch_same_donor_errors.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 27, + "parent_154_page": 29, + "parent_155_page": 30 + }, + { + "kind": "content", + "section": "Figure 3", + "section_label": "Fig. 3b detail", + "title": "Consequence: cells off-manifold", + "layout": "two_figures", + "figures": [ + "figures/sec05_cells_umap.png", + "figures/final_palette_20261013/sec05_cells_metrics.png" + ], + "keywords": [ + "Monocytes: raw 30.6% vs real 16.7%", + "Same noise: 19% of T cells become monocytes", + "Precision: raw 0.36 vs pseudobulk 0.92" + ], + "footnote": "Existing same-noise input diagnostic; not the new LoRA/scSemi OOF embedding. UMAP overlap is not a quantitative score.", + "footnote_full": "Current scope: Existing same-noise input diagnostic; not the new LoRA/scSemi OOF embedding. UMAP overlap is not a quantitative score.\n\nOriginal slide 19: ImMicro, 107 donors. Final model f01 6k 1\u00d7 E16 (DDIM 50, EMA); each cell starts from the same noise under every input. Precision/coverage from one scoring call. Composition L1: raw 0.455, pseudobulk 0.166, cohort-mean null 0.161. The UMAP is fitted on real cells and overstates overlap.", + "sources": "tools/sec05_cells_offmanifold.py; data/sec05_cells_metrics.csv, data/sec05_cells_umap_coords.csv", + "refs": [ + "Kynk\u00e4\u00e4nniemi 2019, NeurIPS", + "Naeem 2020, ICML" + ], + "refs_full": [ + "Kynk\u00e4\u00e4nniemi et al. Improved precision and recall metric for assessing generative models. NeurIPS 2019", + "Naeem et al. Reliable fidelity and diversity metrics for generative models. ICML 2020" + ], + "original_slide": 19, + "original_title": "Consequence: cells off-manifold", + "original_footnote": "ImMicro, 107 donors; same starting noise for every input. Coverage: raw 0.08, sc pseudobulk 0.51.", + "figures_before_palette": [ + "figures/sec05_cells_umap.png", + "figures/sec05_cells_metrics.png" + ], + "source_page_131": 23, + "deck_group": "Figure 3", + "previous_section_label": "Fig. 2b", + "parent_page": 18, + "parent_enriched_page": 28, + "figures_before_final_palette": [ + "figures/sec05_cells_umap.png", + "figures/expanded_master_20261013/palette/sec05_cells_metrics.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 28, + "parent_154_page": 30, + "parent_155_page": 31 + }, + { + "kind": "content", + "title": "Adaptation recovers the observed cell mixture", + "figures": [ + "figures/strengthening_20261013/healthy/same_donors_composition.png" + ], + "layout": "full", + "keywords": [], + "section": "Figure 3", + "deck_group": "Figure 3", + "section_label": "Fig. 3 \u00b7 Adaptation", + "footnote": "Same 107 held-out donors in every row. Full FT / LoRA use raw measured bulk; colours identify cell types.", + "footnote_full": "Same 107 ImMicro healthy donors, pooled five-fold held-out predictions and six frozen broad cell types. Every donor and type retained; display order uses observed monocyte fraction only. Each generated arm has 1500 cells/sample; observed fractions use all available scored observed cells. Full FT and LoRA consume the raw measured-bulk input, as does raw zero-shot. Input calibration is a separate comparator. Adaptation improves generated-cell composition; it does not change the measured bulk or establish that every transcriptional mismatch disappears. Historical training has 78\u201379 train plus 7 inner-validation pairs/fold. No new training, inference, annotation, or scoring; this reuses saved donor metrics.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/pooled/per_donor.csv", + "addition_key": "strengthening:healthy:0", + "parent_154_page": 31, + "parent_155_page": 32 + }, + { + "kind": "content", + "title": "Adaptation: mixture error and expression agreement", + "figures": [ + "figures/strengthening_20261013/healthy/same_donors_output_recovery.png" + ], + "layout": "full", + "keywords": [], + "section": "Figure 3", + "deck_group": "Figure 3", + "section_label": "Fig. 3 \u00b7 Adaptation", + "footnote": "107 matched donors; lines connect the same person. Gene-expression r is a donor pseudobulk Pearson correlation.", + "footnote_full": "Same 107 ImMicro healthy donors, pooled five-fold held-out predictions and six frozen broad cell types. Every donor and type retained; display order uses observed monocyte fraction only. Each generated arm has 1500 cells/sample; observed fractions use all available scored observed cells. Full FT and LoRA consume the raw measured-bulk input, as does raw zero-shot. Input calibration is a separate comparator. Adaptation improves generated-cell composition; it does not change the measured bulk or establish that every transcriptional mismatch disappears. Historical training has 78\u201379 train plus 7 inner-validation pairs/fold. No new training, inference, annotation, or scoring; this reuses saved donor metrics. Error = sum across six types of absolute fraction differences. Expression r is saved pb_r, the per-donor Pearson correlation of log1p-CPM(summed generated expression rates) versus log1p-CPM(observed pseudobulk) on the same 5,508 scorer genes; not cell-wise correspondence. Lines are not diffusion trajectories. No significance tests.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/pooled/per_donor.csv", + "addition_key": "strengthening:healthy:1", + "parent_154_page": 32, + "parent_155_page": 33 + }, + { + "kind": "content", + "title": "Healthy paired data: full FT and LoRA", + "section_label": "Fig. 3b", + "after_original": 3, + "addition_key": "guide:availability", + "footnote": "ImMicro measured bulk-sc pairs; pooled five-fold results. All four methods use the same saved scoring table.", + "footnote_full": "Healthy-only extraction: dataset=immicro, call=pooled; raw, calibrated, full FT and LoRA r16. No disease or fold0 condition-encoder pilot rows. No new statistics.", + "source_page_131": 5, + "deck_group": "Figure 3", + "section": "Figure 3", + "previous_section_label": "Evidence map", + "figures": [ + "figures/final_palette_20261013/native/healthy_ft_comparison.png" + ], + "layout": "full", + "keywords": [], + "sources": "data/sec07_ft_modes_accuracy.csv; data/figure_order_fixed_20261013/healthy_ft_comparison.json", + "parent_page": 19, + "parent_enriched_page": 29, + "figures_before_final_palette": [ + "figures/figure_order_fixed_20261013/healthy_ft_comparison.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 29, + "parent_154_page": 33, + "parent_155_page": 34 + }, + { + "kind": "content", + "title": "Full FT and LoRA ranks: quality and parameters", + "section_label": "Fig. 3c", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/review_style_20261013_rank.png" + ], + "keywords": [], + "footnote": "Original donor-level five-fold comparison. LoRA rank 16 is a quality/parameter compromise.", + "footnote_full": "Additional view from the audited results review. Original donor-level five-fold comparison. LoRA rank 16 is a quality/parameter compromise.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/review_style_20261013_rank.pdf", + "addition_key": "review:8", + "after_original": 27, + "figures_before_palette": [ + "figures/review_style_20261013_rank.png" + ], + "source_page_131": 51, + "deck_group": "Figure 3", + "section": "Figure 3", + "previous_section_label": "Fig. 2c", + "parent_page": 20, + "parent_enriched_page": 30, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/review_style_20261013_rank.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 30, + "parent_154_page": 34, + "parent_155_page": 35 + }, + { + "kind": "content", + "title": "What does \u201call pairs\u201d actually mean?", + "figures": [ + "figures/strengthening_20261013/healthy/fixed_test_split.png" + ], + "layout": "full", + "keywords": [], + "section": "Figure 3", + "deck_group": "Figure 3", + "section_label": "Fig. 3 \u00b7 Adaptation", + "footnote": "Random, nested training subsets; the test fold stays fixed. \u201cAll\u201d = all available training pairs in that fold.", + "footnote_full": "Historical ImMicro CV, not strict total-support experiment. Outer group-stratified donor folds and inner validation use seed 20260927. Saved training subsets use fold-specific seeds 1000, 1001, 1002, 1003, 1004. Within each fold one random permutation yields nested 5/10/20/40 training subsets; exact replay matches all 20 saved subset memberships. Every budget has identical held-out test donors (22,22,21,21,21), and seven validation donors. All uses78,78,79,79,79 training donors. Each of107 donors contributes one held-out prediction to the pooled metric. With n20,58\u201359 remaining training-pool donors are unused, not moved into the test fold. The separate legacy healthy five-draw comparison uses 5 support+0 validation,102 test/draw. The newer strict healthy support campaign instead keeps outer-fold held-out tests and has no extra validation; do not conflate these two total-five designs. No training/test overlap or membership changes were introduced.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/pooled/per_donor.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/immicro/folds.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/immicro/PREPARE.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/prepare_dataset.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/ft_sdd.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/score_ft.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/pooled/summary.json", + "addition_key": "strengthening:healthy:2", + "parent_154_page": 35, + "parent_155_page": 36 + }, + { + "kind": "content", + "key": "pair_budget_fraction_correlation", + "role": "f3_pair_budget_fraction_correlation", + "title": "Healthy pair budget: correlation in all cell types", + "section_label": "Fig. 3d", + "layout": "full", + "keywords": [], + "figures": [ + "figures/pair_budget_celltypes_20261013/fraction_correlation.png" + ], + "palette": "native", + "after_original": 31, + "replaces_figure": "fig2_healthy_pairs", + "footnote": "107 donors, pooled 5-fold CV. Training n + 7 validation pairs; calibration uses the same n + 7. All = 78\u201379 training pairs.", + "footnote_full": "Historical ImMicro healthy five-fold CV: 107 unique donors, one held-out prediction per donor per arm. All six saved cell types are included, with the identical observed fractions and scoring reference across arms. The x-axis is training pairs: 0=raw zero-shot with no pairs; 5/10/20/40/all LoRA training pairs plus 7 inner-validation pairs per fold for early stopping. All means 78\u201379 training pairs. Calibration fits alpha on exactly the same training plus validation pairs, including at all. This is not the strict five-total-pair experiment. Correlation is Pearson between generated and observed donor fractions over pooled 107 donors, not an average of fold correlations. MAE is mean(abs(generated_fraction-observed_fraction))*100 in percentage points, with equal donor weights. No statistical significance or universal improvement is claimed. Missing correlation remains NA; no undefined value is replaced by zero. No donor/type/arm selection by outcome, no new generation, and no scSemiProfiler arm added to the original fig2_healthy_pairs comparison. MAE panels have explicitly labelled per-celltype axis ranges; correlation axes are shared.", + "notes": "Historical ImMicro healthy five-fold CV: 107 unique donors, one held-out prediction per donor per arm. All six saved cell types are included, with the identical observed fractions and scoring reference across arms. The x-axis is training pairs: 0=raw zero-shot with no pairs; 5/10/20/40/all LoRA training pairs plus 7 inner-validation pairs per fold for early stopping. All means 78\u201379 training pairs. Calibration fits alpha on exactly the same training plus validation pairs, including at all. This is not the strict five-total-pair experiment. Correlation is Pearson between generated and observed donor fractions over pooled 107 donors, not an average of fold correlations. MAE is mean(abs(generated_fraction-observed_fraction))*100 in percentage points, with equal donor weights. No statistical significance or universal improvement is claimed. Missing correlation remains NA; no undefined value is replaced by zero. No donor/type/arm selection by outcome, no new generation, and no scSemiProfiler arm added to the original fig2_healthy_pairs comparison. MAE panels have explicitly labelled per-celltype axis ranges; correlation axes are shared.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/pooled/per_donor.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/cal_budget_20261001/pooled/per_donor.csv", + "deck_group": "Figure 3", + "section": "Figure 3", + "addition_key": "all_celltypes:correlation", + "parent_142_page": 31, + "parent_154_page": 36, + "parent_155_page": 37 + }, + { + "kind": "content", + "key": "pair_budget_fraction_mae", + "role": "f3_pair_budget_fraction_mae", + "title": "Healthy pair budget: error in all cell types", + "section_label": "Fig. 3d detail", + "layout": "full", + "keywords": [], + "figures": [ + "figures/pair_budget_celltypes_20261013/fraction_mae.png" + ], + "palette": "native", + "after_original": 31, + "replaces_figure": "fig2_healthy_pairs", + "footnote": "107 donors, pooled 5-fold CV. Training n + 7 validation pairs; calibration uses the same n + 7. All = 78\u201379 training pairs.", + "footnote_full": "Historical ImMicro healthy five-fold CV: 107 unique donors, one held-out prediction per donor per arm. All six saved cell types are included, with the identical observed fractions and scoring reference across arms. The x-axis is training pairs: 0=raw zero-shot with no pairs; 5/10/20/40/all LoRA training pairs plus 7 inner-validation pairs per fold for early stopping. All means 78\u201379 training pairs. Calibration fits alpha on exactly the same training plus validation pairs, including at all. This is not the strict five-total-pair experiment. Correlation is Pearson between generated and observed donor fractions over pooled 107 donors, not an average of fold correlations. MAE is mean(abs(generated_fraction-observed_fraction))*100 in percentage points, with equal donor weights. No statistical significance or universal improvement is claimed. Missing correlation remains NA; no undefined value is replaced by zero. No donor/type/arm selection by outcome, no new generation, and no scSemiProfiler arm added to the original fig2_healthy_pairs comparison. MAE panels have explicitly labelled per-celltype axis ranges; correlation axes are shared.", + "notes": "Historical ImMicro healthy five-fold CV: 107 unique donors, one held-out prediction per donor per arm. All six saved cell types are included, with the identical observed fractions and scoring reference across arms. The x-axis is training pairs: 0=raw zero-shot with no pairs; 5/10/20/40/all LoRA training pairs plus 7 inner-validation pairs per fold for early stopping. All means 78\u201379 training pairs. Calibration fits alpha on exactly the same training plus validation pairs, including at all. This is not the strict five-total-pair experiment. Correlation is Pearson between generated and observed donor fractions over pooled 107 donors, not an average of fold correlations. MAE is mean(abs(generated_fraction-observed_fraction))*100 in percentage points, with equal donor weights. No statistical significance or universal improvement is claimed. Missing correlation remains NA; no undefined value is replaced by zero. No donor/type/arm selection by outcome, no new generation, and no scSemiProfiler arm added to the original fig2_healthy_pairs comparison. MAE panels have explicitly labelled per-celltype axis ranges; correlation axes are shared.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/pooled/per_donor.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/cal_budget_20261001/pooled/per_donor.csv", + "deck_group": "Figure 3", + "section": "Figure 3", + "addition_key": "all_celltypes:mae", + "parent_142_page": 32, + "parent_154_page": 37, + "parent_155_page": 38 + }, + { + "kind": "content", + "section": "Figure 3", + "section_label": "Fig. 3e", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 5 + }, + "title": "Five pairs, no extra validation: LoRA vs scSemi", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig3_healthy_fewpair.png" + ], + "keywords": [ + "LoRA-5 ahead on 6 of 6 metrics", + "Coverage 0.79 vs 0.04", + "E_N 1.5 vs 3.6 (scSemiProfiler's own metric)" + ], + "footnote": "ImMicro: exactly 5 pairs, no extra validation, 3000-step LoRA; five draws, remaining 102 donors per draw.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs": [ + "Wang 2024, Nat Commun", + "Lopez-Paz 2017, ICLR" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1", + "Lopez-Paz & Oquab. Revisiting classifier two-sample tests. ICLR 2017" + ], + "say": "With the same five healthy donors, LoRA-5 beats scSemiProfiler on all six cell-level metrics, including scSemiProfiler's own error metric.", + "footnote_full": "Current scope: ImMicro: exactly 5 pairs, no extra validation, 3000-step LoRA; five draws, remaining 102 donors per draw.\n\nOriginal slide 30: Main comparison on healthy donors: per draw, LoRA-5 and scSemiProfiler get the same 5 paired donors (scSemiProfiler: as its representatives) and are tested on the other 102 from their real bulk; 5 draws. Zero-shot (no pairs) shown for reference. Dashed: real vs real cells. Values (LoRA-5 / scSemiProfiler-5): coverage 0.792 / 0.036; precision 0.952 / 0.518; comp_l1 0.176 / 0.233; c2st_knn_auc 0.943 / 0.999; mmd2_bw1 0.023 / 0.186; en_all_sq 1.547 / 3.649. Source: data/sec09_fewpair_cells.csv, data/sec09_extra_fewpair.csv.", + "numbers": [ + { + "value": "5 donors (mean of 5 draws) coverage zs_raw 0.19; 5 donors (mean of 5 draws) coverage scsemi 0.03583; 5 donors (mean of 5 draws) coverage lora 0.7925; 5 donors (mean of 5 draws) precision zs_raw 0.6772; 5 donors (mean of 5 draws) precision scsemi 0.518; 5 donors (mean of 5 draws) precision lora 0.9521; 5 donors (mean of 5 draws) comp_l1 zs_raw 0.4451; 5 donors (mean of 5 draws) comp_l1 scsemi 0.2325; 5 donors (mean of 5 draws) comp_l1 lora 0.1756; 5 donors (mean of 5 draws) c2st_knn_auc zs_raw 0.9966; 5 donors (mean of 5 draws) c2st_knn_auc scsemi 0.9988; 5 donors (mean of 5 draws) c2st_knn_auc lora 0.9427; 5 donors (mean of 5 draws) mmd2_bw1 zs_raw 0.1026; 5 donors (mean of 5 draws) mmd2_bw1 scsemi 0.1865; 5 donors (mean of 5 draws) mmd2_bw1 lora 0.02308; 5 donors (mean of 5 draws) en_all_sq zs_raw 7.236; 5 donors (mean of 5 draws) en_all_sq scsemi 3.649; 5 donors (mean of 5 draws) en_all_sq lora 1.547", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 30, + "original_title": "Healthy donors, 5 pairs: LoRA-5 vs scSemiProfiler-5", + "original_footnote": "ImMicro, 107 healthy donors; same 5 donors for both methods, 5 draws (dots).", + "figures_before_palette": [ + "figures/fig3_healthy_fewpair.png" + ], + "source_page_131": 54, + "deck_group": "Figure 3", + "previous_section_label": "Fig. 3a", + "parent_page": 22, + "parent_enriched_page": 32, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig3_healthy_fewpair.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 33, + "parent_154_page": 38, + "parent_155_page": 39 + }, + { + "kind": "content", + "title": "Composition: error and donor-level wins", + "section_label": "Fig. 3f", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/review_style_20261013_healthy_composition.png" + ], + "keywords": [], + "footnote": "Exactly five pairs per draw. Absolute errors averaged within each held-out donor; 107 donors, equal weight. Descriptive.", + "footnote_full": "Additional view from the audited results review. Exactly five pairs per draw. Absolute errors averaged within each held-out donor; 107 donors, equal weight. Descriptive.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/review_style_20261013_healthy_composition.pdf", + "addition_key": "review:13", + "after_original": 30, + "figures_before_palette": [ + "figures/review_style_20261013_healthy_composition.png" + ], + "source_page_131": 55, + "deck_group": "Figure 3", + "section": "Figure 3", + "previous_section_label": "Fig. 3c", + "parent_page": 23, + "parent_enriched_page": 33, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/review_style_20261013_healthy_composition.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 34, + "parent_154_page": 39, + "parent_155_page": 40 + }, + { + "kind": "divider", + "title": "Figure 4. Disease signal preservation", + "original_slide": 35, + "original_title": "Figure 4. Disease signal on paired cohorts", + "original_footnote": "", + "source_page_131": 60, + "deck_group": "Figure 4", + "section": "Figure 4", + "previous_section_label": null, + "section_label": "Figure 4", + "parent_page": 24, + "parent_enriched_page": 34, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 35, + "parent_154_page": 40, + "parent_155_page": 41 + }, + { + "kind": "plan_card", + "section": "Figure 4", + "section_label": "Figure 4 plan", + "title": "Figure 4: disease signal preservation", + "footnote": "Paired disease benchmarks: pooled historical CV and strict five-support results remain separate; observed controls are cell resampling, not donor ceilings.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "Paired disease benchmarks: pooled historical CV and strict five-support results remain separate; observed controls are cell resampling, not donor ceilings.", + "original_slide": 36, + "original_title": "Figure 4 plan: Disease signal on paired cohorts", + "original_footnote": "LoRA cells recover the paired single cells' disease signal; compared with other PBISC versions, scSemiProfiler and deconvolution. 6 of 6 panels ready.", + "source_page_131": 61, + "deck_group": "Figure 4", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_fig4.png" + ], + "rows": [ + [ + "4a", + "Disease contrasts: five-support and five-fold", + "pp. 44, 45" + ], + [ + "4b", + "Complete CV-CAT, including scSemiProfiler", + "p. 48" + ], + [ + "4c", + "LogFC / partial correlation / RBO", + "p. 49" + ], + [ + "4d", + "Strict Sepsis: five support draws and donor-aware DEG", + "p. 50" + ], + [ + "4e", + "Deconvolution comparator endpoints", + "p. 47" + ] + ], + "parent_page": 25, + "parent_enriched_page": 35, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 36, + "parent_154_page": 41, + "parent_155_page": 42 + }, + { + "kind": "content", + "section": "Figure 4", + "section_label": "Fig. 4a", + "text_source": { + "file": "data/fig4_disease_slide_text.json", + "index": 0 + }, + "title": "Disease signals with five training donors", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig4_disease_fewpair.png" + ], + "keywords": [ + "vs scSemiProfiler-5: logFC r 13/20, partial r 14/20 draws", + "Beyond bulk: partial r 0.60 vs 0.43 (Imperial)", + "Above zero-shot and calibration in 4/4 cohorts" + ], + "footnote": "Existing five-donor protocol; sample counts differ by cohort. Original sample-level estimator; corrected update separate.", + "sources": "tools/fig4_disease.py; data/fig4_disease.csv", + "refs": [ + "Wang 2024, Nat Commun", + "Vi\u00f1as Torn\u00e9 2026, Nat Biotechnol" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1", + "Vi\u00f1as Torn\u00e9 et al. Systema: a framework for evaluating genetic perturbation response prediction beyond systematic variation. Nat Biotechnol 44:1050-1059 (2026). doi:10.1038/s41587-025-02777-8" + ], + "say": "The main comparison uses five pairs, the situation a lab is in. With the same five donors, LoRA-5 is above both zero-shot versions in every cohort and ahead of scSemiProfiler in most draws. The bold panel is partial r: the cell-type signal the bulk itself does not carry, which is what single cells add.", + "footnote_full": "Current scope: Existing five-donor protocol; sample counts differ by cohort. Original sample-level estimator; corrected update separate.\n\nOriginal slide 38: Main comparison: 5 donors per draw, LoRA\u20135 vs scSemiProfiler with the same 5 donors as representatives (and the two zero\u2013shot PBISC versions), 5 draws per cohort, every other usable sample tested. logFC r = Pearson r of case\u2013vs\u2013control logFC (mean CP10K) with the observed cells, mean over T, B, monocyte and NK. Partial r = the same after both logFCs are residualised on the input bulk's logFC: the cell\u2013type signal the bulk does not carry. Top\u201350 DEG: LoRA\u20135 ahead in 12/20 draws, tied in 8 (Giroux and Ji: few DEG in the observed cells, mostly 0 for every method). Giroux's 5\u2013donor test sets hold 10\u201312 samples from 4 subjects (2 vs 2), so its disease metrics are noisy; there scSemiProfiler\u20135 is ahead. Giroux: sex\u2013linked genes excluded (severity is confounded with sex). Per cohort, mean over draws (r / partial r / DEG): Sepsis LoRA\u20135 0.35 / 0.30 / 16.4, scSemiProfiler\u20135 0.40 / 0.33 / 5.2; Imperial LoRA\u20135 0.75 / 0.60 / 113.6, scSemiProfiler\u20135 0.56 / 0.43 / 46.8; Giroux LoRA\u20135 0.30 / 0.19 / 0.2, scSemiProfiler\u20135 0.40 / 0.24 / 0.0; Ji LoRA\u20135 0.33 / 0.30 / 0.2, scSemiProfiler\u20135 0.15 / 0.09 / 0.0. Source: data/fig4_disease.csv (scores[_nosex]//fewshot_n5/draw[_scsemi5]).", + "numbers": [ + { + "value": "Sepsis LoRA\u20135 0.35 / 0.30 / 16.4, scSemiProfiler\u20135 0.40 / 0.33 / 5.2; Imperial LoRA\u20135 0.75 / 0.60 / 113.6, scSemiProfiler\u20135 0.56 / 0.43 / 46.8; Giroux LoRA\u20135 0.30 / 0.19 / 0.2, scSemiProfiler\u20135 0.40 / 0.24 / 0.0; Ji LoRA\u20135 0.33 / 0.30 / 0.2, scSemiProfiler\u20135 0.15 / 0.09 / 0.0", + "source": "data/fig4_disease.csv" + } + ], + "original_slide": 38, + "original_title": "Disease signal with 5 pairs: LoRA-5 vs others", + "original_footnote": "5 donors, 5 draws; bars = mean, dots = draws. Partial r = cell-type signal beyond the bulk.", + "figures_before_palette": [ + "figures/fig4_disease_fewpair.png" + ], + "source_page_131": 63, + "deck_group": "Figure 4", + "previous_section_label": "Fig. 4b", + "parent_page": 26, + "parent_enriched_page": 36, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig4_disease_fewpair.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 37, + "parent_154_page": 42, + "parent_155_page": 43 + }, + { + "kind": "content", + "section": "Figure 4", + "section_label": "Fig. 4b", + "text_source": { + "file": "data/fig4_disease_slide_text.json", + "index": 1 + }, + "title": "5-fold CV: disease signal", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig4_disease_fivefold.png" + ], + "keywords": [ + "logFC r up to 0.82 (Imperial)", + "vs scSemiProfiler: r 3/4, partial r 3/4, DEG 4/4 cohorts", + "Top PBISC version on logFC r in 4/4 cohorts" + ], + "footnote": "Original donor-level split and original signal estimator. These results are distinct from the exact-five update.", + "sources": "tools/fig4_disease.py; data/fig4_disease.csv", + "refs": [ + "Wang 2024, Nat Commun", + "Vi\u00f1as Torn\u00e9 2026, Nat Biotechnol" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1", + "Vi\u00f1as Torn\u00e9 et al. Systema: a framework for evaluating genetic perturbation response prediction beyond systematic variation. Nat Biotechnol 44:1050-1059 (2026). doi:10.1038/s41587-025-02777-8" + ], + "say": "With every fold's pairs, LoRA reaches a disease logFC r of 0.82 in Imperial and is the best PBISC version on logFC r in every cohort. This is reported separately from the five-pair comparison.", + "footnote_full": "Current scope: Original donor-level split and original signal estimator. These results are distinct from the exact-five update.\n\nOriginal slide 39: Reported separately from the 5-pair comparison: donor-level 5-fold CV, each sample generated by the fold model that never trained on it, one pooled scoring call per method; scSemiProfiler uses exactly LoRA's training pairs as representatives. Per cohort LoRA vs scSemiProfiler (r / partial r / DEG): Sepsis 0.53 / 0.47 / 23 vs 0.44 / 0.36 / 6; Imperial 0.82 / 0.69 / 143 vs 0.71 / 0.54 / 87; Giroux 0.46 / 0.32 / 5 vs 0.54 / 0.39 / 0; Ji 0.48 / 0.45 / 15 vs 0.27 / 0.23 / 0. Giroux: sex-linked genes excluded. Source: data/fig4_disease.csv (scores[_nosex]//pooled, scsemi_cv_20261002/pooled; sepsis DEG from pooled_deg).", + "numbers": [ + { + "value": "Sepsis 0.53 / 0.47 / 23 vs 0.44 / 0.36 / 6; Imperial 0.82 / 0.69 / 143 vs 0.71 / 0.54 / 87; Giroux 0.46 / 0.32 / 5 vs 0.54 / 0.39 / 0; Ji 0.48 / 0.45 / 15 vs 0.27 / 0.23 / 0", + "source": "data/fig4_disease.csv" + } + ], + "original_slide": 39, + "original_title": "5-fold CV: disease signal", + "original_footnote": "Donor-level 5-fold CV, pooled; reported separately from the 5-pair comparison.", + "figures_before_palette": [ + "figures/fig4_disease_fivefold.png" + ], + "source_page_131": 67, + "deck_group": "Figure 4", + "previous_section_label": "Fig. 4c", + "parent_page": 27, + "parent_enriched_page": 37, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig4_disease_fivefold.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 38, + "parent_154_page": 43, + "parent_155_page": 44 + }, + { + "kind": "content", + "section": "Figure 4", + "section_label": "Fig. 4c", + "text_source": { + "file": "data/sec09_cat_curves_slide_text.json", + "index": 0 + }, + "title": "Disease genes at every list length (CAT)", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_cat_curves.png" + ], + "keywords": [ + "LoRA top model: 3/4 (5 donors), 2/4 (5-fold)", + "LoRA-5 above scSemiProfiler-5: 3/4 cohorts (k = 100)", + "Chance at k = 100: 0.012\u20130.016" + ], + "footnote": "Existing CAT curves; ranking agreement and chance line. Patient-aware reanalysis reuses the same predictions.", + "sources": "tools/sec09_cat_curves.py; data/sec09_cat_curves.csv", + "refs": [ + "Irizarry 2005, Nat Methods", + "Webber 2010, ACM TOIS", + "Squair 2021, Nat Commun" + ], + "refs_full": [ + "Irizarry et al. Multiple-laboratory comparison of microarray platforms. Nat Methods 2:345-350 (2005). doi:10.1038/nmeth756", + "Webber, Moffat & Zobel. A similarity measure for indefinite rankings. ACM Trans Inf Syst 28(4):20 (2010). doi:10.1145/1852102.1852106", + "Squair et al. Confronting false discoveries in single-cell differential expression. Nat Commun 12:5692 (2021). doi:10.1038/s41467-021-25960-2" + ], + "footnote_full": "Current scope: Existing CAT curves; ranking agreement and chance line. Patient-aware reanalysis reuses the same predictions.\n\nOriginal slide 40: Correspondence at the top (Irizarry et al. 2005): per cell type, every gene tested on both sides is ranked by |limma-voom moderated t| (case vs control, same donor x type pseudobulks as the DEG panels); CAT(k) = share of the observed top-k genes that are also in the generated top-k with the same logFC sign. Mean over T / B / Mono / NK; 5 donors = mean over 5 draws (scSemiProfiler: its 25 draws), the main comparison; 5-fold = pooled call, shown separately. Chance = k / (2G), G = genes tested on both sides (0.012\u20130.016 at k = 100). No significance threshold, so cohorts where few genes pass FDR (Giroux, Ji) still get a curve. Rank-biased overlap (Webber et al. 2010, p = 0.98) gives the same order: scores_contrast/{fewshot,pooled}_rbo.csv. Giroux: sex-linked genes excluded. Sepsis CAT100 5 donors LoRA 0.092 / scSemi 0.037 / calibration 0.069 / raw 0.028 / chance 0.012; 5-fold LoRA 0.163; Imperial CAT100 5 donors LoRA 0.243 / scSemi 0.133 / calibration 0.198 / raw 0.179 / chance 0.012; 5-fold LoRA 0.270; Giroux CAT100 5 donors LoRA 0.048 / scSemi 0.075 / calibration 0.050 / raw 0.048 / chance 0.016; 5-fold LoRA 0.068; Ji CAT100 5 donors LoRA 0.091 / scSemi 0.030 / calibration 0.021 / raw 0.021 / chance 0.014; 5-fold LoRA 0.152.", + "numbers": [ + { + "value": "Sepsis CAT100 5 donors LoRA 0.092 / scSemi 0.037 / calibration 0.069 / raw 0.028 / chance 0.012; 5-fold LoRA 0.163; Imperial CAT100 5 donors LoRA 0.243 / scSemi 0.133 / calibration 0.198 / raw 0.179 / chance 0.012; 5-fold LoRA 0.270; Giroux CAT100 5 donors LoRA 0.048 / scSemi 0.075 / calibration 0.050 / raw 0.048 / chance 0.016; 5-fold LoRA 0.068; Ji CAT100 5 donors LoRA 0.091 / scSemi 0.030 / calibration 0.021 / raw 0.021 / chance 0.014; 5-fold LoRA 0.152", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/fewshot_cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/pooled_cat.csv" + } + ], + "original_slide": 40, + "original_title": "Disease genes at every list length (CAT)", + "original_footnote": "Same-sign overlap of observed and generated top-k genes per cell type; no FDR cut-off.", + "figures_before_palette": [ + "figures/sec09_cat_curves.png" + ], + "source_page_131": 68, + "deck_group": "Figure 4", + "previous_section_label": "Fig. 4d", + "parent_page": 28, + "parent_enriched_page": 38, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_cat_curves.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 39, + "parent_154_page": 44, + "parent_155_page": 45 + }, + { + "kind": "content", + "section": "Figure 4", + "section_label": "Fig. 4d", + "text_source": { + "file": "data/secS_deconvolution_slide_text.json", + "index": 1 + }, + "title": "Cell-type deconvolution, same pairs: disease signal", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/secS_deconv_disease.png" + ], + "keywords": [ + "logFC r: PBISC ahead in 15/20 draws (BayesPrism), 17/17 (CIBERSORTx)", + "Beyond bulk: PBISC ahead in 18/20 and 17/17", + "Top-50 DEG mixed: PBISC ahead in 6/20 and 6/17" + ], + "footnote": "Completed BayesPrism/CIBERSORTx comparisons; same five-donor support, method-supported expression metrics.", + "sources": "tools/secS_deconvolution.py; data/secS_deconvolution.csv; plans/bayesprism_paired_20261006/results/{RESULTS_full.md, RESULTS_cibersortx.md, TABLE_*_full.csv, TABLE_*_cibersortx.csv}", + "refs": [ + "Chu 2022, Nat Cancer", + "Newman 2019, Nat Biotechnol" + ], + "refs_full": [ + "Chu et al. Cell type and gene expression deconvolution with BayesPrism enables Bayesian integrative analysis across bulk and single-cell RNA sequencing in oncology. Nat Cancer 3:505-517 (2022). doi:10.1038/s43018-022-00356-3", + "Newman et al. Determining cell type abundance and expression from bulk tissues with digital cytometry. Nat Biotechnol 37:773-782 (2019). doi:10.1038/s41587-019-0114-2" + ], + "say": "Both methods also estimate expression per cell type: BayesPrism's posterior and CIBERSORTx HiRes. Their case-versus-control changes per cell type agree less with the truth than PBISC's in most draws (15 of 20 for BayesPrism, all 17 scored for CIBERSORTx; BayesPrism is ahead in Giroux). Beyond what the bulk itself carries, they add little: partial r at most 0.19, against 0.19 to 0.60 for PBISC. Top-50 DEG overlap is mixed: deconvolution is ahead in sepsis, PBISC in Imperial, and Ji and Giroux are near zero for every method.", + "footnote_full": "Current scope: Completed BayesPrism/CIBERSORTx comparisons; same five-donor support, method-supported expression metrics.\n\nOriginal slide 41: Cell-type deconvolution, not single cells. Case-vs-control per-cell-type logFC (mean CP10K; T, B, Monocyte, NK) from BayesPrism's posterior cell-type expression Z and CIBERSORTx HiRes, scored with the benchmark's own functions; partial r = the same r after both logFCs are residualised on the input bulk's logFC (the cell-type signal the bulk does not carry); top-50 DEG = Welch, summed over types (many ties at 0 in Ji and Giroux). 5-donor draws (the 5-pair comparison), mean over draws with a score. Giroux excludes sex-linked genes, as everywhere in the deck. CIBERSORTx: official Docker images (x86, run under qemu emulation on server_1), single-cell reference mode with S-mode batch correction (up to 400 reference cells per type), fractions in cell units (divided by each type's mean UMI per reference cell). CIBERSORTx HiRes gives per-sample cell-type expression for the genes it imputes per type (others are written as 1 and count as zero here); it imputed no gene of the four contrast types in 1 Ji ChILI and 2 Giroux draws, which therefore have no CIBERSORTx disease score. Source: plans/bayesprism_paired_20261006/results/RESULTS_full.md, RESULTS_cibersortx.md.", + "numbers": [ + { + "value": "Sepsis: logFC r PBISC 0.350 / BayesPrism 0.241 / CIBERSORTx 0.069; partial r 0.305 / 0.095 / -0.010; DEG 16.4 / 27.0 / 25.4; Ji ChILI: logFC r PBISC 0.333 / BayesPrism 0.193 / CIBERSORTx 0.158; partial r 0.300 / 0.032 / 0.005; DEG 0.2 / 2.0 / 0.0; Imperial: logFC r PBISC 0.749 / BayesPrism 0.598 / CIBERSORTx 0.283; partial r 0.603 / 0.187 / 0.061; DEG 113.6 / 72.4 / 30.8; Giroux: logFC r PBISC 0.297 / BayesPrism 0.371 / CIBERSORTx 0.095; partial r 0.189 / 0.117 / 0.029; DEG 0.2 / 0.2 / 0.0", + "source": "data/secS_deconvolution.csv" + }, + { + "value": "draws where PBISC LoRA-5 beats BayesPrism: logFC r 15/20, partial r 18/20, top-50 DEG 6 (BayesPrism ahead 8, ties 6); beats CIBERSORTx: logFC r 17/17, partial r 17/17, top-50 DEG 6 (CIBERSORTx ahead 4, ties 7)", + "source": "plans/bayesprism_paired_20261006/results/TABLE_fewshot_full.csv, TABLE_fewshot_cibersortx.csv" + } + ], + "original_slide": 41, + "original_title": "Cell-type deconvolution, same pairs: disease signal", + "original_footnote": "5-donor draws, mean over draws. Partial r = cell-type signal beyond the bulk.", + "figures_before_palette": [ + "figures/secS_deconv_disease.png" + ], + "source_page_131": 71, + "deck_group": "Figure 4", + "previous_section_label": "Fig. 4e", + "parent_page": 29, + "parent_enriched_page": 39, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/secS_deconv_disease.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 40, + "parent_154_page": 45, + "parent_155_page": 46 + }, + { + "kind": "content", + "layout": "full", + "key": "strengthen_cv_complete_cat", + "role": "f4_disease_metrics", + "title": "Disease gene rankings: complete CV comparison", + "subtitle": "Same-sign correspondence at the top; all four disease cohorts", + "section_label": "Fig. 4 \u00b7 Disease integrity", + "figures": [ + "figures/strengthening_20261013/disease_metrics/cv_complete_cat.png" + ], + "figure": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/strengthening_20261013/disease_metrics/cv_complete_cat.png", + "palette": "native", + "keywords": [], + "footnote": "Historical 5-fold CV; mean over five types. All saved scSemi curves included. Imperial retains visit-level sampling; not a strict-five result.", + "footnote_full": "Historical pooled five-fold CV, all available training pairs with inner-validation access; not strict-five support. Methods share each cohort's scored sample set, disease contrast, reference, observed detection mask and gene exclusions. Imperial repeated sample/visit rows remain a historical estimand, not independent patients. Giroux excludes recorded sex-linked genes but this does not remove sex/severity confounding or establish causal disease effects. The common deployable comparator family is raw/calibrated zero-shot, all-pair LoRA and scSemiProfiler, with raw-bulk broadcast as a null; Sepsis full FT is additionally shown. Other budget/recipe/sc-only/oracle arms are recorded in the inventory and retained in the original deck. BayesPrism/CIBERSORTx results belong to different saved support/endpoint contracts and are not silently merged. CAT and RBO average five types (T/B/Monocyte/NK/DC); Pearson and partial-r average four types (T/B/Monocyte/NK). CAT ranks limma-voom primary nonzero-filter genes by absolute moderated t, requires matching logFC sign, and is plotted at k=10/20/50/100/200/500. Tested ranking universes can differ by method; counts are retained. RBO is extrapolated Eq.32 at p=.98, averaging ascending/descending t rankings. Partial r removes each arm's own condition-specific bulk-broadcast vector by linear residualization; broadcast versus itself is undefined and omitted, not zero. Observed-repeat r splits the same samples' cells, so it is a conditional cell-resampling reference, not an independent-patient ceiling or CI. No new test, fit or generation.", + "notes": "Historical pooled five-fold CV, all available training pairs with inner-validation access; not strict-five support. Methods share each cohort's scored sample set, disease contrast, reference, observed detection mask and gene exclusions. Imperial repeated sample/visit rows remain a historical estimand, not independent patients. Giroux excludes recorded sex-linked genes but this does not remove sex/severity confounding or establish causal disease effects. The common deployable comparator family is raw/calibrated zero-shot, all-pair LoRA and scSemiProfiler, with raw-bulk broadcast as a null; Sepsis full FT is additionally shown. Other budget/recipe/sc-only/oracle arms are recorded in the inventory and retained in the original deck. BayesPrism/CIBERSORTx results belong to different saved support/endpoint contracts and are not silently merged. CAT and RBO average five types (T/B/Monocyte/NK/DC); Pearson and partial-r average four types (T/B/Monocyte/NK). CAT ranks limma-voom primary nonzero-filter genes by absolute moderated t, requires matching logFC sign, and is plotted at k=10/20/50/100/200/500. Tested ranking universes can differ by method; counts are retained. RBO is extrapolated Eq.32 at p=.98, averaging ascending/descending t rankings. Partial r removes each arm's own condition-specific bulk-broadcast vector by linear residualization; broadcast versus itself is undefined and omitted, not zero. Observed-repeat r splits the same samples' cells, so it is a conditional cell-resampling reference, not an independent-patient ceiling or CI. No new test, fit or generation.", + "note": "Historical pooled five-fold CV, all available training pairs with inner-validation access; not strict-five support. Methods share each cohort's scored sample set, disease contrast, reference, observed detection mask and gene exclusions. Imperial repeated sample/visit rows remain a historical estimand, not independent patients. Giroux excludes recorded sex-linked genes but this does not remove sex/severity confounding or establish causal disease effects. The common deployable comparator family is raw/calibrated zero-shot, all-pair LoRA and scSemiProfiler, with raw-bulk broadcast as a null; Sepsis full FT is additionally shown. Other budget/recipe/sc-only/oracle arms are recorded in the inventory and retained in the original deck. BayesPrism/CIBERSORTx results belong to different saved support/endpoint contracts and are not silently merged. CAT and RBO average five types (T/B/Monocyte/NK/DC); Pearson and partial-r average four types (T/B/Monocyte/NK). CAT ranks limma-voom primary nonzero-filter genes by absolute moderated t, requires matching logFC sign, and is plotted at k=10/20/50/100/200/500. Tested ranking universes can differ by method; counts are retained. RBO is extrapolated Eq.32 at p=.98, averaging ascending/descending t rankings. Partial r removes each arm's own condition-specific bulk-broadcast vector by linear residualization; broadcast versus itself is undefined and omitted, not zero. Observed-repeat r splits the same samples' cells, so it is a conditional cell-resampling reference, not an independent-patient ceiling or CI. No new test, fit or generation.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_disease_metrics_20261013.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/style.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_assignment_20261013/AUTHOR_SCOPE.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/strengthening_20261013/STRENGTHENING_PROPOSAL_before_revision.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/contrast_extra.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/score_strict.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/donor_statistics.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/donor_limma.R; /home/sj_server_1/PBISC/paper-release-20260928/reference-papers/evaluation-metrics/Irizarry2005_Multiple_laboratory_comparison_microarray_platforms.pdf; /home/sj_server_1/PBISC/paper-release-20260928/reference-papers/evaluation-metrics/Webber2010_Rank_biased_overlap_similarity_indefinite_rankings.pdf; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/ceiling.csv", + "metric_references": [ + { + "metric": "CAT", + "citation": "Irizarry et al. Nature Methods 2005;2:345\u2013350", + "url": "https://doi.org/10.1038/nmeth756", + "verification": "Read local original paper: CAT is top-list overlap versus list size. Same-sign agreement and moderated-t ranking are this benchmark's explicit extensions." + }, + { + "metric": "RBO", + "citation": "Webber, Moffat & Zobel. ACM TOIS 2010;28:20", + "url": "https://doi.org/10.1145/1852102.1852106", + "verification": "Read local original paper Eq. 32; saved scorer uses extrapolated RBO, p=0.98. Displayed value averages descending and ascending t rankings, then five cell types." + }, + { + "metric": "limma-voom", + "citation": "Law et al. Genome Biology 2014;15:R29", + "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4053721/", + "verification": "Primary full text checked: log-count precision weights enter limma. Saved donor-aware implementation uses edgeR TMM, voom, lmFit and eBayes." + }, + { + "metric": "biological replication", + "citation": "Squair et al. Nature Communications 2021;12:5692", + "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8479118/", + "verification": "Primary paper supports using biological replicates rather than independent-cell inference. Existing donor-aware DEG tables are reused; no new fit or CI is computed." + } + ], + "refs": [ + "Irizarry et al. 2005; Law et al. 2014" + ], + "refs_full": [ + "Irizarry et al. Nature Methods 2005;2:345\u2013350. https://doi.org/10.1038/nmeth756 \u2014 Read local original paper: CAT is top-list overlap versus list size. Same-sign agreement and moderated-t ranking are this benchmark's explicit extensions.", + "Webber, Moffat & Zobel. ACM TOIS 2010;28:20. https://doi.org/10.1145/1852102.1852106 \u2014 Read local original paper Eq. 32; saved scorer uses extrapolated RBO, p=0.98. Displayed value averages descending and ascending t rankings, then five cell types.", + "Law et al. Genome Biology 2014;15:R29. https://pmc.ncbi.nlm.nih.gov/articles/PMC4053721/ \u2014 Primary full text checked: log-count precision weights enter limma. Saved donor-aware implementation uses edgeR TMM, voom, lmFit and eBayes.", + "Squair et al. Nature Communications 2021;12:5692. https://pmc.ncbi.nlm.nih.gov/articles/PMC8479118/ \u2014 Primary paper supports using biological replicates rather than independent-cell inference. Existing donor-aware DEG tables are reused; no new fit or CI is computed." + ], + "deck_group": "Figure 4", + "section": "Figure 4", + "addition_key": "strengthening:disease_metrics:0", + "parent_154_page": 46, + "parent_155_page": 47 + }, + { + "kind": "content", + "layout": "full", + "key": "strengthen_cv_multimetric", + "role": "f4_disease_metrics", + "title": "Disease integrity depends on cohort and endpoint", + "subtitle": "Correlation, bulk-adjusted correlation and rank agreement", + "section_label": "Fig. 4 \u00b7 Disease integrity", + "figures": [ + "figures/strengthening_20261013/disease_metrics/cv_multimetric.png" + ], + "figure": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/strengthening_20261013/disease_metrics/cv_multimetric.png", + "palette": "native", + "keywords": [], + "footnote": "Historical 5-fold CV. r: four types; RBO: five types, p=0.98. Gray repeat markers reflect same-patient cell resampling, not independent ceilings.", + "footnote_full": "Historical pooled five-fold CV, all available training pairs with inner-validation access; not strict-five support. Methods share each cohort's scored sample set, disease contrast, reference, observed detection mask and gene exclusions. Imperial repeated sample/visit rows remain a historical estimand, not independent patients. Giroux excludes recorded sex-linked genes but this does not remove sex/severity confounding or establish causal disease effects. The common deployable comparator family is raw/calibrated zero-shot, all-pair LoRA and scSemiProfiler, with raw-bulk broadcast as a null; Sepsis full FT is additionally shown. Other budget/recipe/sc-only/oracle arms are recorded in the inventory and retained in the original deck. BayesPrism/CIBERSORTx results belong to different saved support/endpoint contracts and are not silently merged. CAT and RBO average five types (T/B/Monocyte/NK/DC); Pearson and partial-r average four types (T/B/Monocyte/NK). CAT ranks limma-voom primary nonzero-filter genes by absolute moderated t, requires matching logFC sign, and is plotted at k=10/20/50/100/200/500. Tested ranking universes can differ by method; counts are retained. RBO is extrapolated Eq.32 at p=.98, averaging ascending/descending t rankings. Partial r removes each arm's own condition-specific bulk-broadcast vector by linear residualization; broadcast versus itself is undefined and omitted, not zero. Observed-repeat r splits the same samples' cells, so it is a conditional cell-resampling reference, not an independent-patient ceiling or CI. No new test, fit or generation.", + "notes": "Historical pooled five-fold CV, all available training pairs with inner-validation access; not strict-five support. Methods share each cohort's scored sample set, disease contrast, reference, observed detection mask and gene exclusions. Imperial repeated sample/visit rows remain a historical estimand, not independent patients. Giroux excludes recorded sex-linked genes but this does not remove sex/severity confounding or establish causal disease effects. The common deployable comparator family is raw/calibrated zero-shot, all-pair LoRA and scSemiProfiler, with raw-bulk broadcast as a null; Sepsis full FT is additionally shown. Other budget/recipe/sc-only/oracle arms are recorded in the inventory and retained in the original deck. BayesPrism/CIBERSORTx results belong to different saved support/endpoint contracts and are not silently merged. CAT and RBO average five types (T/B/Monocyte/NK/DC); Pearson and partial-r average four types (T/B/Monocyte/NK). CAT ranks limma-voom primary nonzero-filter genes by absolute moderated t, requires matching logFC sign, and is plotted at k=10/20/50/100/200/500. Tested ranking universes can differ by method; counts are retained. RBO is extrapolated Eq.32 at p=.98, averaging ascending/descending t rankings. Partial r removes each arm's own condition-specific bulk-broadcast vector by linear residualization; broadcast versus itself is undefined and omitted, not zero. Observed-repeat r splits the same samples' cells, so it is a conditional cell-resampling reference, not an independent-patient ceiling or CI. No new test, fit or generation.", + "note": "Historical pooled five-fold CV, all available training pairs with inner-validation access; not strict-five support. Methods share each cohort's scored sample set, disease contrast, reference, observed detection mask and gene exclusions. Imperial repeated sample/visit rows remain a historical estimand, not independent patients. Giroux excludes recorded sex-linked genes but this does not remove sex/severity confounding or establish causal disease effects. The common deployable comparator family is raw/calibrated zero-shot, all-pair LoRA and scSemiProfiler, with raw-bulk broadcast as a null; Sepsis full FT is additionally shown. Other budget/recipe/sc-only/oracle arms are recorded in the inventory and retained in the original deck. BayesPrism/CIBERSORTx results belong to different saved support/endpoint contracts and are not silently merged. CAT and RBO average five types (T/B/Monocyte/NK/DC); Pearson and partial-r average four types (T/B/Monocyte/NK). CAT ranks limma-voom primary nonzero-filter genes by absolute moderated t, requires matching logFC sign, and is plotted at k=10/20/50/100/200/500. Tested ranking universes can differ by method; counts are retained. RBO is extrapolated Eq.32 at p=.98, averaging ascending/descending t rankings. Partial r removes each arm's own condition-specific bulk-broadcast vector by linear residualization; broadcast versus itself is undefined and omitted, not zero. Observed-repeat r splits the same samples' cells, so it is a conditional cell-resampling reference, not an independent-patient ceiling or CI. No new test, fit or generation.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_disease_metrics_20261013.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/style.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_assignment_20261013/AUTHOR_SCOPE.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/strengthening_20261013/STRENGTHENING_PROPOSAL_before_revision.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/contrast_extra.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/score_strict.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/donor_statistics.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/donor_limma.R; /home/sj_server_1/PBISC/paper-release-20260928/reference-papers/evaluation-metrics/Irizarry2005_Multiple_laboratory_comparison_microarray_platforms.pdf; /home/sj_server_1/PBISC/paper-release-20260928/reference-papers/evaluation-metrics/Webber2010_Rank_biased_overlap_similarity_indefinite_rankings.pdf; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/ceiling.csv", + "metric_references": [ + { + "metric": "CAT", + "citation": "Irizarry et al. Nature Methods 2005;2:345\u2013350", + "url": "https://doi.org/10.1038/nmeth756", + "verification": "Read local original paper: CAT is top-list overlap versus list size. Same-sign agreement and moderated-t ranking are this benchmark's explicit extensions." + }, + { + "metric": "RBO", + "citation": "Webber, Moffat & Zobel. ACM TOIS 2010;28:20", + "url": "https://doi.org/10.1145/1852102.1852106", + "verification": "Read local original paper Eq. 32; saved scorer uses extrapolated RBO, p=0.98. Displayed value averages descending and ascending t rankings, then five cell types." + }, + { + "metric": "limma-voom", + "citation": "Law et al. Genome Biology 2014;15:R29", + "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4053721/", + "verification": "Primary full text checked: log-count precision weights enter limma. Saved donor-aware implementation uses edgeR TMM, voom, lmFit and eBayes." + }, + { + "metric": "biological replication", + "citation": "Squair et al. Nature Communications 2021;12:5692", + "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8479118/", + "verification": "Primary paper supports using biological replicates rather than independent-cell inference. Existing donor-aware DEG tables are reused; no new fit or CI is computed." + } + ], + "refs": [ + "Webber et al. 2010" + ], + "refs_full": [ + "Irizarry et al. Nature Methods 2005;2:345\u2013350. https://doi.org/10.1038/nmeth756 \u2014 Read local original paper: CAT is top-list overlap versus list size. Same-sign agreement and moderated-t ranking are this benchmark's explicit extensions.", + "Webber, Moffat & Zobel. ACM TOIS 2010;28:20. https://doi.org/10.1145/1852102.1852106 \u2014 Read local original paper Eq. 32; saved scorer uses extrapolated RBO, p=0.98. Displayed value averages descending and ascending t rankings, then five cell types.", + "Law et al. Genome Biology 2014;15:R29. https://pmc.ncbi.nlm.nih.gov/articles/PMC4053721/ \u2014 Primary full text checked: log-count precision weights enter limma. Saved donor-aware implementation uses edgeR TMM, voom, lmFit and eBayes.", + "Squair et al. Nature Communications 2021;12:5692. https://pmc.ncbi.nlm.nih.gov/articles/PMC8479118/ \u2014 Primary paper supports using biological replicates rather than independent-cell inference. Existing donor-aware DEG tables are reused; no new fit or CI is computed." + ], + "deck_group": "Figure 4", + "section": "Figure 4", + "addition_key": "strengthening:disease_metrics:1", + "parent_154_page": 47, + "parent_155_page": 48 + }, + { + "kind": "content", + "layout": "full", + "key": "strengthen_strict_sepsis_multimetric", + "role": "f4_disease_metrics", + "title": "Five-pair Sepsis: metrics across all support draws", + "subtitle": "Complete Sepsis subset; no inference across unfinished cohorts", + "section_label": "Fig. 4 \u00b7 Disease integrity", + "figures": [ + "figures/strengthening_20261013/disease_metrics/strict_sepsis_multimetric.png" + ], + "figure": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/strengthening_20261013/disease_metrics/strict_sepsis_multimetric.png", + "palette": "native", + "keywords": [], + "footnote": "Sepsis draws 0\u20134 only; five total support pairs. Points show overlapping support draws, not independent replicates or CI. Donor-aware DEG reused.", + "footnote_full": "Frozen Sepsis draws 0\u20134 only; every draw has a completed scoring receipt with verified byte sizes and SHA-256 values. Each support has exactly five unique samples from five unique donors, no held-out donor overlap. Strict protocol uses fixed 3000 LoRA steps, one sample per donor, no extra target validation selection; paired calibration/scSemi share support. Raw and external-reference calibration are zero-target-pair controls. Reference calibration additionally anchors on unlabelled target bulks, so information access differs as specified in the saved protocol. Each point is one support draw; donors overlap across draws and dots are not independent biological replicates. Diamonds are descriptive means, not CI or statistical tests. All five saved generator arms included. Pearson/partial r average four cell types; CAT100 averages five. DEG sums top-50 overlap over T/B/Monocyte/NK/DC after existing donor-condition pseudobulk edgeR-TMM limma-voom fits, nonzero gene filter and BH FDR<.05, ranked by absolute logFC; overlaps do not themselves enforce same direction. The theoretical upper bound is 250, but the actual observed significant-list total is saved for each draw and can be lower. CAT is same-sign moderated-t ranking agreement, not FDR significance. Ranking universes may differ by method. No new DEG fit, array scoring, training or inference. Receipt outputs directly bind status and contrast manifests; the referenced numerical CSVs are additionally frozen and SHA-bound by this figure bundle, not claimed to have an original receipt hash they never had. Other cohorts and deconvolution endpoint families are not combined with this completed Sepsis subset.", + "notes": "Frozen Sepsis draws 0\u20134 only; every draw has a completed scoring receipt with verified byte sizes and SHA-256 values. Each support has exactly five unique samples from five unique donors, no held-out donor overlap. Strict protocol uses fixed 3000 LoRA steps, one sample per donor, no extra target validation selection; paired calibration/scSemi share support. Raw and external-reference calibration are zero-target-pair controls. Reference calibration additionally anchors on unlabelled target bulks, so information access differs as specified in the saved protocol. Each point is one support draw; donors overlap across draws and dots are not independent biological replicates. Diamonds are descriptive means, not CI or statistical tests. All five saved generator arms included. Pearson/partial r average four cell types; CAT100 averages five. DEG sums top-50 overlap over T/B/Monocyte/NK/DC after existing donor-condition pseudobulk edgeR-TMM limma-voom fits, nonzero gene filter and BH FDR<.05, ranked by absolute logFC; overlaps do not themselves enforce same direction. The theoretical upper bound is 250, but the actual observed significant-list total is saved for each draw and can be lower. CAT is same-sign moderated-t ranking agreement, not FDR significance. Ranking universes may differ by method. No new DEG fit, array scoring, training or inference. Receipt outputs directly bind status and contrast manifests; the referenced numerical CSVs are additionally frozen and SHA-bound by this figure bundle, not claimed to have an original receipt hash they never had. Other cohorts and deconvolution endpoint families are not combined with this completed Sepsis subset.", + "note": "Frozen Sepsis draws 0\u20134 only; every draw has a completed scoring receipt with verified byte sizes and SHA-256 values. Each support has exactly five unique samples from five unique donors, no held-out donor overlap. Strict protocol uses fixed 3000 LoRA steps, one sample per donor, no extra target validation selection; paired calibration/scSemi share support. Raw and external-reference calibration are zero-target-pair controls. Reference calibration additionally anchors on unlabelled target bulks, so information access differs as specified in the saved protocol. Each point is one support draw; donors overlap across draws and dots are not independent biological replicates. Diamonds are descriptive means, not CI or statistical tests. All five saved generator arms included. Pearson/partial r average four cell types; CAT100 averages five. DEG sums top-50 overlap over T/B/Monocyte/NK/DC after existing donor-condition pseudobulk edgeR-TMM limma-voom fits, nonzero gene filter and BH FDR<.05, ranked by absolute logFC; overlaps do not themselves enforce same direction. The theoretical upper bound is 250, but the actual observed significant-list total is saved for each draw and can be lower. CAT is same-sign moderated-t ranking agreement, not FDR significance. Ranking universes may differ by method. No new DEG fit, array scoring, training or inference. Receipt outputs directly bind status and contrast manifests; the referenced numerical CSVs are additionally frozen and SHA-bound by this figure bundle, not claimed to have an original receipt hash they never had. Other cohorts and deconvolution endpoint families are not combined with this completed Sepsis subset.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/strengthen_disease_metrics_20261013.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/style.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/PBISC-Diffusion_color_palette_20261013.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/figure_assignment_20261013/AUTHOR_SCOPE.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/strengthening_20261013/STRENGTHENING_PROPOSAL_before_revision.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/contrast_extra.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/score_strict.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/donor_statistics.py; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/tools/donor_limma.R; /home/sj_server_1/PBISC/paper-release-20260928/reference-papers/evaluation-metrics/Irizarry2005_Multiple_laboratory_comparison_microarray_platforms.pdf; /home/sj_server_1/PBISC/paper-release-20260928/reference-papers/evaluation-metrics/Webber2010_Rank_biased_overlap_similarity_indefinite_rankings.pdf; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ye2026_sepsis/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/imperial_eskd/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/giroux2022/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/cat.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/rbo.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_contrast/ji2025_chili/scsemi_cv_20261002/pooled/ceiling.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/protocol/PROTOCOL.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/score/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/status.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/donor_bootstrap_2000.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/contrast/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/protocol/splits/ye2026_sepsis/disease/draw0/support.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/contrast/ye2026_sepsis/strict/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/score/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/status.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/donor_bootstrap_2000.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/contrast/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/protocol/splits/ye2026_sepsis/disease/draw1/support.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/contrast/ye2026_sepsis/strict/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/donor_deg/donor_deg_summary.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/score/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/status.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/donor_bootstrap_2000.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/contrast/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/protocol/splits/ye2026_sepsis/disease/draw2/support.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/contrast/ye2026_sepsis/strict/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/donor_deg/donor_deg_summary.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/score/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/status.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/donor_bootstrap_2000.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/contrast/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/protocol/splits/ye2026_sepsis/disease/draw3/support.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/contrast/ye2026_sepsis/strict/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/donor_deg/donor_deg_summary.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/score/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/status.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/donor_bootstrap_2000.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/contrast/ye2026_sepsis/strict/manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/protocol/splits/ye2026_sepsis/disease/draw4/support.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/contrast/ye2026_sepsis/strict/headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/donor_deg/donor_deg_summary.csv", + "metric_references": [ + { + "metric": "CAT", + "citation": "Irizarry et al. Nature Methods 2005;2:345\u2013350", + "url": "https://doi.org/10.1038/nmeth756", + "verification": "Read local original paper: CAT is top-list overlap versus list size. Same-sign agreement and moderated-t ranking are this benchmark's explicit extensions." + }, + { + "metric": "RBO", + "citation": "Webber, Moffat & Zobel. ACM TOIS 2010;28:20", + "url": "https://doi.org/10.1145/1852102.1852106", + "verification": "Read local original paper Eq. 32; saved scorer uses extrapolated RBO, p=0.98. Displayed value averages descending and ascending t rankings, then five cell types." + }, + { + "metric": "limma-voom", + "citation": "Law et al. Genome Biology 2014;15:R29", + "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4053721/", + "verification": "Primary full text checked: log-count precision weights enter limma. Saved donor-aware implementation uses edgeR TMM, voom, lmFit and eBayes." + }, + { + "metric": "biological replication", + "citation": "Squair et al. Nature Communications 2021;12:5692", + "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8479118/", + "verification": "Primary paper supports using biological replicates rather than independent-cell inference. Existing donor-aware DEG tables are reused; no new fit or CI is computed." + } + ], + "refs": [ + "Law et al. 2014; Squair et al. 2021" + ], + "refs_full": [ + "Irizarry et al. Nature Methods 2005;2:345\u2013350. https://doi.org/10.1038/nmeth756 \u2014 Read local original paper: CAT is top-list overlap versus list size. Same-sign agreement and moderated-t ranking are this benchmark's explicit extensions.", + "Webber, Moffat & Zobel. ACM TOIS 2010;28:20. https://doi.org/10.1145/1852102.1852106 \u2014 Read local original paper Eq. 32; saved scorer uses extrapolated RBO, p=0.98. Displayed value averages descending and ascending t rankings, then five cell types.", + "Law et al. Genome Biology 2014;15:R29. https://pmc.ncbi.nlm.nih.gov/articles/PMC4053721/ \u2014 Primary full text checked: log-count precision weights enter limma. Saved donor-aware implementation uses edgeR TMM, voom, lmFit and eBayes.", + "Squair et al. Nature Communications 2021;12:5692. https://pmc.ncbi.nlm.nih.gov/articles/PMC8479118/ \u2014 Primary paper supports using biological replicates rather than independent-cell inference. Existing donor-aware DEG tables are reused; no new fit or CI is computed." + ], + "deck_group": "Figure 4", + "section": "Figure 4", + "addition_key": "strengthening:disease_metrics:2", + "parent_154_page": 48, + "parent_155_page": 49 + }, + { + "kind": "divider", + "title": "Figure 5. Rare-population identity", + "original_slide": 47, + "original_title": "Figure 5. Rare cell states in disease cohorts", + "original_footnote": "", + "source_page_131": 86, + "deck_group": "Figure 5", + "section": "Figure 5", + "previous_section_label": null, + "section_label": "Figure 5", + "parent_page": 30, + "parent_enriched_page": 40, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 41, + "parent_154_page": 49, + "parent_155_page": 50 + }, + { + "kind": "plan_card", + "section": "Figure 5", + "section_label": "Figure 5 plan", + "title": "Figure 5: paired rare-population identity", + "footnote": "Observed marker diagnostics are not completed author-cluster reproduction. Matched observed/PBISC/scSemi identity panels remain the next analysis.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "Observed marker diagnostics are not completed author-cluster reproduction. Matched observed/PBISC/scSemi identity panels remain the next analysis.", + "original_slide": 48, + "original_title": "Figure 5 plan: Rare cell states in disease cohorts", + "original_footnote": "LoRA cells carry rare disease states near their observed fractions, in the donors that have them. 3 of 5 panels ready.", + "source_page_131": 87, + "deck_group": "Figure 5", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_fig5.png" + ], + "rows": [ + [ + "5a", + "Original-paper targets and observed-truth gates", + "pp. 53, 54, 55" + ], + [ + "5b", + "Ye: pediatric pulmonary cycling monocytes", + "Observed marker diagnostic; exact cluster TODO" + ], + [ + "5c", + "Imperial: wave2 steroid-associated monocytes", + "Match steroid timing and fine labels: TODO" + ], + [ + "Gate", + "Giroux/Ji: verify a paper-reported rare target first", + "No forced rare-population claim" + ] + ], + "parent_page": 31, + "parent_enriched_page": 41, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 42, + "parent_154_page": 50, + "parent_155_page": 51 + }, + { + "kind": "content", + "key": "rare_literature_published_population_targets", + "role": "f5_original_paper_observed_gate", + "title": "Figure 5 targets begin with the source paper", + "section_label": "Fig. 5 \u00b7 Source-defined populations", + "layout": "full", + "keywords": [], + "figures": [ + "figures/strengthening_20261013/rare_literature/published_population_targets.png" + ], + "palette": "biological", + "footnote": "Ye: Fig. 5 proliferating pediatric monocytes. Imperial: Fig. 5 wave-2 steroid-associated monocytes; fine labels/timing still needed.", + "footnote_full": "Original-cohort retrospective analysis. No new generation, label inference or embedding fit. Literature targets are not automatically rare; rarity and identity require observed evidence. Ye: Fig. 5 proliferating pediatric monocytes. Imperial: Fig. 5 wave-2 steroid-associated monocytes; fine labels/timing still needed. [{\"cohort\": \"ye2026_sepsis\", \"citation\": \"Ye et al., Nature Immunology 2026\", \"doi\": \"10.1038/s41590-025-02345-x\", \"url\": \"https://www.nature.com/articles/s41590-025-02345-x\", \"figure\": \"Fig. 5a\\u2013f; validation Fig. 7f / Extended Data Fig. 9e\", \"population\": \"Mono_c06-CD14-MKI67\", \"kind\": \"Reported discrete proliferating monocyte subset; rarity must be measured locally\", \"setting\": \"Pediatric pulmonary sepsis (paper: within 48 h of onset); pediatric HC/ICU, abdominal sepsis and adults remain controls.\", \"markers\": [\"CD14\", \"MKI67\", \"STMN1\", \"MPO\", \"ELANE\"], \"negative_markers\": [], \"interpretation\": \"MKI67/STMN1/MPO/ELANE jointly support the published proliferative program in a CD14 monocyte context. PCLAF is not required by this original Fig. 5 definition.\", \"gate\": \"No fine-cluster labels found in 133 audited canonical obs schemas or the paired table. Marker co-detection is not a Mono_c06 cluster assignment.\", \"pdf\": \"/home/sj_server_1/PBISC/paper-release-20260928/reference-papers/paired-cohorts/Single-cell multi-omic landscape reveals anatomical-specific immune features in adult and pediatric sepsis.pdf\", \"all_paired_samples\": 85, \"identity_passed_samples\": 82, \"identity_passed_donors\": 82, \"legacy_cache_genes\": 73, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/ye2026_sepsis/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis/cells_obs.parquet\", \"observed_author_labels\": {\"\": 610713}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis/cells_panel6k.npz\", \"wave_counts\": {}, \"sex_by_group\": {}, \"available_cache_markers\": [\"MKI67\", \"MPO\", \"ELANE\"], \"missing_cache_markers\": [\"CD14\", \"STMN1\"]}, {\"cohort\": \"imperial_eskd\", \"citation\": \"Stephenson et al., Cell Genomics 2025\", \"doi\": \"10.1016/j.xgen.2025.100918\", \"url\": \"https://pmc.ncbi.nlm.nih.gov/articles/PMC12366655/\", \"figure\": \"Fig. 5A\\u2013G (especially 5C\\u2013E)\", \"population\": \"Dexamethasone-related monocytes\", \"kind\": \"Reported treatment-associated discrete subset; not automatically rare\", \"setting\": \"Wave 2 COVID-positive patients receiving glucocorticoids; align each blood visit to treatment start.\", \"markers\": [\"CD163\", \"ADAMTS2\", \"SLC1A3\", \"SESN1\", \"FPR1\", \"MTSS1\", \"MFGE8\", \"MRC1\"], \"negative_markers\": [\"JUN\", \"CXCL8\", \"HLA-DRA\", \"HLA-DRB5\"], \"interpretation\": \"The paper contrasts an anti-inflammatory/phagocytic program with lower inflammatory and antigen-presentation expression. Fig. 3 HLA-II suppression is a program, not a rare-cell identity.\", \"gate\": \"Paired metadata include wave 2, but the available author labels collapse these cells into broad monocytes. Treatment dates and fine labels must be joined before frequency recovery.\", \"pdf\": \"/home/sj_server_1/PBISC/paper-release-20260928/reference-papers/paired-cohorts/Stephenson2025_CellGenom_temporal_multiomics_COVID19_ESKD.pdf\", \"all_paired_samples\": 172, \"identity_passed_samples\": 158, \"identity_passed_donors\": 58, \"legacy_cache_genes\": 73, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/imperial_eskd/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/imperial_eskd/cells_obs.parquet\", \"observed_author_labels\": {\"T CD4 Helper\": 85235, \"NK\": 77242, \"T CD8 CTL\": 70657, \"CD14mono\": 53967, \"T CD4 Naive\": 49943, \"B_naive\": 33633, \"CD16mono\": 29056, \"T CD8 CM\": 19962, \"T CD8 Naive\": 18113, \"T CD4 CTL\": 17209, \"B_switched_mem\": 9223, \"T reg\": 9176, \"NK CD56 bright\": 8610, \"Int.mono\": 7136, \"DC2\": 5972, \"NKT\": 4647, \"DC3\": 4318, \"T g/d\": 3911, \"pDC\": 3034, \"B_non-switched_mem\": 2815, \"RBC\": 2604, \"Platelet\": 2576, \"B_ASC_IgA\": 2538, \"B_CD11c\": 2108, \"MAIT\": 1984, \"T CD8 EM\": 1860, \"B_ASC_dividing\": 1785, \"B_ASC_IgG\": 1182, \"HSPC\": 1004, \"DC1\": 754, \"B_ASC_IgM\": 499, \"T CD4 cycling\": 447, \"T CD8 EMRA\": 400, \"ASDC\": 231, \"ILC\": 147, \"B_ASC_other\": 36}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/imperial_eskd/cells_panel6k.npz\", \"wave_counts\": {\"w2\": 83, \"w1\": 75, \"w1|w2\": 14}, \"sex_by_group\": {}, \"available_cache_markers\": [\"CD163\", \"HLA-DRA\"], \"missing_cache_markers\": [\"ADAMTS2\", \"SLC1A3\", \"SESN1\", \"FPR1\", \"MTSS1\", \"MFGE8\", \"MRC1\", \"JUN\", \"CXCL8\", \"HLA-DRB5\"]}]", + "notes": "Original-cohort retrospective analysis. No new generation, label inference or embedding fit. Literature targets are not automatically rare; rarity and identity require observed evidence. Ye: Fig. 5 proliferating pediatric monocytes. Imperial: Fig. 5 wave-2 steroid-associated monocytes; fine labels/timing still needed. [{\"cohort\": \"ye2026_sepsis\", \"citation\": \"Ye et al., Nature Immunology 2026\", \"doi\": \"10.1038/s41590-025-02345-x\", \"url\": \"https://www.nature.com/articles/s41590-025-02345-x\", \"figure\": \"Fig. 5a\\u2013f; validation Fig. 7f / Extended Data Fig. 9e\", \"population\": \"Mono_c06-CD14-MKI67\", \"kind\": \"Reported discrete proliferating monocyte subset; rarity must be measured locally\", \"setting\": \"Pediatric pulmonary sepsis (paper: within 48 h of onset); pediatric HC/ICU, abdominal sepsis and adults remain controls.\", \"markers\": [\"CD14\", \"MKI67\", \"STMN1\", \"MPO\", \"ELANE\"], \"negative_markers\": [], \"interpretation\": \"MKI67/STMN1/MPO/ELANE jointly support the published proliferative program in a CD14 monocyte context. PCLAF is not required by this original Fig. 5 definition.\", \"gate\": \"No fine-cluster labels found in 133 audited canonical obs schemas or the paired table. Marker co-detection is not a Mono_c06 cluster assignment.\", \"pdf\": \"/home/sj_server_1/PBISC/paper-release-20260928/reference-papers/paired-cohorts/Single-cell multi-omic landscape reveals anatomical-specific immune features in adult and pediatric sepsis.pdf\", \"all_paired_samples\": 85, \"identity_passed_samples\": 82, \"identity_passed_donors\": 82, \"legacy_cache_genes\": 73, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/ye2026_sepsis/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis/cells_obs.parquet\", \"observed_author_labels\": {\"\": 610713}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis/cells_panel6k.npz\", \"wave_counts\": {}, \"sex_by_group\": {}, \"available_cache_markers\": [\"MKI67\", \"MPO\", \"ELANE\"], \"missing_cache_markers\": [\"CD14\", \"STMN1\"]}, {\"cohort\": \"imperial_eskd\", \"citation\": \"Stephenson et al., Cell Genomics 2025\", \"doi\": \"10.1016/j.xgen.2025.100918\", \"url\": \"https://pmc.ncbi.nlm.nih.gov/articles/PMC12366655/\", \"figure\": \"Fig. 5A\\u2013G (especially 5C\\u2013E)\", \"population\": \"Dexamethasone-related monocytes\", \"kind\": \"Reported treatment-associated discrete subset; not automatically rare\", \"setting\": \"Wave 2 COVID-positive patients receiving glucocorticoids; align each blood visit to treatment start.\", \"markers\": [\"CD163\", \"ADAMTS2\", \"SLC1A3\", \"SESN1\", \"FPR1\", \"MTSS1\", \"MFGE8\", \"MRC1\"], \"negative_markers\": [\"JUN\", \"CXCL8\", \"HLA-DRA\", \"HLA-DRB5\"], \"interpretation\": \"The paper contrasts an anti-inflammatory/phagocytic program with lower inflammatory and antigen-presentation expression. Fig. 3 HLA-II suppression is a program, not a rare-cell identity.\", \"gate\": \"Paired metadata include wave 2, but the available author labels collapse these cells into broad monocytes. Treatment dates and fine labels must be joined before frequency recovery.\", \"pdf\": \"/home/sj_server_1/PBISC/paper-release-20260928/reference-papers/paired-cohorts/Stephenson2025_CellGenom_temporal_multiomics_COVID19_ESKD.pdf\", \"all_paired_samples\": 172, \"identity_passed_samples\": 158, \"identity_passed_donors\": 58, \"legacy_cache_genes\": 73, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/imperial_eskd/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/imperial_eskd/cells_obs.parquet\", \"observed_author_labels\": {\"T CD4 Helper\": 85235, \"NK\": 77242, \"T CD8 CTL\": 70657, \"CD14mono\": 53967, \"T CD4 Naive\": 49943, \"B_naive\": 33633, \"CD16mono\": 29056, \"T CD8 CM\": 19962, \"T CD8 Naive\": 18113, \"T CD4 CTL\": 17209, \"B_switched_mem\": 9223, \"T reg\": 9176, \"NK CD56 bright\": 8610, \"Int.mono\": 7136, \"DC2\": 5972, \"NKT\": 4647, \"DC3\": 4318, \"T g/d\": 3911, \"pDC\": 3034, \"B_non-switched_mem\": 2815, \"RBC\": 2604, \"Platelet\": 2576, \"B_ASC_IgA\": 2538, \"B_CD11c\": 2108, \"MAIT\": 1984, \"T CD8 EM\": 1860, \"B_ASC_dividing\": 1785, \"B_ASC_IgG\": 1182, \"HSPC\": 1004, \"DC1\": 754, \"B_ASC_IgM\": 499, \"T CD4 cycling\": 447, \"T CD8 EMRA\": 400, \"ASDC\": 231, \"ILC\": 147, \"B_ASC_other\": 36}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/imperial_eskd/cells_panel6k.npz\", \"wave_counts\": {\"w2\": 83, \"w1\": 75, \"w1|w2\": 14}, \"sex_by_group\": {}, \"available_cache_markers\": [\"CD163\", \"HLA-DRA\"], \"missing_cache_markers\": [\"ADAMTS2\", \"SLC1A3\", \"SESN1\", \"FPR1\", \"MTSS1\", \"MFGE8\", \"MRC1\", \"JUN\", \"CXCL8\", \"HLA-DRB5\"]}]", + "sources": "https://www.nature.com/articles/s41590-025-02345-x; https://pmc.ncbi.nlm.nih.gov/articles/PMC12366655/; https://pmc.ncbi.nlm.nih.gov/articles/PMC9271053/; https://pmc.ncbi.nlm.nih.gov/articles/PMC12049347/", + "deck_group": "Figure 5", + "section": "Figure 5", + "addition_key": "strengthening:rare_literature:0", + "parent_154_page": 51, + "parent_155_page": 52 + }, + { + "kind": "content", + "key": "rare_literature_ye_observed_paper_markers", + "role": "f5_original_paper_observed_gate", + "title": "Ye sepsis: published markers in paired scRNA", + "section_label": "Fig. 5 \u00b7 Source-defined populations", + "layout": "full", + "keywords": [], + "figures": [ + "figures/strengthening_20261013/rare_literature/ye_observed_paper_markers.png" + ], + "palette": "biological", + "footnote": "Native-depth diagnostic, not Mono_c06 identity. MKI67/STMN1 mean: pulmonary 2.33%, abdominal 4.10%; pulmonary specificity is not reproduced.", + "footnote_full": "Original-cohort retrospective analysis. No new generation, label inference or embedding fit. Literature targets are not automatically rare; rarity and identity require observed evidence. Native-depth diagnostic, not Mono_c06 identity. MKI67/STMN1 mean: pulmonary 2.33%, abdominal 4.10%; pulmonary specificity is not reproduced. {\"n_cells\": 610713, \"input_genes\": 6000, \"retained_genes\": 5, \"csr_entries_streamed\": 255556515, \"max_entries_per_block\": 1000000, \"no_full_sparse_or_dense_matrix_loaded\": true, \"identity_passed_samples\": 82, \"analyzed_samples\": 82, \"analyzed_monocytes\": 111269, \"all_five_genes_measured_in_all_source_samples\": true, \"labels_npz_rows_exact\": true, \"author_fine_labels_present\": false, \"source_sha256\": \"c814b53d04ffdbe8e257f1fb35df80e7945479db95a924774522b9a967745280\", \"limitations\": [\"Native depth; marker dropout and depth differ between donors.\", \"Parent uses the frozen AIFI label; published Mono_c06 fine labels are absent.\", \"Co-detection patterns are diagnostics, not author-cluster calls.\", \"This is reuse of the original cohort, not independent replication.\"], \"no_new_inference\": true, \"no_embedding_fit\": true, \"observed_pattern_comparison\": {\"pattern\": \"MKI67+STMN1\", \"pediatric_pulmonary_mean_percent\": 2.331039105701644, \"pediatric_abdominal_mean_percent\": 4.103247268955951, \"pulmonary_specificity_reproduced\": false}, \"author_label_inventory\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/strengthening_20261013/rare_literature/ye_author_label_inventory.json\"}", + "notes": "Original-cohort retrospective analysis. No new generation, label inference or embedding fit. Literature targets are not automatically rare; rarity and identity require observed evidence. Native-depth diagnostic, not Mono_c06 identity. MKI67/STMN1 mean: pulmonary 2.33%, abdominal 4.10%; pulmonary specificity is not reproduced. {\"n_cells\": 610713, \"input_genes\": 6000, \"retained_genes\": 5, \"csr_entries_streamed\": 255556515, \"max_entries_per_block\": 1000000, \"no_full_sparse_or_dense_matrix_loaded\": true, \"identity_passed_samples\": 82, \"analyzed_samples\": 82, \"analyzed_monocytes\": 111269, \"all_five_genes_measured_in_all_source_samples\": true, \"labels_npz_rows_exact\": true, \"author_fine_labels_present\": false, \"source_sha256\": \"c814b53d04ffdbe8e257f1fb35df80e7945479db95a924774522b9a967745280\", \"limitations\": [\"Native depth; marker dropout and depth differ between donors.\", \"Parent uses the frozen AIFI label; published Mono_c06 fine labels are absent.\", \"Co-detection patterns are diagnostics, not author-cluster calls.\", \"This is reuse of the original cohort, not independent replication.\"], \"no_new_inference\": true, \"no_embedding_fit\": true, \"observed_pattern_comparison\": {\"pattern\": \"MKI67+STMN1\", \"pediatric_pulmonary_mean_percent\": 2.331039105701644, \"pediatric_abdominal_mean_percent\": 4.103247268955951, \"pulmonary_specificity_reproduced\": false}, \"author_label_inventory\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/strengthening_20261013/rare_literature/ye_author_label_inventory.json\"}", + "sources": "https://www.nature.com/articles/s41590-025-02345-x; https://pmc.ncbi.nlm.nih.gov/articles/PMC12366655/; https://pmc.ncbi.nlm.nih.gov/articles/PMC9271053/; https://pmc.ncbi.nlm.nih.gov/articles/PMC12049347/", + "deck_group": "Figure 5", + "section": "Figure 5", + "addition_key": "strengthening:rare_literature:1", + "parent_154_page": 52, + "parent_155_page": 53 + }, + { + "kind": "content", + "key": "rare_literature_published_program_boundaries", + "role": "f5_original_paper_observed_gate", + "title": "Transcriptional programs and rare-population identity", + "section_label": "Fig. 5 \u00b7 Source-defined populations", + "layout": "full", + "keywords": [], + "figures": [ + "figures/strengthening_20261013/rare_literature/published_program_boundaries.png" + ], + "palette": "biological", + "footnote": "Giroux: pre-seroconversion RNA/regulon states and sex imbalance. Ji: common-cell-type DE/TCR changes; no paper-defined rare identity established.", + "footnote_full": "Original-cohort retrospective analysis. No new generation, label inference or embedding fit. Literature targets are not automatically rare; rarity and identity require observed evidence. Giroux: pre-seroconversion RNA/regulon states and sex imbalance. Ji: common-cell-type DE/TCR changes; no paper-defined rare identity established. [{\"cohort\": \"giroux2022\", \"citation\": \"Giroux et al., Scientific Reports 2022\", \"doi\": \"10.1038/s41598-022-15668-8\", \"url\": \"https://pmc.ncbi.nlm.nih.gov/articles/PMC9271053/\", \"figure\": \"Fig. 5A\\u2013F; Supplementary Tables S8/S9\", \"population\": \"IgG-negative CD14 monocyte clusters 2/4 (mild) and 7 (moderate)\", \"kind\": \"Transcriptomic/regulon states with epigenetic corroboration; rarity not established\", \"setting\": \"Before IgG seroconversion, mild versus moderate outpatients; local severe label must be checked against the paper\\u2019s moderate definition.\", \"markers\": [\"CD14\", \"IRF7\", \"IRF1\", \"STAT1\", \"CEBPB\", \"KLF3\", \"CCR2\"], \"negative_markers\": [], \"interpretation\": \"IRF7/IRF1/STAT1 versus CEBPB/KLF3 are regulatory-program findings, not an exact RNA-only positivity gate. scRNA clusters connect to scATAC states.\", \"gate\": \"No author fine labels locally; IRF1/KLF3 are outside the saved 6k panel. Align serostatus, longitudinal donors and marked sex imbalance; no exact rare-cluster claim.\", \"xml\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/strengthening_20261013/rare_literature/Giroux2022_PMC9271053.xml\", \"all_paired_samples\": 26, \"identity_passed_samples\": 24, \"identity_passed_donors\": 9, \"legacy_cache_genes\": 73, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/giroux2022/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/giroux2022/cells_obs.parquet\", \"observed_author_labels\": {\"\": 276615}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/giroux2022/cells_panel6k.npz\", \"wave_counts\": {}, \"sex_by_group\": {\"FEMALE\": {\"mild\": 0, \"severe\": 11}, \"MALE\": {\"mild\": 13, \"severe\": 2}}, \"available_cache_markers\": [], \"missing_cache_markers\": [\"CD14\", \"IRF7\", \"IRF1\", \"STAT1\", \"CEBPB\", \"KLF3\", \"CCR2\"]}, {\"cohort\": \"ji2025_chili\", \"citation\": \"Ji et al., Cancer Immunology, Immunotherapy 2025\", \"doi\": \"10.1007/s00262-025-04033-z\", \"url\": \"https://pmc.ncbi.nlm.nih.gov/articles/PMC12049347/\", \"figure\": \"Fig. 2 (TCR clonality); Fig. 4A/B and Table 2 (cell-type DE/programs)\", \"population\": \"CD8 TEM, CD8 naive, CD4 TCM and CD16 monocyte programs\", \"kind\": \"Common cell-type expression changes; no specific rare subset established in the original paper\", \"setting\": \"Liver-injury timepoint TOL versus pretreatment BC / control AC; FU1 is follow-up and must not be silently pooled with TOL.\", \"markers\": [], \"negative_markers\": [], \"interpretation\": \"Original HuBMAP-type labels survive locally. DEG counts and pathways can be reproduced as Figure 4 evidence; TCR clonality cannot be recovered from generated 6k expression alone.\", \"gate\": \"Do not substitute a generic cycling/plasmablast rule for the paper\\u2019s main finding. Select a genuinely paper-defined rare target only if supplementary evidence supports one.\", \"pdf\": \"/home/sj_server_1/PBISC/paper-release-20260928/reference-papers/paired-cohorts/Ji2025_CancerImmunolImmunother_checkpoint_inhibitor_liver_injury.pdf\", \"all_paired_samples\": 40, \"identity_passed_samples\": 33, \"identity_passed_donors\": 20, \"legacy_cache_genes\": 72, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/ji2025_chili/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ji2025_chili/cells_obs.parquet\", \"observed_author_labels\": {\"CD4 TCM\": 48811, \"CD14 Mono\": 38304, \"NK\": 19796, \"CD8 TEM\": 15575, \"CD16 Mono\": 6512, \"B naive\": 4822, \"CD8 Naive\": 2736, \"CD4 Naive\": 2559, \"B intermediate\": 2372, \"Platelet\": 2165, \"Treg\": 2001, \"cDC2\": 1259, \"CD8 TCM\": 1134, \"NK_CD56bright\": 1010, \"B memory\": 878, \"Plasmablast\": 852, \"MAIT\": 801, \"HSPC\": 724, \"CD4 TEM\": 600, \"pDC\": 582, \"NK Proliferating\": 358, \"dnT\": 314, \"Eryth\": 230, \"CD4 CTL\": 80, \"ILC\": 68, \"cDC1\": 55, \"ASDC\": 50, \"CD8 Proliferating\": 28, \"gdT\": 14, \"CD4 Proliferating\": 8}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ji2025_chili/cells_panel6k.npz\", \"wave_counts\": {}, \"sex_by_group\": {}, \"available_cache_markers\": [], \"missing_cache_markers\": []}]", + "notes": "Original-cohort retrospective analysis. No new generation, label inference or embedding fit. Literature targets are not automatically rare; rarity and identity require observed evidence. Giroux: pre-seroconversion RNA/regulon states and sex imbalance. Ji: common-cell-type DE/TCR changes; no paper-defined rare identity established. [{\"cohort\": \"giroux2022\", \"citation\": \"Giroux et al., Scientific Reports 2022\", \"doi\": \"10.1038/s41598-022-15668-8\", \"url\": \"https://pmc.ncbi.nlm.nih.gov/articles/PMC9271053/\", \"figure\": \"Fig. 5A\\u2013F; Supplementary Tables S8/S9\", \"population\": \"IgG-negative CD14 monocyte clusters 2/4 (mild) and 7 (moderate)\", \"kind\": \"Transcriptomic/regulon states with epigenetic corroboration; rarity not established\", \"setting\": \"Before IgG seroconversion, mild versus moderate outpatients; local severe label must be checked against the paper\\u2019s moderate definition.\", \"markers\": [\"CD14\", \"IRF7\", \"IRF1\", \"STAT1\", \"CEBPB\", \"KLF3\", \"CCR2\"], \"negative_markers\": [], \"interpretation\": \"IRF7/IRF1/STAT1 versus CEBPB/KLF3 are regulatory-program findings, not an exact RNA-only positivity gate. scRNA clusters connect to scATAC states.\", \"gate\": \"No author fine labels locally; IRF1/KLF3 are outside the saved 6k panel. Align serostatus, longitudinal donors and marked sex imbalance; no exact rare-cluster claim.\", \"xml\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/strengthening_20261013/rare_literature/Giroux2022_PMC9271053.xml\", \"all_paired_samples\": 26, \"identity_passed_samples\": 24, \"identity_passed_donors\": 9, \"legacy_cache_genes\": 73, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/giroux2022/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/giroux2022/cells_obs.parquet\", \"observed_author_labels\": {\"\": 276615}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/giroux2022/cells_panel6k.npz\", \"wave_counts\": {}, \"sex_by_group\": {\"FEMALE\": {\"mild\": 0, \"severe\": 11}, \"MALE\": {\"mild\": 13, \"severe\": 2}}, \"available_cache_markers\": [], \"missing_cache_markers\": [\"CD14\", \"IRF7\", \"IRF1\", \"STAT1\", \"CEBPB\", \"KLF3\", \"CCR2\"]}, {\"cohort\": \"ji2025_chili\", \"citation\": \"Ji et al., Cancer Immunology, Immunotherapy 2025\", \"doi\": \"10.1007/s00262-025-04033-z\", \"url\": \"https://pmc.ncbi.nlm.nih.gov/articles/PMC12049347/\", \"figure\": \"Fig. 2 (TCR clonality); Fig. 4A/B and Table 2 (cell-type DE/programs)\", \"population\": \"CD8 TEM, CD8 naive, CD4 TCM and CD16 monocyte programs\", \"kind\": \"Common cell-type expression changes; no specific rare subset established in the original paper\", \"setting\": \"Liver-injury timepoint TOL versus pretreatment BC / control AC; FU1 is follow-up and must not be silently pooled with TOL.\", \"markers\": [], \"negative_markers\": [], \"interpretation\": \"Original HuBMAP-type labels survive locally. DEG counts and pathways can be reproduced as Figure 4 evidence; TCR clonality cannot be recovered from generated 6k expression alone.\", \"gate\": \"Do not substitute a generic cycling/plasmablast rule for the paper\\u2019s main finding. Select a genuinely paper-defined rare target only if supplementary evidence supports one.\", \"pdf\": \"/home/sj_server_1/PBISC/paper-release-20260928/reference-papers/paired-cohorts/Ji2025_CancerImmunolImmunother_checkpoint_inhibitor_liver_injury.pdf\", \"all_paired_samples\": 40, \"identity_passed_samples\": 33, \"identity_passed_donors\": 20, \"legacy_cache_genes\": 72, \"legacy_cache_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/rare_states_paired_20261008/work/ji2025_chili/draw0/lora5.npz\", \"observed_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ji2025_chili/cells_obs.parquet\", \"observed_author_labels\": {\"CD4 TCM\": 48811, \"CD14 Mono\": 38304, \"NK\": 19796, \"CD8 TEM\": 15575, \"CD16 Mono\": 6512, \"B naive\": 4822, \"CD8 Naive\": 2736, \"CD4 Naive\": 2559, \"B intermediate\": 2372, \"Platelet\": 2165, \"Treg\": 2001, \"cDC2\": 1259, \"CD8 TCM\": 1134, \"NK_CD56bright\": 1010, \"B memory\": 878, \"Plasmablast\": 852, \"MAIT\": 801, \"HSPC\": 724, \"CD4 TEM\": 600, \"pDC\": 582, \"NK Proliferating\": 358, \"dnT\": 314, \"Eryth\": 230, \"CD4 CTL\": 80, \"ILC\": 68, \"cDC1\": 55, \"ASDC\": 50, \"CD8 Proliferating\": 28, \"gdT\": 14, \"CD4 Proliferating\": 8}, \"observed_full_panel_path\": \"/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ji2025_chili/cells_panel6k.npz\", \"wave_counts\": {}, \"sex_by_group\": {}, \"available_cache_markers\": [], \"missing_cache_markers\": []}]", + "sources": "https://www.nature.com/articles/s41590-025-02345-x; https://pmc.ncbi.nlm.nih.gov/articles/PMC12366655/; https://pmc.ncbi.nlm.nih.gov/articles/PMC9271053/; https://pmc.ncbi.nlm.nih.gov/articles/PMC12049347/", + "deck_group": "Figure 5", + "section": "Figure 5", + "addition_key": "strengthening:rare_literature:2", + "parent_154_page": 53, + "parent_155_page": 54 + }, + { + "kind": "divider", + "title": "Figure 6. Published-finding reproduction", + "original_slide": 53, + "original_title": "Figure 6. Discovery from public bulk-only data", + "original_footnote": "", + "source_page_131": 102, + "deck_group": "Figure 6", + "section": "Figure 6", + "previous_section_label": null, + "section_label": "Figure 6", + "parent_page": 32, + "parent_enriched_page": 42, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 43, + "parent_154_page": 54, + "parent_155_page": 55 + }, + { + "kind": "plan_card", + "section": "Figure 6", + "section_label": "Figure 6 plan", + "title": "Figure 6: published-finding reproduction", + "footnote": "Reproduction of an existing finding is the objective. Historical exploratory searches are retained in the archive section.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "Reproduction of an existing finding is the objective. Historical exploratory searches are retained in the archive section.\nAuthor-defined Figure scope, 13 October 2026. Supersedes the former planning card.", + "original_slide": 54, + "original_title": "Figure 6 plan: Discovery from public bulk-only data", + "original_footnote": "New findings from LoRA cells of bulk-only sepsis and COVID dialysis samples (K = 1,000 and 4,000). 0 of 2 panels ready.", + "source_page_131": 103, + "deck_group": "Figure 6", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_fig6.png" + ], + "rows": [ + [ + "6a", + "Define paired adaptation and bulk-only target samples", + "Planned" + ], + [ + "6b", + "Fix one published finding and its exact comparison", + "Planned" + ], + [ + "6c", + "Report the adapted-model bulk-only reproduction", + "Result pending" + ], + [ + "6d", + "Independent support and raw-bulk comparison", + "Result pending" + ] + ], + "parent_page": 33, + "parent_enriched_page": 43, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 44, + "parent_154_page": 55, + "parent_155_page": 56 + }, + { + "kind": "evidence_note", + "title": "Figure 6: what the reproduction must establish", + "section_label": "Figure 6", + "rows": [ + [ + "Adaptation and target samples", + "Record paired supports, bulk-only targets and donor/visit overlap." + ], + [ + "Published endpoint", + "Fix the paper finding, cell population, severity/time definition and evaluation before a new analysis." + ], + [ + "Result and evidence", + "Show the reproduction result and comparison with the bulk signal; distinguish retrospective replay from independent confirmation." + ] + ], + "after_original": 55, + "addition_key": "guide:figure6", + "footnote": "Result pending. Historical candidate searches neither prove nor disprove a prespecified published-finding reproduction.", + "footnote_full": "Result pending. Historical candidate searches neither prove nor disprove a prespecified published-finding reproduction.", + "source_page_131": 107, + "deck_group": "Figure 6", + "section": "Figure 6", + "previous_section_label": "Figure 6", + "parent_page": 34, + "parent_enriched_page": 44, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 45, + "parent_154_page": 56, + "parent_155_page": 57 + }, + { + "kind": "divider", + "title": "Supplementary evidence", + "original_slide": 56, + "original_title": "Supplementary", + "original_footnote": "", + "source_page_131": 108, + "deck_group": "Supplementary", + "section": "Supplementary", + "previous_section_label": null, + "section_label": "Supplementary", + "parent_page": 35, + "parent_enriched_page": 45, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 46, + "parent_154_page": 57, + "parent_155_page": 58 + }, + { + "kind": "plan_card", + "section": "Supplementary", + "section_label": "Supplementary", + "title": "Supplementary evidence map", + "footnote": "Width-validation curves moved from supplementary into Figure 2; the prior figure remains in the preserved source deck.", + "sources": "data/figure_assignment_20261013/AUTHOR_SCOPE.md; data/figure_order_fixed_20261013/page_map.csv", + "footnote_full": "Width-validation curves moved from supplementary into Figure 2; the prior figure remains in the preserved source deck.", + "original_slide": 57, + "original_title": "Supplementary plan", + "original_footnote": "Bulk deconvolution methods appear here only: PBISC generates single cells, a different task.", + "source_page_131": 109, + "deck_group": "Supplementary", + "previous_section_label": "Original roadmap", + "superseded_figures": [ + "figures/figplan_supp.png" + ], + "rows": [ + [ + "S01-S02", + "Methods and additional SLE controls", + "pp. 62-72" + ], + [ + "S03-S06", + "Input diagnostics, evaluation, healthy quality and budgets", + "pp. 73-105" + ], + [ + "S07-S09", + "Cohorts, disease quality, corrected DEG and diagnostics", + "pp. 106-130" + ], + [ + "S10-S11", + "Paired rare-state quantities and external screens", + "pp. 131-141" + ], + [ + "S12-S14", + "Transfer, scratch, readout and bulk-only cohort shift", + "pp. 142-148" + ], + [ + "Archive", + "Historical scoring and earlier bulk-only plans", + "pp. 149-154" + ] + ], + "parent_page": 36, + "parent_enriched_page": 46, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 47, + "parent_154_page": 58, + "parent_155_page": 59 + }, + { + "kind": "content", + "section": "Supp S01", + "section_label": "Supp. S01", + "text_source": { + "file": "data/figure_plan_slide_text.json", + "index": 9 + }, + "title": "Corpus reference: assay and source composition", + "layout": "full", + "figures": [ + "figures/fig1d_corpus.png" + ], + "keywords": [], + "footnote": "Original corpus overview. Healthy 88.5% is a cell share; the revised condition detail is in Figure 1.", + "sources": "tools/figure_plan.py; manuscript figures/fig01_corpus_nature.png, source_data/corpus/", + "footnote_full": "Original corpus overview. Healthy 88.5% is a cell share; the revised condition detail is in Figure 1.\nCurrent scope: Corpus labels describe the data; they are not model inputs. Dataset composition is not sampler exposure.\n\nOriginal slide 9: Source: paper-release-20260928/manuscript/source_data/corpus/ (build_corpus_tables.py) and figures/fig01_corpus_nature.png (build_figure1_corpus.py).", + "numbers": [ + { + "value": "41,395,184 cells; 22 datasets; 3,199 donors; healthy 88.5%; in vitro 23.6%", + "source": "manuscript/source_data/corpus/corpus_summary.json" + } + ], + "original_slide": 9, + "original_title": "Pretraining corpus", + "original_footnote": "Detailed version: Supplementary Fig. S1 of the manuscript.", + "figures_before_palette": [ + "figures/fig1d_corpus.png" + ], + "source_page_131": 11, + "deck_group": "Supp S01", + "previous_section_label": "Fig. 1d", + "parent_page": 8, + "parent_enriched_page": 8, + "figures_before_final_palette": [ + "figures/fig1d_corpus.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 8, + "parent_154_page": 59, + "parent_155_page": 60 + }, + { + "kind": "content", + "section": "Supp S01", + "section_label": "Supp. S01", + "title": "Backbone: SDD \u2192 PBISC-Diffusion", + "layout": "diagram", + "figures": [ + "figures/final_palette_20261013/native/sec02_backbone_compare.png" + ], + "keywords": [ + "Kept from SDD: value + bit, skip-MLP", + "Changed: bulk condition, masked loss, Bernoulli read-out" + ], + "footnote": "SDD: Ostheimer et al., arXiv 2502.02448. scDiffusion: Luo et al., Bioinformatics 2024.", + "sources": "tools/sec02_backbone.py; data/sec02_backbone_compare.csv, data/sec02_slide_text.csv, data/refs/", + "footnote_full": "Original slide 6: Luo et al., Bioinformatics 2024; Ostheimer et al., arXiv 2502.02448 (code 1d5423c). PBISC: final model f01 6k 1x, part 16. \u03c3 = sigmoid; Bernoulli is the final read-out; the logit > 0 cut is also reported.", + "refs": [ + "Ostheimer 2025, arXiv", + "Luo 2024, Bioinformatics" + ], + "refs_full": [ + "Ostheimer et al. Sparse Data Diffusion for scientific simulations in biology and physics. arXiv:2502.02448 (2025)", + "Luo et al. scDiffusion: conditional generation of high-quality single-cell data using diffusion model. Bioinformatics 40(9):btae518 (2024)" + ], + "original_slide": 6, + "original_title": "Backbone: SDD \u2192 PBISC-Diffusion", + "original_footnote": "SDD: Ostheimer et al., arXiv 2502.02448. scDiffusion: Luo et al., Bioinformatics 2024.", + "figures_before_palette": [ + "figures/sec02_backbone_compare.png" + ], + "source_page_131": 8, + "deck_group": "Supp S01", + "previous_section_label": "Fig. 1a", + "parent_page": 37, + "parent_enriched_page": 47, + "figures_before_final_palette": [ + "figures/sec02_backbone_compare.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 48, + "parent_154_page": 60, + "parent_155_page": 61 + }, + { + "kind": "content", + "section": "Supp S01", + "section_label": "Supp. S01", + "title": "Bulk conditioning", + "layout": "diagram", + "figures": [ + "figures/sec02_bulk_conditioning.png" + ], + "keywords": [ + "Same bulk condition for every generated cell", + "Final: no cell types, proportions, sum constraint", + "18,006 \u2192 512 context into every block" + ], + "footnote": "Final 6k model. Recipe size K conditions the bulk; the number of generated cells is chosen separately.", + "sources": "tools/sec02_backbone.py; data/sec02_bulk_conditioning.csv, data/sec02_slide_text.csv", + "footnote_full": "Current scope: Final 6k model. Recipe size K conditions the bulk; the number of generated cells is chosen separately.\n\nOriginal slide 8: Final model f01 6k 1x, part 16. Bar heights schematic.", + "original_slide": 8, + "original_title": "Bulk conditioning", + "original_footnote": "Final model f01 6k 1x. Bar heights schematic.", + "figures_before_palette": [ + "figures/sec02_bulk_conditioning.png" + ], + "source_page_131": 10, + "deck_group": "Supp S01", + "previous_section_label": "Fig. 1c", + "parent_page": 38, + "parent_enriched_page": 48, + "figures_before_final_palette": [ + "figures/sec02_bulk_conditioning.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 49, + "parent_154_page": 61, + "parent_155_page": 62 + }, + { + "kind": "content", + "section": "Supp S01", + "section_label": "Supp. S01", + "title": "Denoiser: joint value + support diffusion", + "layout": "diagram", + "figures": [ + "figures/sec01_fig1a_denoiser.png" + ], + "keywords": [ + "6,000 genes \u00b7 54.2M params \u00b7 Bernoulli support draw" + ], + "footnote": "Fig. 1a redraw. Final model f01 6k 1x, epoch 16, EMA; 50 DDIM steps.", + "footnote_full": "Original slide 14: Deck redraw of manuscript Fig. 1a (larger labels, legend dropped). Purple dashed = soft self-conditioning feedback; gold dashed = context to all blocks. Final model f01_6k_1x_seed20260627, checkpoint_epoch_016, EMA; 7 blocks 1024-1024-512-256-512-1024-1024 (effective_config.json, RUN_STATUS.json). Sampling: 50 DDIM steps, \u03b7 = 0, drawn support read-out, keep gene if u < \u03c3(logit) (methods.tex l.58, l.99). 16 epochs = 16 disjoint 8,192-recipe slots = one pass over the 131,072-recipe pool (manuscript: part 16).", + "sources": "tools/sec01_overview.py; data/sec01_model_numbers.csv, data/sec01_fig1_label_pt.csv", + "refs": [ + "Ostheimer 2025, arXiv" + ], + "refs_full": [ + "Ostheimer et al. Sparse Data Diffusion for scientific simulations in biology and physics. arXiv:2502.02448 (2025)" + ], + "original_slide": 14, + "original_title": "Denoiser: joint value + support diffusion", + "original_footnote": "Fig. 1a redraw. Final model f01 6k 1x, epoch 16, EMA; 50 DDIM steps.", + "figures_before_palette": [ + "figures/sec01_fig1a_denoiser.png" + ], + "source_page_131": 17, + "deck_group": "Supp S01", + "previous_section_label": "Fig. 1 details", + "parent_page": 39, + "parent_enriched_page": 49, + "figures_before_final_palette": [ + "figures/sec01_fig1a_denoiser.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 50, + "parent_154_page": 62, + "parent_155_page": 63 + }, + { + "kind": "content", + "section": "Supp S01", + "section_label": "Supp. S01", + "title": "Fine-tuning with paired bulk + sc", + "layout": "diagram", + "figures": [ + "figures/final_palette_20261013/sec06_ft_method.png" + ], + "keywords": [ + "LoRA 2\u201314 min vs pretraining 15.9 h" + ], + "footnote": "Base: final model f01 6k 1x E16. Times from 40 LoRA runs on this host.", + "sources": "tools/sec06_finetune_method.py; data/sec06_ft_method.csv", + "footnote_full": "Original slide 20: Receipts: immicro/fold0/real_lora and ye2026_sepsis/fewshot_n5/draw0/lora RUN.json. Base: f01_6k_1x E16. Times: 40 local LoRA runs.", + "refs": [ + "Hu 2022, ICLR" + ], + "refs_full": [ + "Hu et al. LoRA: Low-Rank Adaptation of Large Language Models. ICLR 2022" + ], + "original_slide": 20, + "original_title": "Fine-tuning with paired bulk + sc", + "original_footnote": "Base: final model f01 6k 1x E16. Times from 40 LoRA runs on this host.", + "figures_before_palette": [ + "figures/sec06_ft_method.png" + ], + "source_page_131": 38, + "deck_group": "Supp S01", + "previous_section_label": "Fig. 2c", + "parent_page": 40, + "parent_enriched_page": 50, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec06_ft_method.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 51, + "parent_154_page": 63, + "parent_155_page": 64 + }, + { + "kind": "content", + "section": "Supp S01", + "section_label": "Supp. S01", + "title": "Evaluation design", + "layout": "full", + "figures": [ + "figures/sec06_eval_design.png" + ], + "keywords": [ + "Donor-level splits; test donors never trained", + "5-fold: other folds \u00b7 few: 5 donors", + "Giroux few-pair trains 12\u201314 of 24 pairs" + ], + "footnote": "Bars: Ye 2026 sepsis. QC-failed pairs excluded.", + "sources": "tools/sec06_finetune_method.py; data/sec06_eval_design.csv, data/sec06_eval_design_sepsis_roles.csv", + "footnote_full": "Original slide 21: Counts: data//folds.csv and fewshot_n5/DRAWS.json; QC-failed pairs (fold \u22121) excluded. Bars: Ye 2026 sepsis.", + "original_slide": 21, + "original_title": "Evaluation design", + "original_footnote": "Bars: Ye 2026 sepsis. QC-failed pairs excluded.", + "figures_before_palette": [ + "figures/sec06_eval_design.png" + ], + "source_page_131": 39, + "deck_group": "Supp S01", + "previous_section_label": "Fig. 2c", + "parent_page": 41, + "parent_enriched_page": 51, + "figures_before_final_palette": [ + "figures/sec06_eval_design.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 52, + "parent_154_page": 64, + "parent_155_page": 65 + }, + { + "kind": "content", + "section": "Supp S01", + "title": "How we score generated cells", + "layout": "full", + "figures": [ + "figures/figure_enrichment_20261013/labels_metrics_overview.png" + ], + "keywords": [ + "Cells, composition and disease signal, separately", + "Every metric next to its control or ceiling", + "Same donors and reference cells for every method" + ], + "footnote": "Definitions follow the cited papers; scorers: tools/score_ft.py, extra_metrics.py, contrast_extra.py.", + "sources": "tools/sec04_metrics_overview.py; data/sec04_metrics_overview.csv", + "refs_full": [ + "Kynk\u00e4\u00e4nniemi et al. Improved precision and recall metric for assessing generative models. NeurIPS 2019", + "Naeem et al. Reliable fidelity and diversity metrics for generative models. ICML 2020", + "Lopez-Paz & Oquab. Revisiting classifier two-sample tests. ICLR 2017", + "Gretton et al. A kernel two-sample test. JMLR 13:723-773 (2012)", + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1", + "Roohani, Huang & Leskovec. Predicting transcriptional outcomes of novel multigene perturbations with GEARS. Nat Biotechnol 42:927-935 (2024). doi:10.1038/s41587-023-01905-6", + "Vi\u00f1as Torn\u00e9 et al. Systema: a framework for evaluating genetic perturbation response prediction beyond systematic variation. Nat Biotechnol 44:1050-1059 (2026). doi:10.1038/s41587-025-02777-8", + "Li et al. A systematic comparison of single-cell perturbation response prediction models. Sci Adv 12:eaed3414 (2026). doi:10.1126/sciadv.aed3414", + "Squair et al. Confronting false discoveries in single-cell differential expression. Nat Commun 12:5692 (2021). doi:10.1038/s41467-021-25960-2", + "Law, Chen, Shi & Smyth. voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol 15:R29 (2014)", + "Irizarry et al. Multiple-laboratory comparison of microarray platforms. Nat Methods 2:345-350 (2005). doi:10.1038/nmeth756" + ], + "section_label": "Supp. S01", + "original_slide": 22, + "original_title": "How we score generated cells", + "original_footnote": "Definitions follow the cited papers; scorers: tools/score_ft.py, extra_metrics.py, contrast_extra.py.", + "footnote_full": "Original slide 22: Definitions follow the cited papers; scorers: tools/score_ft.py, extra_metrics.py, contrast_extra.py.", + "figures_before_palette": [ + "figures/sec04_metrics_overview.png" + ], + "source_page_131": 41, + "deck_group": "Supp S01", + "previous_section_label": "Fig. 2d", + "parent_page": 42, + "previous_figures": [ + "figures/sec04_metrics_overview.png" + ], + "parent_enriched_page": 52, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/labels_metrics_overview.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 53, + "parent_154_page": 65, + "parent_155_page": 66 + }, + { + "kind": "content", + "key": "healthy_sle_conditioning_interventions", + "after_original": 14, + "title": "Does the model follow its bulk conditioning?", + "section_label": "Supp. S02", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sle_conditioning_interventions.png" + ], + "keywords": [], + "footnote": "Existing SLE pseudobulk controls: same noise, 42 donors, 2,279 genes. This is a conditioning intervention, not real-bulk validation.", + "footnote_full": "SLE development benchmark: 42 donors, 2,279 genes, f01 6k E16 EMA, 1,500 generated cells/donor, 50 DDIM steps, seed20260913. Only conditioning differs; noise is matched. These are pseudobulk inputs, not an independent paired real-bulk validation. Donor-source caption uses saved uncentered nearest-match counts; centered flip retrieval is 27/42, permutation42/42. Displayed benchmark metrics use the common six-type gate. The pre-registered condition-use criterion and earlier-model tolerance context are in source RESULTS.md.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/conditioning_2279/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/conditioning_2279/paired_cells.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/metrics_2279/summary_tidy.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/conditioning_2279/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/conditioning_2279/paired_cells.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/metrics_2279/summary_tidy.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/conditioning_2279/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/conditioning_2279/paired_cells.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle/metrics_2279/summary_tidy.csv", + "palette": "native", + "addition_key": "healthy_sle_conditioning_interventions", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/sle_conditioning_interventions.png" + ], + "source_page_131": 18, + "deck_group": "Supp S02", + "section": "Supp S02", + "previous_section_label": "Control evidence", + "parent_page": 44, + "parent_enriched_page": 53, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/sle_conditioning_interventions.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 54, + "parent_154_page": 66, + "parent_155_page": 67 + }, + { + "kind": "content", + "section": "Supp S02", + "section_label": "Supp. S02", + "title": "SLE benchmark: PBISC vs scSemiProfiler", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec04_sle_benchmark.png" + ], + "keywords": [ + "42 held-out donors, sc-pseudobulk input", + "Beats scSemiProfiler on all four metrics", + "Beyond bulk: partial r 0.21 vs 0.17" + ], + "footnote": "Completed pseudobulk benchmark; separate from external paired real-bulk validation. PBISC five seeds, scSemi one run.", + "sources": "tools/sec04_sle_benchmark.py; data/sec04_sle_benchmark.csv, data/sec04_sle_numbers.csv", + "footnote_full": "Current scope: Completed pseudobulk benchmark; separate from external paired real-bulk validation. PBISC five seeds, scSemi one run.\n\nOriginal slide 42: GSE135779, 42 held-out donors (30 SLE / 12 HC). PBISC f01 6k E16, 5 seeds (mean \u00b1 SD); baselines 1 draw. scSemiProfiler also gets real sc of 14 in-cohort donors (1,500 cells each). Partial r: both logFCs residualised on the input bulk's logFC (the cell-type signal the bulk does not carry), 1,006 genes: PBISC 0.208 \u00b1 0.012, scSemiProfiler 0.167.", + "refs": [ + "Wang 2024, Nat Commun", + "Xu 2021, Mol Syst Biol" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1", + "Xu et al. Probabilistic harmonization and annotation of single-cell transcriptomics data with deep generative models. Mol Syst Biol 17:e9620 (2021)" + ], + "original_slide": 42, + "original_title": "SLE benchmark: PBISC vs scSemiProfiler", + "original_footnote": "GSE135779, 42 held-out donors (30 SLE / 12 HC). PBISC: 5 seeds, mean \u00b1 SD; scSemiProfiler: 1 run.", + "figures_before_palette": [ + "figures/sec04_sle_benchmark.png" + ], + "source_page_131": 72, + "deck_group": "Supp S02", + "previous_section_label": "Fig. 4f", + "parent_page": 45, + "parent_enriched_page": 54, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec04_sle_benchmark.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 55, + "parent_154_page": 67, + "parent_155_page": 68 + }, + { + "kind": "content", + "section": "Supp S02", + "section_label": "Supp. S02", + "title": "SLE benchmark: per cell type", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec04_sle_per_celltype.png" + ], + "keywords": [ + "Beats scSemiProfiler in T, B, Mono, NK", + "Loses in Other; DC r ties, fewer DEGs", + "DC: ~15 generated cells per donor" + ], + "footnote": "Same 42 donors, 1,006 genes. PBISC: 5 seeds, mean \u00b1 SD; scSemiProfiler: 1 run. Other: top-50 of 12.", + "say": "PBISC's DC top list in the base seed has only 13 genes, because few DC cells are generated.", + "sources": "tools/sec04_sle_benchmark.py; data/sec04_sle_per_celltype.csv, data/sec04_sle_numbers.csv", + "footnote_full": "Original slide 43: Same 42 donors, 1,006 genes. PBISC 5 seeds (mean \u00b1 SD), baselines 1 draw. Tick %: share of 251,179 real cells. Other top-50 is out of 12 truth genes.", + "refs": [ + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "original_slide": 43, + "original_title": "SLE benchmark: per cell type", + "original_footnote": "Same 42 donors, 1,006 genes. PBISC: 5 seeds, mean \u00b1 SD; scSemiProfiler: 1 run. Other: top-50 of 12.", + "figures_before_palette": [ + "figures/sec04_sle_per_celltype.png" + ], + "source_page_131": 73, + "deck_group": "Supp S02", + "previous_section_label": "Fig. 4f", + "parent_page": 46, + "parent_enriched_page": 55, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec04_sle_per_celltype.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 56, + "parent_154_page": 68, + "parent_155_page": 69 + }, + { + "kind": "content", + "key": "healthy_sle_disjoint_truth_control", + "after_original": 43, + "title": "Disjoint-cell controls test the bulk-broadcast baseline", + "section_label": "Supp. S02", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/native/sle_disjoint_truth_control.png" + ], + "keywords": [], + "footnote": "SLE pseudobulk benchmark; five existing half-cell splits. Error bars are saved split SDs. No new cell splitting or scoring.", + "footnote_full": "Existing five random half-A/half-B splits of the same 42 held-out SLE donors. Bulk A is formed from cells disjoint from per-type half-B truth. The full-bulk control includes B. The observed type-specific half-A predictor is a split-half reference. Bars and SDs are saved summary values; no split, expression analysis, or hypothesis test is rerun. PBISC values summarize its five existing sampling seeds.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle_broadcast/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle_broadcast/tables/splithalf_summary.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle_broadcast/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle_broadcast/tables/splithalf_summary.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle_broadcast/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/sle_broadcast/tables/splithalf_summary.csv", + "palette": "native", + "addition_key": "healthy_sle_disjoint_truth_control", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/sle_disjoint_truth_control.png" + ], + "source_page_131": 74, + "deck_group": "Supp S02", + "section": "Supp S02", + "previous_section_label": "Control evidence", + "parent_page": 47, + "parent_enriched_page": 56, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/sle_disjoint_truth_control.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 57, + "parent_154_page": 69, + "parent_155_page": 70 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_sle_broadcast", + "title": "SLE: what is added beyond the bulk broadcast?", + "section_label": "Supp. S02", + "figures": [ + "figures/final_palette_20261013/sec04_sle_broadcast.png" + ], + "keywords": [], + "footnote": "Existing SLE pseudobulk development benchmark; bulk broadcast is an explicit comparator. Separate from paired real-bulk validation.", + "footnote_full": "Existing figure omitted from the original 76-slide specification. Retain its gene axis, cell-type units and seed counts. This is a pseudobulk-conditioned benchmark, not a new external real-bulk result.\nSources: presentations/2026-09-30--paper-summary--r00/data/sec04_sle_broadcast.csv; presentations/2026-09-30--paper-summary--r00/tools/sec04_sle_benchmark.py\nEvidence source: presentations/2026-09-30--paper-summary--r00/data/sec04_sle_broadcast.csv; presentations/2026-09-30--paper-summary--r00/tools/sec04_sle_benchmark.py", + "sources": "presentations/2026-09-30--paper-summary--r00/data/sec04_sle_broadcast.csv; presentations/2026-09-30--paper-summary--r00/tools/sec04_sle_benchmark.py", + "notes": "Existing figure omitted from the original 76-slide specification. Retain its gene axis, cell-type units and seed counts. This is a pseudobulk-conditioned benchmark, not a new external real-bulk result.\nSources: presentations/2026-09-30--paper-summary--r00/data/sec04_sle_broadcast.csv; presentations/2026-09-30--paper-summary--r00/tools/sec04_sle_benchmark.py", + "after_original": 43, + "palette": "biological", + "new_experiment": false, + "addition_key": "disease_sle_broadcast", + "figures_before_palette": [ + "figures/sec04_sle_broadcast.png" + ], + "source_page_131": 75, + "deck_group": "Supp S02", + "section": "Supp S02", + "previous_section_label": "Development benchmark", + "parent_page": 48, + "parent_enriched_page": 57, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec04_sle_broadcast.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 58, + "parent_154_page": 70, + "parent_155_page": 71 + }, + { + "kind": "content", + "section": "Supp S03", + "section_label": "Supp. S03", + "title": "Same shift in three paired cohorts", + "layout": "full", + "figures": [ + "figures/sec05_shift_cohorts.png" + ], + "keywords": [ + "Shared shift in all 3 cohorts" + ], + "footnote": "Model inputs only, no generation. PCA fitted per cohort on sc pseudobulks.", + "sources": "tools/sec05_shift_cohorts.py; data/sec05_shift_cohorts.csv", + "footnote_full": "Original slide 18: Model inputs only, no generation. PCA is fitted on each cohort's sc pseudobulks (Arunachalam: 11 PCs, n = 12). D uses same-condition pairs only for Randolph and Arunachalam. Shift removal is a leave-one-out diagnostic.", + "original_slide": 18, + "original_title": "Same shift in three paired cohorts", + "original_footnote": "Model inputs only, no generation. PCA fitted per cohort on sc pseudobulks.", + "figures_before_palette": [ + "figures/sec05_shift_cohorts.png" + ], + "source_page_131": 22, + "deck_group": "Supp S03", + "previous_section_label": "Fig. 2a", + "parent_page": 49, + "parent_enriched_page": 58, + "figures_before_final_palette": [ + "figures/sec05_shift_cohorts.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 59, + "parent_154_page": 71, + "parent_155_page": 72 + }, + { + "kind": "content", + "key": "healthy_stored_real_reference_umap", + "after_original": 19, + "title": "Input corrections on one real-reference UMAP", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/expanded_master_20261013/healthy/stored_real_reference_umap.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nReplots only the stored interactive coordinates: 3,000 displayed cells per set. The reference PCA/UMAP was previously fitted to observed reference. UMAP overlap is illustrative, not a score. Cell-type colors are preserved; no new labels or embedding were computed. Shared display bounds cover all saved coordinates plus 2.5% padding; the narrower old illustrative extent is not reused.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/interactive/cells_5arm.json\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/interactive/cells_5arm.json", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/interactive/cells_5arm.json", + "palette": "preserved", + "addition_key": "healthy_stored_real_reference_umap", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/stored_real_reference_umap.png" + ], + "source_page_131": 24, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 50, + "parent_enriched_page": 59, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/stored_real_reference_umap.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 60, + "parent_154_page": 72, + "parent_155_page": 73 + }, + { + "kind": "content", + "key": "healthy_saved_density_difference", + "after_original": 19, + "title": "Where generated-cell density differs from real cells", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/expanded_master_20261013/healthy/saved_density_difference.png" + ], + "keywords": [], + "footnote": "Original high-resolution density map; zoom for panel labels. Quantitative colors preserved. Historical native-depth controls.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nThe original high-resolution PNG is copied byte-for-byte, with no crop, pixel recoloring, density-grid recomputation, or fresh embedding. Its diverging colors encode signed density differences, not PBISC/scSemiProfiler identities. Retained as a full-width working-master reference; zoom may be needed for small panel labels. UMAP-space density is a qualitative localization diagnostic, not a calibrated biological score. \n\nSource files:\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/figures/fig2_density_difference.png\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/figures/fig2_density_difference.png", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/figures/fig2_density_difference.png", + "palette": "preserved", + "addition_key": "healthy_saved_density_difference", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/saved_density_difference.png" + ], + "source_page_131": 25, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 51, + "parent_enriched_page": 60, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/saved_density_difference.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 61, + "parent_154_page": 73, + "parent_155_page": 74 + }, + { + "kind": "content", + "key": "healthy_input_control_composition", + "after_original": 19, + "title": "Input calibration changes cell-type composition", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/expanded_master_20261013/healthy/input_control_composition.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nSaved fractions from 150 cells/donor; displayed as percent. Color identifies biological cell type, not method. Do not interpret reference calibration's composition agreement as proof of within-type state recovery.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/b_composition_l1.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/b_composition_l1.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/b_composition_l1.csv", + "palette": "preserved", + "addition_key": "healthy_input_control_composition", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/input_control_composition.png" + ], + "source_page_131": 26, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 52, + "parent_enriched_page": 61, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/input_control_composition.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 62, + "parent_154_page": 74, + "parent_155_page": 75 + }, + { + "kind": "content", + "key": "healthy_stored_matched_trajectories", + "after_original": 19, + "title": "Same noise, different input: matched-cell shifts", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/stored_matched_trajectories.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nDisplay subset: first 12 stored pair IDs within each pseudobulk-predicted type, at most 72 pairs/arm, selected deterministically from the previously stored display sample. No embedding or labels recalculated. Lines connect generated samples sharing noise; they are not longitudinal biological transitions. Exact displayed pairs are saved. Shared display bounds cover all saved coordinates plus 2.5% padding.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/interactive/cells_5arm.json\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/interactive/cells_5arm.json", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/interactive/cells_5arm.json", + "palette": "preserved", + "addition_key": "healthy_stored_matched_trajectories", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/stored_matched_trajectories.png" + ], + "source_page_131": 27, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 53, + "parent_enriched_page": 62, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/stored_matched_trajectories.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 63, + "parent_154_page": 75, + "parent_155_page": 76 + }, + { + "kind": "content", + "key": "healthy_matched_similarity", + "after_original": 19, + "title": "Matched noise separates cell state from cell identity", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/matched_similarity.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nOriginal model-output comparisons over 160,500 paired generated cells. Expression-correlation whiskers show saved 25th\u201375th percentiles, not confidence intervals. All comparisons are against the pseudobulk-conditioned twin, not observed individual cells.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_matched_cells.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_matched_cells.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_matched_cells.csv", + "palette": "native", + "addition_key": "healthy_matched_similarity", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/matched_similarity.png" + ], + "source_page_131": 28, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 54, + "parent_enriched_page": 63, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/matched_similarity.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 64, + "parent_154_page": 76, + "parent_155_page": 77 + }, + { + "kind": "content", + "key": "healthy_label_transitions", + "after_original": 19, + "title": "Input changes can switch the generated cell label", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/expanded_master_20261013/healthy/label_transitions.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nExisting matched-cell transition counts, normalized within each pseudobulk-label row for display. Numbers are percentages. Frozen model labels, not proof of biological transdifferentiation. Greyscale represents a quantitative transition fraction.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_cal_counts.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_raw_counts.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_refcal_counts.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_tpm_counts.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_cal_counts.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_raw_counts.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_refcal_counts.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_tpm_counts.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_cal_counts.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_raw_counts.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_refcal_counts.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/c_label_transition_pseudo_to_tpm_counts.csv", + "palette": "preserved", + "addition_key": "healthy_label_transitions", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/label_transitions.png" + ], + "source_page_131": 29, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 55, + "parent_enriched_page": 64, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/label_transitions.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 65, + "parent_154_page": 77, + "parent_155_page": 78 + }, + { + "kind": "content", + "key": "healthy_nearest_real_by_type", + "after_original": 19, + "title": "Nearest real cells: distance and label agreement", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/nearest_real_by_type.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nSaved PCA50 nearest-reference-cell metrics from 150 cells/donor. Different cell types have different geometry; distances are not calibrated probabilities or donor-level effect estimates.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/f7_nearest_real_by_type.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/f7_nearest_real_by_type.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/f7_nearest_real_by_type.csv", + "palette": "native", + "addition_key": "healthy_nearest_real_by_type", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/nearest_real_by_type.png" + ], + "source_page_131": 30, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 56, + "parent_enriched_page": 65, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/nearest_real_by_type.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 66, + "parent_154_page": 78, + "parent_155_page": 79 + }, + { + "kind": "content", + "key": "healthy_within_type_centroids", + "after_original": 19, + "title": "Composition agreement does not ensure state agreement", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/labels_within_type_centroids.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nCentroids and real within-type RMS normalization are previously computed. No new embedding, centroids or annotation are fitted. Six broad types; small Other/DC groups require caution.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/b_type_centroid_shift_l1.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/b_type_centroid_shift_l1.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/b_type_centroid_shift_l1.csv", + "palette": "native", + "addition_key": "healthy_within_type_centroids", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/within_type_centroids.png" + ], + "source_page_131": 31, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 57, + "previous_figures": [ + "figures/expanded_master_20261013/healthy/within_type_centroids.png" + ], + "parent_enriched_page": 66, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/labels_within_type_centroids.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 67, + "parent_154_page": 79, + "parent_155_page": 80 + }, + { + "kind": "content", + "key": "healthy_gene_length_input", + "after_original": 19, + "title": "Gene-length effects in the bulk input", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/native/gene_length_input.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nOnly log10 axis conversion and plotting of saved per-gene shifts. The displayed Spearman values come directly from g_length_stats.csv; no correlation or regression is recomputed. Y values are input cpm_log1p differences. Gene length accounts for only part of assay mismatch; these are diagnostic associations.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_gene_shift_vs_length.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_length_stats.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_gene_shift_vs_length.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_length_stats.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_gene_shift_vs_length.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_length_stats.csv", + "palette": "native", + "addition_key": "healthy_gene_length_input", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/gene_length_input.png" + ], + "source_page_131": 32, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 58, + "parent_enriched_page": 67, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/gene_length_input.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 68, + "parent_154_page": 80, + "parent_155_page": 81 + }, + { + "kind": "content", + "key": "healthy_gene_length_stable", + "after_original": 19, + "title": "Gene-length effects within label-stable cells", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/native/gene_length_stable.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nOnly log10 axis conversion and plotting of saved per-gene shifts. The displayed Spearman values come directly from g_length_stats.csv; no correlation or regression is recomputed. Label-stable means the frozen broad label agrees between generated twins; it does not establish identical biological state. Y values are generated log1p CP10K differences.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_gene_shift_vs_length.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_length_stats.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_gene_shift_vs_length.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_length_stats.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_gene_shift_vs_length.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_length_stats.csv", + "palette": "native", + "addition_key": "healthy_gene_length_stable", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/gene_length_stable.png" + ], + "source_page_131": 33, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 59, + "parent_enriched_page": 68, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/gene_length_stable.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 69, + "parent_154_page": 81, + "parent_155_page": 82 + }, + { + "kind": "content", + "key": "healthy_raw_tpm_top_genes", + "after_original": 19, + "title": "Raw and TPM shifts persist within cell labels", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/raw_tpm_top_genes.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nExisting ranked gene lists only; no DEG testing. Displayed ranks are original saved ranks within direction and label-stable cells. Immediate-early/stress differences may reflect handling or cohort state; that interpretation is untested. Input and output shifts use different normalization and should not be interpreted as an identical quantitative scale.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_top_genes_raw_tpm.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_top_genes_raw_tpm.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_top_genes_raw_tpm.csv", + "palette": "native", + "addition_key": "healthy_raw_tpm_top_genes", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/raw_tpm_top_genes.png" + ], + "source_page_131": 34, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 60, + "parent_enriched_page": 69, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/raw_tpm_top_genes.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 70, + "parent_154_page": 82, + "parent_155_page": 83 + }, + { + "kind": "content", + "key": "healthy_refcal_top_genes", + "after_original": 19, + "title": "Reference calibration leaves specific gene-state shifts", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/native/refcal_top_genes.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nExisting ranked gene lists only; no DEG testing. Displayed ranks are original saved ranks within direction and label-stable cells. Immediate-early/stress differences may reflect handling or cohort state; that interpretation is untested. Input and output shifts use different normalization and should not be interpreted as an identical quantitative scale.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_top_genes_refcal.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_top_genes_refcal.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/g_top_genes_refcal.csv", + "palette": "native", + "addition_key": "healthy_refcal_top_genes", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/refcal_top_genes.png" + ], + "source_page_131": 35, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 61, + "parent_enriched_page": 70, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/refcal_top_genes.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 71, + "parent_154_page": 83, + "parent_155_page": 84 + }, + { + "kind": "content", + "key": "healthy_donor_axes", + "after_original": 19, + "title": "Donor trait axes and donor identification differ", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/donor_axes.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nAll Spearman and retrieval values are copied from the saved summary. Y-gene/sex calls and IFN/KLRC2 axes are expression-derived proxies, not independent clinical-label validation. Top-1 chance line is 1/107. High trait-axis agreement does not establish donor identity recovery.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/scorer_summary_5arm.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/scorer_summary_5arm.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/scorer_summary_5arm.csv", + "palette": "native", + "addition_key": "healthy_donor_axes", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/donor_axes.png" + ], + "source_page_131": 36, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 62, + "parent_enriched_page": 71, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/donor_axes.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 72, + "parent_154_page": 84, + "parent_155_page": 85 + }, + { + "kind": "content", + "key": "healthy_reference_choice", + "after_original": 19, + "title": "Reference choice changes different quality endpoints", + "section_label": "Supp. S03", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/reference_choice.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nAll eight generated arms are already scored in immicro_table.csv. No generation or rescore. AIDA and the training-mixture reference have their own pretraining overlap; this is not a held-out reference experiment. The composition dashed line is the observed donor-mean leave-one-out null, NOT a real-cell repeat composition error. No method-independent win claim is made across different endpoints.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/input_path_compare/immicro_behcet/tables/immicro_table.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/input_path_compare/immicro_behcet/tables/immicro_table.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/input_path_compare/immicro_behcet/tables/immicro_table.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md", + "palette": "native", + "addition_key": "healthy_reference_choice", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/reference_choice.png" + ], + "source_page_131": 37, + "deck_group": "Supp S03", + "section": "Supp S03", + "previous_section_label": "Control evidence", + "parent_page": 63, + "parent_enriched_page": 72, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/reference_choice.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 73, + "parent_154_page": 85, + "parent_155_page": 86 + }, + { + "kind": "evidence_note", + "title": "Why the historical and strict protocols differ", + "section_label": "Supp. S04", + "rows": [ + [ + "Historical healthy learning curves", + "LoRA used n training plus 7 validation pairs. Historical scSemi budget curves also reused CV-derived highly variable genes." + ], + [ + "Five donors versus five samples", + "Sepsis support is five samples; multi-visit Imperial, Giroux and Ji support can include more than five samples." + ], + [ + "New strict evaluation", + "Fixed-step, exact-sample support is a distinct protocol. Its partial results are labelled by completed cohort and draw." + ] + ], + "after_original": 21, + "addition_key": "guide:protocol", + "footnote": "Support draws overlap in donors. Displayed draw variation is not independent population replication.", + "footnote_full": "Support draws overlap in donors. Displayed draw variation is not independent population replication.", + "source_page_131": 40, + "deck_group": "Supp S04", + "section": "Supp S04", + "previous_section_label": "Protocol boundary", + "parent_page": 64, + "parent_enriched_page": 73, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 74, + "parent_154_page": 86, + "parent_155_page": 87 + }, + { + "kind": "content", + "key": "healthy_immicro_replicate_controls", + "after_original": 22, + "title": "Technical repeats and follow-up are different controls", + "section_label": "Supp. S04", + "layout": "full", + "figures": [ + "figures/expanded_master_20261013/healthy/immicro_replicate_controls.png" + ], + "keywords": [], + "footnote": "Observed ImMicro controls. Technical remeasurement and follow-up: 16 donors each. Follow-up is not a technical replicate.", + "footnote_full": "Previously processed ImMicro observations. Technical remeasurement and follow-up have 16 donors; capture-half control has its own recorded sample count. Follow-up is biological longitudinal variation, not a technical replicate. Correlation and top-1 retrieval are saved aggregates; no new subsampling, correlation or retrieval analysis. Use the recorded follow-up gene-support rule in observed_summary.json.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/build_manifest.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/observed_summary.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/replicate_pseudobulk_per_donor.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/build_manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/observed_summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/replicate_pseudobulk_per_donor.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/build_manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/observed_summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/replicate_pseudobulk_per_donor.csv", + "palette": "preserved", + "addition_key": "healthy_immicro_replicate_controls", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/immicro_replicate_controls.png" + ], + "source_page_131": 42, + "deck_group": "Supp S04", + "section": "Supp S04", + "previous_section_label": "Control evidence", + "parent_page": 65, + "parent_enriched_page": 74, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/immicro_replicate_controls.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 75, + "parent_154_page": 87, + "parent_155_page": 88 + }, + { + "kind": "content", + "key": "healthy_annotation_agreement", + "after_original": 22, + "title": "Annotation agreement and difficult categories", + "section_label": "Supp. S04", + "layout": "full", + "figures": [ + "figures/expanded_master_20261013/healthy/annotation_agreement.png" + ], + "keywords": [], + "footnote": "Observed cells only. Overall agreement uses the supported six-type gate; author mixed/unlabelled cells are separate. Other is difficult.", + "footnote_full": "Saved observed-cell comparison, not generated-cell model performance. Overall 97.17% agreement uses 365,404 author-labelled cells in the supported six-type gate; mixed/unlabelled categories are displayed in the matrix but excluded from that agreement figure. Other has weak correspondence between taxonomies. The frozen annotation's comparison to author labels is not an independent gold-standard validation. Heatmap is a count-row normalization for display; precision/recall are saved summary metrics.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/build_manifest.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/confusion_author_vs_frozen_baseline.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/observed_summary.json\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/build_manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/confusion_author_vs_frozen_baseline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/observed_summary.json", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/build_manifest.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/confusion_author_vs_frozen_baseline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/out/observed_summary.json", + "palette": "preserved", + "addition_key": "healthy_annotation_agreement", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/annotation_agreement.png" + ], + "source_page_131": 43, + "deck_group": "Supp S04", + "section": "Supp S04", + "previous_section_label": "Control evidence", + "parent_page": 66, + "parent_enriched_page": 75, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/annotation_agreement.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 76, + "parent_154_page": 88, + "parent_155_page": 89 + }, + { + "kind": "content", + "key": "healthy_depth_matched_quality", + "after_original": 24, + "title": "Depth matching changes precision and coverage", + "section_label": "Supp. S04", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/depth_matched_quality.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nBoth tables already exist. The matched-depth procedure thins all seven source sets to common detected-gene depth, as documented in the original analysis. Real-cell repeat is a held-out real-real control; neighborhood k=10 only. Display hatching, not a new sensitivity run.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/a_precision_coverage_pca50.csv\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/e_depth_matched_precision_coverage_pca50.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/a_precision_coverage_pca50.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/e_depth_matched_precision_coverage_pca50.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/a_precision_coverage_pca50.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/e_depth_matched_precision_coverage_pca50.csv", + "palette": "native", + "addition_key": "healthy_depth_matched_quality", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/depth_matched_quality.png" + ], + "source_page_131": 46, + "deck_group": "Supp S04", + "section": "Supp S04", + "previous_section_label": "Control evidence", + "parent_page": 67, + "parent_enriched_page": 76, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/healthy/depth_matched_quality.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 77, + "parent_154_page": 89, + "parent_155_page": 90 + }, + { + "kind": "content", + "key": "healthy_detection_and_library", + "after_original": 24, + "title": "Detection and library size are separate diagnostics", + "section_label": "Supp. S04", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/labels_detection_and_library.png" + ], + "keywords": [], + "footnote": "Historical native-depth diagnostic: sc-derived depth/masks; 10-pair calibration; AIFI overlaps pretraining. One seed.", + "footnote_full": "Historical zero-shot input diagnostics, not strict-five LoRA/scSemiProfiler comparison. Final f01 6k 1x E16 EMA; 107 ImMicro donors; 1,500 generated cells/donor; 50 DDIM steps; Bernoulli support; generation seed 20260922 shared across input arms. Native depth and cell count, and the output gene-availability mask, use each donor's sc data. Paired calibration uses 10 paired samples. AIFI alpha uses no paired ImMicro sc expression, but 85 of its 108 reference donors overlap pretraining; this is not evidence of a completely sc-free deployment. Most cell-space metrics use 150 cells/donor and 5,508 common measured genes; PCA50/UMAP fitted only on observed reference. Generated counts are Poisson readouts, not observed UMIs. One cohort and one seed; no new uncertainty tests. The source manifest records exact reproduction of the earlier four-arm controls.\n\nObserved counts and the previously computed Poisson readout of generated cells. Whiskers are saved 10th\u201390th cell percentiles, not donor confidence intervals. Scoring-gene libraries are not full-transcriptome UMI totals.\n\nSource files (existing results, read only):\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md\n/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/e_detected_genes.csv\nEvidence source: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/e_detected_genes.csv", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/MANIFEST.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/RESULTS.md; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tables/e_detected_genes.csv", + "palette": "native", + "addition_key": "healthy_detection_and_library", + "figures_before_palette": [ + "figures/expanded_master_20261013/healthy/detection_and_library.png" + ], + "source_page_131": 47, + "deck_group": "Supp S04", + "section": "Supp S04", + "previous_section_label": "Control evidence", + "parent_page": 68, + "previous_figures": [ + "figures/expanded_master_20261013/healthy/detection_and_library.png" + ], + "parent_enriched_page": 77, + "figures_before_final_palette": [ + "figures/figure_enrichment_20261013/labels_detection_and_library.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 78, + "parent_154_page": 90, + "parent_155_page": 91 + }, + { + "kind": "content", + "section": "Supp S05", + "section_label": "Supp. S05", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 0 + }, + "title": "Healthy cohort: LoRA cells cover the real ones", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig2_healthy_cells.png" + ], + "keywords": [ + "5 donors: coverage 0.19 \u2192 0.79", + "Precision 0.95; composition L1 0.18", + "5-fold: coverage 0.78, L1 0.16" + ], + "footnote": "ImMicro, 107 healthy donors. 5 donors: mean of 5 draws (dots); 5-fold CV shown separately.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs": [ + "Kynk\u00e4\u00e4nniemi 2019, NeurIPS", + "Naeem 2020, ICML" + ], + "refs_full": [ + "Kynk\u00e4\u00e4nniemi et al. Improved precision and recall metric for assessing generative models. NeurIPS 2019", + "Naeem et al. Reliable fidelity and diversity metrics for generative models. ICML 2020" + ], + "say": "On healthy donors, five pairs and a LoRA of minutes take coverage from about 0.2 to about 0.8: the generated cells now cover the real cells.", + "footnote_full": "Original slide 23: ImMicro, 107 healthy donors (no disease metric on this slide). 5 donors: LoRA-5 fine-tuned on 5 paired donors, tested on the other 102, 5 draws (dots, bar = mean). 5-fold CV: every donor generated by the fold model that never trained on it (pooled call), reported separately. Coverage / precision: k = 10 nearest-neighbour balls on held-out real cells in the scorer's PCA-50 (Kynk\u00e4\u00e4nniemi 2019; Naeem 2020); composition L1 over six scANVI types. Values: coverage 5 donors Zero-shot, real bulk 0.190, Zero-shot + calibration 0.535, PBISC LoRA 0.792; 5-fold Zero-shot, real bulk 0.192, Zero-shot + calibration 0.527, sc-only fine-tune (+ cal.) 0.623, PBISC LoRA 0.784; precision 5 donors Zero-shot, real bulk 0.677, Zero-shot + calibration 0.904, PBISC LoRA 0.952; 5-fold Zero-shot, real bulk 0.680, Zero-shot + calibration 0.905, sc-only fine-tune (+ cal.) 0.805, PBISC LoRA 0.924; comp_l1 5 donors Zero-shot, real bulk 0.445, Zero-shot + calibration 0.203, PBISC LoRA 0.176; 5-fold Zero-shot, real bulk 0.455, Zero-shot + calibration 0.199, sc-only fine-tune (+ cal.) 0.240, PBISC LoRA 0.162. Sources: data/sec09_fewpair_cells.csv, data/sec07_ft_modes_accuracy.csv.", + "numbers": [ + { + "value": "5 donors (mean of 5 draws) coverage zs_raw 0.19; 5 donors (mean of 5 draws) coverage zs_cal 0.5352; 5 donors (mean of 5 draws) coverage lora 0.7925; 5-fold pooled coverage zs_raw 0.1918; 5-fold pooled coverage zs_cal 0.5272; 5-fold pooled coverage sconly_cal 0.6233; 5-fold pooled coverage lora 0.7843; 5 donors (mean of 5 draws) precision zs_raw 0.6772; 5 donors (mean of 5 draws) precision zs_cal 0.9037; 5 donors (mean of 5 draws) precision lora 0.9521; 5-fold pooled precision zs_raw 0.6804; 5-fold pooled precision zs_cal 0.905; 5-fold pooled precision sconly_cal 0.8047; 5-fold pooled precision lora 0.9241; 5 donors (mean of 5 draws) comp_l1 zs_raw 0.4451; 5 donors (mean of 5 draws) comp_l1 zs_cal 0.2028; 5 donors (mean of 5 draws) comp_l1 lora 0.1756; 5-fold pooled comp_l1 zs_raw 0.4549; 5-fold pooled comp_l1 zs_cal 0.1985; 5-fold pooled comp_l1 sconly_cal 0.2403; 5-fold pooled comp_l1 lora 0.1617", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 23, + "original_title": "Healthy cohort: LoRA cells cover the real ones", + "original_footnote": "ImMicro, 107 healthy donors. 5 donors: mean of 5 draws (dots); 5-fold CV shown separately.", + "figures_before_palette": [ + "figures/fig2_healthy_cells.png" + ], + "source_page_131": 44, + "deck_group": "Supp S05", + "previous_section_label": "Fig. 2e", + "parent_page": 69, + "parent_enriched_page": 78, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig2_healthy_cells.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 79, + "parent_154_page": 91, + "parent_155_page": 92 + }, + { + "kind": "content", + "section": "Supp S05", + "section_label": "Supp. S05", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 1 + }, + "title": "Healthy cohort: realism beyond precision/coverage", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig2_healthy_realism.png" + ], + "keywords": [ + "Every metric: LoRA best of the PBISC versions", + "E_N 1.5 vs zero-shot 7.2 (5 donors)", + "Classifier AUC 0.94: still separable" + ], + "footnote": "ImMicro, 107 healthy donors. Dashed = real vs real cells of the same donors.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs": [ + "Lopez-Paz 2017, ICLR", + "Gretton 2012, JMLR" + ], + "refs_full": [ + "Lopez-Paz & Oquab. Revisiting classifier two-sample tests. ICLR 2017", + "Gretton et al. A kernel two-sample test. JMLR 13:723-773 (2012)" + ], + "say": "The same holds for the stricter realism metrics; LoRA is the best PBISC version, although a classifier can still tell generated from real cells.", + "footnote_full": "Original slide 24: ImMicro. Classifier two-sample test (kNN, 5-fold CV ROC AUC; 0.5 = indistinguishable; Lopez-Paz & Oquab 2017), MMD\u00b2 (RBF, median bandwidth; Gretton 2012) and scSemiProfiler's E_N (Wang 2024), all on the scorer's cells and embedding. Dashed: real vs real (held-out cells of the same donors). 5 donors: mean of 5 draws (dots); 5-fold reported separately. Source: data/sec09_extra_fewpair.csv, data/sec09_extra_fivefold.csv.", + "numbers": [ + { + "value": "5 donors (mean of 5 draws) c2st_knn_auc zs_raw 0.9966; 5 donors (mean of 5 draws) c2st_knn_auc zs_cal 0.9732; 5 donors (mean of 5 draws) c2st_knn_auc lora 0.9427; 5-fold pooled c2st_knn_auc zs_raw 0.9964; 5-fold pooled c2st_knn_auc zs_cal 0.9728; 5-fold pooled c2st_knn_auc sconly_cal 0.9524; 5-fold pooled c2st_knn_auc lora 0.9409; 5 donors (mean of 5 draws) mmd2_bw1 zs_raw 0.1026; 5 donors (mean of 5 draws) mmd2_bw1 zs_cal 0.04304; 5 donors (mean of 5 draws) mmd2_bw1 lora 0.02308; 5-fold pooled mmd2_bw1 zs_raw 0.1019; 5-fold pooled mmd2_bw1 zs_cal 0.04253; 5-fold pooled mmd2_bw1 sconly_cal 0.03419; 5-fold pooled mmd2_bw1 lora 0.0264; 5 donors (mean of 5 draws) en_all_sq zs_raw 7.236; 5 donors (mean of 5 draws) en_all_sq zs_cal 2.855; 5 donors (mean of 5 draws) en_all_sq lora 1.547; 5-fold pooled en_all_sq zs_raw 4.832; 5-fold pooled en_all_sq zs_cal 2.263; 5-fold pooled en_all_sq sconly_cal 1.767; 5-fold pooled en_all_sq lora 1.56", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 24, + "original_title": "Healthy cohort: realism beyond precision/coverage", + "original_footnote": "ImMicro, 107 healthy donors. Dashed = real vs real cells of the same donors.", + "figures_before_palette": [ + "figures/fig2_healthy_realism.png" + ], + "source_page_131": 45, + "deck_group": "Supp S05", + "previous_section_label": "Fig. 2f", + "parent_page": 70, + "parent_enriched_page": 79, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig2_healthy_realism.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 80, + "parent_154_page": 92, + "parent_155_page": 93 + }, + { + "kind": "content", + "section": "Supp S05", + "section_label": "Supp. S05", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 2 + }, + "title": "Healthy cohort: full vs LoRA vs condition encoder", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig2_healthy_modes.png" + ], + "keywords": [ + "LoRA \u2248 full fine-tune with 42\u00d7 fewer weights", + "LoRA coverage 0.78 vs full 0.70 (5-fold)", + "Condition encoder alone falls short" + ], + "footnote": "ImMicro, 107 healthy donors; cond.-encoder-only ran on fold 0 (22 donors) only.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs": [ + "Hu 2022, ICLR" + ], + "refs_full": [ + "Hu et al. LoRA: Low-Rank Adaptation of Large Language Models. ICLR 2022" + ], + "say": "LoRA matches full fine-tuning with 42 times fewer trainable weights; tuning only the condition encoder is not enough.", + "footnote_full": "Original slide 25: ImMicro. Full fine-tune trains all 54.2 M weights; LoRA r16 1.30 M; condition-encoder-only 9.48 M (fold 0 pilot only). Values: pooled Full fine-tune: L1 0.157, coverage 0.695, precision 0.890; pooled PBISC LoRA: L1 0.162, coverage 0.784, precision 0.924; fold0 Full fine-tune: L1 0.147, coverage 0.841, precision 0.987; fold0 PBISC LoRA: L1 0.153, coverage 0.915, precision 0.992; fold0 Cond.-encoder-only: L1 0.204, coverage 0.728, precision 0.979. Source: data/sec07_ft_modes_accuracy.csv, data/sec07_ft_modes_compute.csv.", + "numbers": [ + { + "value": "pooled comp_l1 full 0.1572; pooled comp_l1 lora 0.1617; fold0 comp_l1 full 0.147; fold0 comp_l1 lora 0.1529; fold0 comp_l1 condenc 0.2042; pooled coverage full 0.6954; pooled coverage lora 0.7843; fold0 coverage full 0.8412; fold0 coverage lora 0.9148; fold0 coverage condenc 0.7285; pooled precision full 0.8895; pooled precision lora 0.9241; fold0 precision full 0.987; fold0 precision lora 0.9915; fold0 precision condenc 0.9794; trainable_M full 54.19; trainable_M lora 1.296; trainable_M condenc 9.483", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 25, + "original_title": "Healthy cohort: full vs LoRA vs condition encoder", + "original_footnote": "ImMicro, 107 healthy donors; cond.-encoder-only ran on fold 0 (22 donors) only.", + "figures_before_palette": [ + "figures/fig2_healthy_modes.png" + ], + "source_page_131": 48, + "deck_group": "Supp S05", + "previous_section_label": "Fig. 2g", + "parent_page": 71, + "parent_enriched_page": 80, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig2_healthy_modes.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 81, + "parent_154_page": 93, + "parent_155_page": 94 + }, + { + "kind": "content", + "section": "Supp S05", + "section_label": "Supp. S05", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 7 + }, + "title": "Representative-dependent composition", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig3_healthy_copy.png" + ], + "keywords": [ + "scSemiProfiler 0.233 vs copying 0.187 (L1)", + "Its error tracks the copy's: \u03c1 0.57", + "LoRA-5 below the copy: 0.176" + ], + "footnote": "Existing healthy five-draw diagnostic. This evaluates composition dependence, not literal copying of every cell.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs": [ + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "say": "scSemiProfiler does worse than simply copying its representative's composition, and its error follows that representative; LoRA does better than the copy.", + "footnote_full": "Current scope: Existing healthy five-draw diagnostic. This evaluates composition dependence, not literal copying of every cell.\n\nOriginal slide 32: ImMicro, 5 donors. scSemiProfiler assigns each test sample one representative; 'copy' uses that representative's observed composition as the prediction. Right: Spearman over test samples between each method's composition error and the copy's error. Source: data/sec09_scsemi_representative.csv (scsemi/analysis/representative_dependence_20261002).", + "numbers": [ + { + "value": "5 donors (mean of 5 draws) comp_l1 copy 0.1871; 5 donors (mean of 5 draws) comp_l1 scsemi 0.2325; 5 donors (mean of 5 draws) comp_l1 lora 0.1756; spearman_with_copy_error scsemi 0.5736; spearman_with_copy_error lora 0.3509", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 32, + "original_title": "Healthy donors: scSemiProfiler copies its representative", + "original_footnote": "ImMicro, 102 test donors per draw, 5 draws. Copy = the assigned representative's composition.", + "figures_before_palette": [ + "figures/fig3_healthy_copy.png" + ], + "source_page_131": 57, + "deck_group": "Supp S05", + "previous_section_label": "Fig. 3c", + "parent_page": 72, + "parent_enriched_page": 81, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig3_healthy_copy.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 82, + "parent_154_page": 94, + "parent_155_page": 95 + }, + { + "kind": "content", + "section": "Supp S05", + "section_label": "Supp. S05", + "title": "Full vs LoRA vs cond.-encoder-only", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec07_ft_modes.png" + ], + "keywords": [ + "LoRA \u2248 full FT, 42\u00d7 fewer params", + "Cond.-encoder-only \u2248 ZS + cal. on ImMicro", + "LoRA beats both on coverage" + ], + "footnote": "Fold 0, cond.-enc. vs ZS + cal.: coverage 0.73 vs 0.73 / 0.74 vs 0.63, L1 p 0.06 / 0.93 (ImMicro / sepsis).", + "sources": "tools/sec07_ft_modes.py; data/sec07_ft_modes_accuracy.csv, data/sec07_ft_modes_compute.csv, data/sec07_ft_modes_wilcoxon.csv", + "footnote_full": "Original slide 64: Base f01_6k_1x E16. ImMicro / sepsis; paired Wilcoxon on donor values, uncorrected. Full vs LoRA L1 p 0.25 / 0.71 (pooled); 42\u00d7 fewer trainable params (1.3 M). Zero-shot + calibration: \u03b1_g = mean CPM(pseudobulk_g) / mean CPM(bulk_g) over the fold's 85\u201386 / 65\u201366 paired training donors (train + inner-val); bulk CPM \u00d7 \u03b1, renormalised. The sc-only FT never saw real bulk, so it is fed calibrated bulk. Raw \u2192 calibrated zero-shot (pooled): L1 0.45 \u2192 0.20 / 0.51 \u2192 0.40, coverage 0.19 \u2192 0.53 / 0.07 \u2192 0.43, sepsis logFC r \u22120.01 \u2192 0.12. LoRA vs ZS + cal. (pooled): coverage 0.78 vs 0.53 / 0.83 vs 0.43 (p 7e-19 / 4e-15); L1 0.16 vs 0.20 (p 2e-7) / 0.36 vs 0.40 (p 0.19); logFC r 0.53 vs 0.12. Fold 0, cond.-enc. vs ZS + cal.: L1 0.20 vs 0.17 (p 0.06) / 0.38 vs 0.39 (p 0.93), coverage 0.73 vs 0.73 (p 0.68) / 0.74 vs 0.63 (p 7e-4). Fold 0, LoRA: coverage 0.91 / 0.86; vs cond.-enc. p 4e-5 / 1e-3, vs ZS + cal. p 6e-5 / 3e-4; L1 vs cond.-enc. p 3e-5 / 0.52, vs ZS + cal. p 0.29 / 0.96. sc-only FT (+ cal.), pooled: coverage 0.62 / 0.73, logFC r 0.30. Open = fold-0 pilot; compare same marker only. Per-step time: runs early-stop differently.", + "refs": [ + "Hu 2022, ICLR" + ], + "refs_full": [ + "Hu et al. LoRA: Low-Rank Adaptation of Large Language Models. ICLR 2022" + ], + "original_slide": 64, + "original_title": "Full vs LoRA vs cond.-encoder-only", + "original_footnote": "Fold 0, cond.-enc. vs ZS + cal.: coverage 0.73 vs 0.73 / 0.74 vs 0.63, L1 p 0.06 / 0.93 (ImMicro / sepsis).", + "figures_before_palette": [ + "figures/sec07_ft_modes.png" + ], + "source_page_131": 116, + "deck_group": "Supp S05", + "previous_section_label": "Supplementary", + "parent_page": 73, + "parent_enriched_page": 82, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec07_ft_modes.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 83, + "parent_154_page": 95, + "parent_155_page": 96 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 3 + }, + "title": "Pair-budget summary: composition, coverage and steps", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig2_healthy_pairs.png" + ], + "keywords": [], + "footnote": "Historical n training + 7 validation. All six cell types are shown in the main pair-budget panels; this is the original summary.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs_full": [ + "Kynk\u00e4\u00e4nniemi et al. Improved precision and recall metric for assessing generative models. NeurIPS 2019", + "Naeem et al. Reliable fidelity and diversity metrics for generative models. ICML 2020" + ], + "footnote_full": "Historical n training + 7 validation. All six cell types are shown in the main pair-budget panels; this is the original summary.\nCurrent scope: Historical n + 7 validation design. The fixed-3000-step strict curve is a separate protocol update.\n\nOriginal slide 26: ImMicro, donor-level 5-fold CV, pooled scoring (107 donors); x = LoRA training pairs per fold (the last point is every training donor of the fold, 78\u201379). Calibration fitted on exactly the pairs the LoRA point saw. Selected step = early-stopping step on 7 inner-validation pairs (cap 20,000). Values: coverage LoRA 0.84 / 0.79 / 0.76 / 0.78 / 0.78; NK fraction r LoRA 0.20 / 0.40 / 0.56 / 0.70 / 0.70 at 5 / 10 / 20 / 40 / all pairs. Source: data/sec08_pairs_realism.csv, data/sec08_pairs_signal.csv, data/sec08_selected_steps_per_fold.csv.", + "numbers": [ + { + "value": "pairs 0 comp_l1 zero_shot 0.4549; pairs 5 comp_l1 zs_cal_budget 0.1916; pairs 10 comp_l1 zs_cal_budget 0.1996; pairs 20 comp_l1 zs_cal_budget 0.1911; pairs 40 comp_l1 zs_cal_budget 0.196; pairs all comp_l1 zs_cal_budget 0.1985; pairs 5 comp_l1 lora 0.1715; pairs 10 comp_l1 lora 0.1879; pairs 20 comp_l1 lora 0.1723; pairs 40 comp_l1 lora 0.1714; pairs all comp_l1 lora 0.1617; pairs 0 coverage zero_shot 0.1918; pairs 5 coverage zs_cal_budget 0.527; pairs 10 coverage zs_cal_budget 0.5287; pairs 20 coverage zs_cal_budget 0.5287; pairs 40 coverage zs_cal_budget 0.5224; pairs all coverage zs_cal_budget 0.5272; pairs 5 coverage lora 0.8392; pairs 10 coverage lora 0.7857; pairs 20 coverage lora 0.7552; pairs 40 coverage lora 0.7784; pairs all coverage lora 0.7843; pairs 0 nk_fraction_r zero_shot 0.4939; pairs 5 nk_fraction_r zs_cal_budget 0.5019; pairs 10 nk_fraction_r zs_cal_budget 0.515; pairs 20 nk_fraction_r zs_cal_budget 0.5251; pairs 40 nk_fraction_r zs_cal_budget 0.5; pairs all nk_fraction_r zs_cal_budget 0.5004; pairs 5 nk_fraction_r lora 0.1993; pairs 10 nk_fraction_r lora 0.3972; pairs 20 nk_fraction_r lora 0.5644; pairs 40 nk_fraction_r lora 0.6993; pairs all nk_fraction_r lora 0.7019; pairs 5 selected_step_median_k lora 3.25; pairs 10 selected_step_median_k lora 5.75; pairs 20 selected_step_median_k lora 19.5; pairs 40 selected_step_median_k lora 20; pairs all selected_step_median_k lora 20", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 26, + "original_title": "Healthy cohort: how many pairs", + "original_footnote": "ImMicro, 107 donors, 5-fold CV pooled at each pair count. 0 = no pairs.", + "figures_before_palette": [ + "figures/fig2_healthy_pairs.png" + ], + "source_page_131": 49, + "deck_group": "Supp S06", + "previous_section_label": "Fig. 2h", + "parent_page": 21, + "parent_enriched_page": 31, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig2_healthy_pairs.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 84, + "parent_154_page": 96, + "parent_155_page": 97 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 4 + }, + "title": "Healthy cohort: LoRA rank and K", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig2_healthy_rank.png" + ], + "keywords": [ + "r64 vs r16: coverage 0.78 \u2192 0.82", + "r4 vs r16: L1 0.162 \u2192 0.186", + "Multi-K, recipes: no gain; kept r16, K 1,000" + ], + "footnote": "ImMicro, 107 donors, 5-fold CV pooled. Open symbols: K choices with every training pair.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs": [ + "Hu 2022, ICLR" + ], + "refs_full": [ + "Hu et al. LoRA: Low-Rank Adaptation of Large Language Models. ICLR 2022" + ], + "say": "Rank 16 is a good default; rank 64 or extra K choices change little.", + "footnote_full": "Original slide 27: ImMicro, 5-fold CV pooled. LoRA rank 4 / 16 / 64 at 5, 20 and all training pairs; multi-K (K = 250\u20134,000 conditions) and half-batch recipes with every training pair. Kept: r = 16, K = 1,000. Source: data/sec07_rank_sweep.csv, data/sec07_k_aug.csv.", + "numbers": [ + { + "value": "pairs 5 comp_l1 lora_r4 0.1622; pairs 20 comp_l1 lora_r4 0.171; pairs all comp_l1 lora_r4 0.1858; pairs 5 comp_l1 lora_r16 0.1715; pairs 20 comp_l1 lora_r16 0.1723; pairs all comp_l1 lora_r16 0.1617; pairs 5 comp_l1 lora_r64 0.1765; pairs 20 comp_l1 lora_r64 0.1717; pairs all comp_l1 lora_r64 0.161; pairs all comp_l1 lora_multik__raw 0.1678; pairs all comp_l1 lora_recipe__raw 0.1609; pairs 5 coverage lora_r4 0.7891; pairs 20 coverage lora_r4 0.715; pairs all coverage lora_r4 0.7123; pairs 5 coverage lora_r16 0.8392; pairs 20 coverage lora_r16 0.7552; pairs all coverage lora_r16 0.7843; pairs 5 coverage lora_r64 0.8473; pairs 20 coverage lora_r64 0.7937; pairs all coverage lora_r64 0.8159; pairs all coverage lora_multik__raw 0.7748; pairs all coverage lora_recipe__raw 0.7655; pairs 5 precision lora_r4 0.9369; pairs 20 precision lora_r4 0.9247; pairs all precision lora_r4 0.9333; pairs 5 precision lora_r16 0.9463; pairs 20 precision lora_r16 0.9204; pairs all precision lora_r16 0.9241; pairs 5 precision lora_r64 0.9455; pairs 20 precision lora_r64 0.9053; pairs all precision lora_r64 0.9044; pairs all precision lora_multik__raw 0.9284; pairs all precision lora_recipe__raw 0.933", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 27, + "original_title": "Healthy cohort: LoRA rank and K", + "original_footnote": "ImMicro, 107 donors, 5-fold CV pooled. Open symbols: K choices with every training pair.", + "figures_before_palette": [ + "figures/fig2_healthy_rank.png" + ], + "source_page_131": 50, + "deck_group": "Supp S06", + "previous_section_label": "Fig. 2i", + "parent_page": 74, + "parent_enriched_page": 83, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig2_healthy_rank.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 85, + "parent_154_page": 97, + "parent_155_page": 98 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "text_source": { + "file": "data/fig23_healthy_slide_text.json", + "index": 6 + }, + "title": "Healthy paired-data learning curves", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig3_healthy_pairs.png" + ], + "keywords": [ + "Coverage at every n: LoRA 0.76\u20130.84, scSemiProfiler 0.04", + "Composition L1 lower for LoRA at every n", + "5-fold: LoRA 0.16 vs 0.22 L1" + ], + "footnote": "LoRA: n training + 7 validation. scSemi: n + 7 representatives, with CV-HVG reuse. Not the strict total-n protocol.", + "sources": "tools/fig23_healthy.py; data/fig23_healthy.csv", + "refs": [ + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "say": "Giving scSemiProfiler exactly LoRA's pairs, from five up to every training donor, does not close the gap in coverage or composition.", + "footnote_full": "Current scope: LoRA: n training + 7 validation. scSemi: n + 7 representatives, with CV-HVG reuse. Not the strict total-n protocol.\n\nOriginal slide 31: ImMicro, donor-level 5-fold CV pooled at each pair count: for every fold, scSemiProfiler's representatives are exactly the n pairs LoRA trained on (plus the inner-validation pairs); 'all' (open symbols, right of the dotted line) = every training donor of the fold, i.e. the 5-fold result, reported separately from the 5-pair comparison. Source: data/sec08_pairs_realism.csv, data/sec08_pairs_signal.csv, data/sec09_same_pairs.csv.", + "numbers": [ + { + "value": "pairs 5 comp_l1 scsemi 0.2184; pairs 10 comp_l1 scsemi 0.2269; pairs 20 comp_l1 scsemi 0.2232; pairs 40 comp_l1 scsemi 0.2213; pairs all comp_l1 scsemi 0.217; pairs 5 comp_l1 lora 0.1715; pairs 10 comp_l1 lora 0.1879; pairs 20 comp_l1 lora 0.1723; pairs 40 comp_l1 lora 0.1714; pairs all comp_l1 lora 0.1617; pairs 5 coverage scsemi 0.03819; pairs 10 coverage scsemi 0.03801; pairs 20 coverage scsemi 0.0372; pairs 40 coverage scsemi 0.03844; pairs all coverage scsemi 0.03844; pairs 5 coverage lora 0.8392; pairs 10 coverage lora 0.7857; pairs 20 coverage lora 0.7552; pairs 40 coverage lora 0.7784; pairs all coverage lora 0.7843; pairs 5 nk_fraction_r scsemi 0.2651; pairs 10 nk_fraction_r scsemi 0.33; pairs 20 nk_fraction_r scsemi 0.4762; pairs 40 nk_fraction_r scsemi 0.4937; pairs all nk_fraction_r scsemi 0.4245; pairs 5 nk_fraction_r lora 0.1993; pairs 10 nk_fraction_r lora 0.3972; pairs 20 nk_fraction_r lora 0.5644; pairs 40 nk_fraction_r lora 0.6993; pairs all nk_fraction_r lora 0.7019", + "source": "data/fig23_healthy.csv" + } + ], + "original_slide": 31, + "original_title": "Healthy donors: the same pairs, 5 to all", + "original_footnote": "ImMicro, 107 donors, 5-fold CV design; scSemiProfiler gets exactly LoRA's pairs.", + "figures_before_palette": [ + "figures/fig3_healthy_pairs.png" + ], + "source_page_131": 56, + "deck_group": "Supp S06", + "previous_section_label": "Fig. 3b", + "parent_page": 75, + "parent_enriched_page": 84, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig3_healthy_pairs.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 86, + "parent_154_page": 98, + "parent_155_page": 99 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "text_source": { + "file": "data/fig3_time_days_slide_text.json", + "index": 0 + }, + "title": "Time: days for scSemiProfiler, minutes for LoRA", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig3_time_days.png" + ], + "keywords": [ + "5 donors, 5 draws: 5\u201320 h vs 13\u201354 min", + "5-fold: up to 1.6 days vs 37 min (Sepsis)", + "All runs: 7.7 days vs 5.7 h" + ], + "footnote": "One GB10 GPU, job alone on it. GPU time from receipts; waits for the GPU excluded.", + "sources": "tools/fig3_time_days.py; data/fig3_time_days.csv; data/sec09_same_pairs.csv; data/sec09_time_to_apply.csv; scsemi/work//cv/fold/events.jsonl", + "refs": [ + "Wang 2024, Nat Commun", + "Hu 2022, ICLR" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1", + "Hu et al. LoRA: Low-Rank Adaptation of Large Language Models. ICLR 2022" + ], + "say": "Time is the practical difference. scSemiProfiler trains a model for every representative and then one more model for every test sample. With five paired donors, one cohort's five draws take it five to twenty GPU hours; LoRA takes under an hour. A 5-fold run takes scSemiProfiler up to one and a half GPU-days and LoRA about half an hour. Counting only per-cohort work on both sides, all our runs took almost eight GPU-days for scSemiProfiler and under six hours for LoRA.", + "footnote_full": "Original slide 33: GPU time per cohort on one GB10 GPU with every step alone on the GPU, from each run's receipts; waits for the GPU (our environment) are excluded. Main comparison: 5 donors, i.e. LoRA\u20135 vs scSemiProfiler with the same 5 pairs as representatives, all 5 draws per cohort. 5\u2013fold CV (the five fold models, every usable sample tested once) is shown separately. scSemiProfiler = pretraining of every representative (two stages) + one model per test sample (upstream c2aea02, default settings); PBISC LoRA = fine\u2013tune + sampling 1,500 cells per test sample. Ratio per cohort: 5 donors 17\u201359x, 5\u2013fold 19\u201362x. scSemiProfiler 5\u2013fold running time including its CPU steps: ImMicro (healthy) 27 h; Sepsis 41 h; Imperial 33 h; Giroux 15 h; Ji (whole blood) 8 h. All runs = 25 five\u2013donor draws + 25 fold models: scSemiProfiler 184 h, LoRA 5.7 h. Not counted on either side: data preparation, scoring, and what each method provides (PBISC's pretrained model, scSemiProfiler's code). The equal\u2013pair\u2013budget scSemiProfiler runs are not included (they reuse the 5\u2013fold run's representative models). Sources: data/sec09_same_pairs.csv, data/sec09_time_to_apply.csv, scsemi/work//cv/fold/events.jsonl.", + "numbers": [ + { + "value": "ImMicro (healthy): 5 donors (5 draws) scSemiProfiler 11.7 h vs LoRA 0.70 h; 5-fold 23.7 h (0.99 days) vs 1.22 h (19x); Sepsis: 5 donors (5 draws) scSemiProfiler 20.3 h vs LoRA 0.56 h; 5-fold 38.5 h (1.60 days) vs 0.62 h (62x); Imperial: 5 donors (5 draws) scSemiProfiler 19.1 h vs LoRA 0.90 h; 5-fold 30.9 h (1.29 days) vs 0.52 h (60x); Giroux: 5 donors (5 draws) scSemiProfiler 13.0 h vs LoRA 0.22 h; 5-fold 14.4 h (0.60 days) vs 0.30 h (49x); Ji (whole blood): 5 donors (5 draws) scSemiProfiler 4.8 h vs LoRA 0.28 h; 5-fold 7.4 h (0.31 days) vs 0.37 h (20x)", + "source": "data/fig3_time_days.csv" + }, + { + "value": "all runs: scSemiProfiler 183.8 GPU h (7.66 days); PBISC LoRA 5.68 h", + "source": "data/fig3_time_days.csv" + } + ], + "original_slide": 33, + "original_title": "Time: days for scSemiProfiler, minutes for LoRA", + "original_footnote": "One GB10 GPU, job alone on it. GPU time from receipts; waits for the GPU excluded.", + "figures_before_palette": [ + "figures/fig3_time_days.png" + ], + "source_page_131": 58, + "deck_group": "Supp S06", + "previous_section_label": "Fig. 3d", + "parent_page": 76, + "parent_enriched_page": 85, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig3_time_days.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 87, + "parent_154_page": 99, + "parent_155_page": 100 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "text_source": { + "file": "data/sec09_time_to_apply_slide_text.json", + "index": 0 + }, + "title": "5 donors: time to apply", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_time_to_apply.png" + ], + "keywords": [ + "LoRA-5: 3\u201311 min per draw, every test sample", + "scSemiProfiler-5: 57 min\u20134.1 h per draw", + "17\u201359\u00d7 less GPU time, same 5 donors" + ], + "footnote": "Existing measurements: adaptation + target generation; pretraining, GPU wait and scoring excluded.", + "sources": "tools/sec09_time_to_apply.py; data/sec09_time_to_apply.csv, data/sec09_time_to_apply_runs.csv; data/sec06_ft_method.csv (pretrain_wall_hours)", + "footnote_full": "Current scope: Existing measurements: adaptation + target generation; pretraining, GPU wait and scoring excluded.\n\nOriginal slide 34: Same 25 draws as the '5 donors' slides (5 cohorts x 5 draws; same 5 training donors and test samples for both methods). GPU time on one GB10, every step alone on the GPU (shared lock). LoRA-5: fine-tune = RUN.json wall_seconds of the 3,000-step LoRA; generate = wall clock of the sampling job in the queue log (model load + DDIM 50, 1,500 cells per test sample). scSemiProfiler-5 (upstream c2aea02, default budget): stage 1 + stage 2 pretraining of every representative (model receipts) + its per-target inference (5 ministages x 150 epochs, export receipt total_target_seconds). Not counted on either side: CPU data preparation, scoring, work lost to interruptions, and PBISC's one-time pretraining (15.9 h; 'Fine-tuning with paired bulk + sc' slide), which every cohort reuses. scSemiProfiler time grows with the number of test samples because it trains one model per sample. Per-draw ratio scSemiProfiler / LoRA: ImMicro 15\u201319\u00d7; Sepsis 26\u201347\u00d7; Imperial 19\u201326\u00d7; Giroux 54\u201363\u00d7; Ji 11\u201320\u00d7 (all 25 draws 11\u201363\u00d7; keyword and figure = mean of each cohort's 5 draws).", + "numbers": [ + { + "value": "ImMicro: LoRA-5 8.4 min (fine-tune 1.9 + generate 6.5), scSemiProfiler-5 141 min (representatives 15 + per-sample models 126); 15\u201319\u00d7 per draw; per test sample 4.9 s vs 83 s", + "source": "data/sec09_time_to_apply.csv" + }, + { + "value": "Sepsis: LoRA-5 6.7 min (fine-tune 1.8 + generate 4.9), scSemiProfiler-5 244 min (representatives 30 + per-sample models 214); 26\u201347\u00d7 per draw; per test sample 5.2 s vs 190 s", + "source": "data/sec09_time_to_apply.csv" + }, + { + "value": "Imperial: LoRA-5 10.8 min (fine-tune 1.8 + generate 9.0), scSemiProfiler-5 229 min (representatives 32 + per-sample models 197); 19\u201326\u00d7 per draw; per test sample 4.5 s vs 96 s", + "source": "data/sec09_time_to_apply.csv" + }, + { + "value": "Giroux: LoRA-5 2.6 min (fine-tune 1.9 + generate 0.7), scSemiProfiler-5 156 min (representatives 117 + per-sample models 38); 54\u201363\u00d7 per draw; per test sample 14.8 s vs 872 s", + "source": "data/sec09_time_to_apply.csv" + }, + { + "value": "Ji: LoRA-5 3.4 min (fine-tune 1.8 + generate 1.6), scSemiProfiler-5 57 min (representatives 24 + per-sample models 33); 11\u201320\u00d7 per draw; per test sample 8.1 s vs 137 s", + "source": "data/sec09_time_to_apply.csv" + }, + { + "value": "PBISC one-time pretraining 15.9 h (not counted per draw)", + "source": "data/sec06_ft_method.csv (pretrain_wall_hours)" + } + ], + "refs": [ + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "original_slide": 34, + "original_title": "5 donors: time to apply", + "original_footnote": "One GB10 GPU; mean of 5 draws. Data prep and scoring not counted. PBISC pretraining (15.9 h) is done once.", + "figures_before_palette": [ + "figures/sec09_time_to_apply.png" + ], + "source_page_131": 59, + "deck_group": "Supp S06", + "previous_section_label": "Fig. 3d", + "parent_page": 77, + "parent_enriched_page": 86, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_time_to_apply.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 88, + "parent_154_page": 100, + "parent_155_page": 101 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "title": "Pairs needed: composition, coverage, training steps", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec08_pairs_realism.png" + ], + "keywords": [ + "No pairs: coverage 0.19 / 0.07; length fix 0.50 / 0.40", + "Calibration flat 0.53 / 0.43; scSemi 0.04 / 0.10", + "LoRA with 5 pairs: coverage 0.84 / 0.77" + ], + "footnote": "ImMicro / sepsis, 5-fold CV (107 / 82 donors). 0 = no pairs; squares = calibration fitted on the same n pairs.", + "footnote_full": "Original slide 65: ImMicro / sepsis, donor-level 5-fold CV, pooled scoring (107 / 82 donors). x = LoRA training pairs per fold; each run also early-stops on 7\u20138 paired inner-val donors; the last point is every training donor of the fold (78\u201379 / 58\u201359; the other four folds minus inner-val, never the test fold). x = 0 uses no paired data: the base model fed the real bulk as measured (pale) and after exon-length correction (sand: counts divided by Ensembl union-exon kb before the CPM step; scores//zeroshot_tpm_20261002/pooled). Squares: zero-shot + gene-wise calibration with alpha fitted on exactly the pairs that LoRA point saw (cal_budget_20261001; the last square uses every training pair of the fold). Coverage ImMicro / sepsis: as measured 0.19 / 0.07, length-corrected 0.50 / 0.40, calibration 0.52\u20130.53 / 0.43\u20130.44 at every n, LoRA 5 pairs 0.84 / 0.77, every training donor 0.78 / 0.83. Composition L1: length correction does not fix it (ImMicro 0.45 \u2192 0.42, sepsis 0.51 \u2192 0.56). Diamonds (ImMicro / sepsis): scSemiProfiler with exactly LoRA's n training + inner-val pairs as representatives (scores//scsemi_budget_20261002/pooled), and with every training donor at the last point (scsemi_cv_20261002/pooled); same donors and reference as the LoRA points. scSemiProfiler coverage 0.04 / 0.10 and composition L1 0.22\u20130.23 / 0.52\u20130.54 at n = 5\u201340 (0.22 / 0.50 with every training donor).", + "say": "Each LoRA run also early-stops on 7 inner-validation pairs (8 in 2 of the 5 sepsis folds).", + "sources": "tools/sec08_pairs_needed.py; data/sec08_pairs_realism.csv, data/sec08_selected_steps.csv, data/sec08_slide_numbers.csv", + "refs": [ + "Kynk\u00e4\u00e4nniemi 2019, NeurIPS", + "Naeem 2020, ICML" + ], + "refs_full": [ + "Kynk\u00e4\u00e4nniemi et al. Improved precision and recall metric for assessing generative models. NeurIPS 2019", + "Naeem et al. Reliable fidelity and diversity metrics for generative models. ICML 2020" + ], + "original_slide": 65, + "original_title": "Pairs needed: composition, coverage, training steps", + "original_footnote": "ImMicro / sepsis, 5-fold CV (107 / 82 donors). 0 = no pairs; squares = calibration fitted on the same n pairs.", + "figures_before_palette": [ + "figures/sec08_pairs_realism.png" + ], + "source_page_131": 117, + "deck_group": "Supp S06", + "previous_section_label": "Supplementary", + "parent_page": 78, + "parent_enriched_page": 87, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec08_pairs_realism.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 89, + "parent_154_page": 101, + "parent_155_page": 102 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "text_source": { + "file": "data/sec07_rank_sweep_slide_text.json", + "index": 0 + }, + "title": "LoRA rank: r = 4, 16, 64", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec07_rank_sweep.png" + ], + "keywords": [ + "ImMicro r64 vs r16: coverage 0.78 \u2192 0.82", + "ImMicro r4 vs r16: L1 0.162 \u2192 0.186", + "Kept r16; r64 precision 0.92 \u2192 0.90" + ], + "footnote": "Sepsis, 58\u201359 pairs/fold: r64 logFC r 0.53 \u2192 0.55, DEG 23 \u2192 26; r4 0.48, DEG 18. Pooled 5-fold scores.", + "footnote_full": "Original slide 66: LoRA fine-tune of the final model (f01 6k\u00b71x\u00b7E16) with rank r and alpha = r, same recipe as the deck's r = 16 runs (early stop on 7\u20138 inner-val pairs, 20k step cap); n = training pairs, nested subsets; donor-level 5-fold CV, each pair scored once by its fold model (pooled call). Trainable parameters r4 324,056, r16 1,296,224, r64 5,184,896. ImMicro n = 5: r4: L1 0.162, P/C 0.937 / 0.789 (L1 vs r16 -0.009, Wilcoxon p 0.00088, 107 donors); r16: L1 0.171, P/C 0.946 / 0.839; r64: L1 0.177, P/C 0.946 / 0.847 (L1 vs r16 +0.005, Wilcoxon p 0.0076, 107 donors) | ImMicro n = 20: r4: L1 0.171, P/C 0.925 / 0.715 (L1 vs r16 -0.001, Wilcoxon p 0.32, 107 donors); r16: L1 0.172, P/C 0.920 / 0.755; r64: L1 0.172, P/C 0.905 / 0.794 (L1 vs r16 -0.001, Wilcoxon p 0.66, 107 donors) | ImMicro n = 78\u201379 (every training donor of the fold): r4: L1 0.186, P/C 0.933 / 0.712 (L1 vs r16 +0.024, Wilcoxon p 9.8e-09, 107 donors); r16: L1 0.162, P/C 0.924 / 0.784; r64: L1 0.161, P/C 0.904 / 0.816 (L1 vs r16 -0.001, Wilcoxon p 0.89, 107 donors) | Sepsis n = 5: r4: L1 0.426, P/C 0.907 / 0.742, logFC r 0.27, DEG 2 (L1 vs r16 -0.002, Wilcoxon p 0.97, 82 donors); r16: L1 0.428, P/C 0.911 / 0.767, logFC r 0.29, DEG 5; r64: L1 0.429, P/C 0.906 / 0.766, logFC r 0.31, DEG 1 (L1 vs r16 +0.001, Wilcoxon p 0.66, 82 donors) | Sepsis n = 20: r4: L1 0.378, P/C 0.929 / 0.769, logFC r 0.31, DEG 14 (L1 vs r16 -0.003, Wilcoxon p 0.58, 82 donors); r16: L1 0.380, P/C 0.927 / 0.807, logFC r 0.37, DEG 18; r64: L1 0.376, P/C 0.911 / 0.808, logFC r 0.43, DEG 18 (L1 vs r16 -0.004, Wilcoxon p 0.046, 82 donors) | Sepsis n = 58\u201359 (every training donor of the fold): r4: L1 0.363, P/C 0.923 / 0.789, logFC r 0.48, DEG 18 (L1 vs r16 +0.002, Wilcoxon p 0.77, 82 donors); r16: L1 0.362, P/C 0.926 / 0.829, logFC r 0.53, DEG 23; r64: L1 0.367, P/C 0.907 / 0.824, logFC r 0.55, DEG 26 (L1 vs r16 +0.005, Wilcoxon p 0.66, 82 donors). r64 vs r16 brings small gains in coverage and disease signal but slightly lower precision and no composition change; r4 is worse with many pairs; at n = 5 the ranks are mixed. r = 16 is kept. Scores: scores//lora_rank_20261001/pooled (r4, r64) and scores//pooled{,_deg} (r16); table: scores//rankcal_20261001.md.", + "sources": "tools/sec07_rank_sweep.py; data/sec07_rank_sweep.csv, data/sec07_rank_sweep_runs.csv; plans/paired_finetune_intradataset_20260927/scores//rankcal_20261001.md", + "numbers": [ + { + "value": "ImMicro 78\u201379 pairs/fold coverage r16 0.784 -> r64 0.816; precision 0.924 -> 0.904; L1 r16 0.162, r4 0.186 (Wilcoxon p 9.8e-09)", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/pooled/summary.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/immicro/lora_rank_20261001/pooled/summary.json, per_donor.csv" + }, + { + "value": "Sepsis 58\u201359 pairs/fold logFC r r16 0.527, r64 0.553, r4 0.476; DEG 23 / 26 / 18", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/pooled_deg/summary.json, contrast.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/lora_rank_20261001/pooled/summary.json, contrast.csv" + }, + { + "value": "trainable parameters: r4 324056, r16 1296224, r64 5184896", + "source": "runs//fold//RUN.json (trainable.n_trainable)" + } + ], + "refs": [ + "Hu 2022, ICLR" + ], + "refs_full": [ + "Hu et al. LoRA: Low-Rank Adaptation of Large Language Models. ICLR 2022" + ], + "original_slide": 66, + "original_title": "LoRA rank: r = 4, 16, 64", + "original_footnote": "Sepsis, 58\u201359 pairs/fold: r64 logFC r 0.53 \u2192 0.55, DEG 23 \u2192 26; r4 0.48, DEG 18. Pooled 5-fold scores.", + "figures_before_palette": [ + "figures/sec07_rank_sweep.png" + ], + "source_page_131": 118, + "deck_group": "Supp S06", + "previous_section_label": "Supplementary", + "parent_page": 79, + "parent_enriched_page": 88, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec07_rank_sweep.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 90, + "parent_154_page": 102, + "parent_155_page": 103 + }, + { + "kind": "content", + "section": "Supp S06", + "section_label": "Supp. S06", + "title": "K diversification: no gain", + "layout": "figure_keywords", + "figures": [ + "figures/final_palette_20261013/sec07_k_aug.png" + ], + "keywords": [ + "No gain; keep single K 1000", + "Multi-K slightly worse on ImMicro (p\u00a00.04)", + "Inner-val loss \u0394 \u2264 1.2e-4" + ], + "footnote": "LoRA r16, 5-fold pooled. p = paired Wilcoxon on donor L1, uncorrected.", + "sources": "tools/sec07_k_aug.py; data/sec07_k_aug.csv, data/sec07_k_aug_wilcoxon.csv, data/sec07_k_aug_inner_val_delta.csv", + "footnote_full": "Original slide 67: LoRA r16 on f01_6k_1x E16, 5-fold pooled. p: paired Wilcoxon, donor L1 vs plain LoRA, uncorrected. Recipes arm: K 1000 real bulk + \u00bd-batch pretraining-style pseudobulk recipes made from the training donors' own cells (recipe K drawn from the pretraining grid 250\u20134,000; RUN.json augmentation); the real bulk itself is not multi-K.", + "original_slide": 67, + "original_title": "K diversification: no gain", + "original_footnote": "LoRA r16, 5-fold pooled. p = paired Wilcoxon on donor L1, uncorrected.", + "figures_before_palette": [ + "figures/sec07_k_aug.png" + ], + "source_page_131": 119, + "deck_group": "Supp S06", + "previous_section_label": "Supplementary", + "parent_page": 80, + "parent_enriched_page": 89, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec07_k_aug.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 91, + "parent_154_page": 103, + "parent_155_page": 104 + }, + { + "kind": "content", + "section": "Supp S07", + "section_label": "Supp. S07", + "sec09_index": 0, + "title": "Five paired cohorts", + "layout": "full", + "figures": [ + "figures/sec09_cohorts.png" + ], + "keywords": [ + "403 paired samples, 275 donors", + "Identity screening where available; donor-level splits", + "Ji bulk: whole blood, not PBMC" + ], + "footnote": "Existing paired-cohort map. Five donors means 5 pairs in Sepsis, 12-15 Imperial, 12-14 Giroux, 7-9 Ji.", + "footnote_full": "Current scope: Existing paired-cohort map. Five donors means 5 pairs in Sepsis, 12-15 Imperial, 12-14 Giroux, 7-9 Ji.\n\nOriginal slide 37: Usable pair = passed the pairing identity check and >= 200 cells (fold >= 0 in data//folds.csv); dropped 15 identity-fail, 11 borderline and 1 low-cell pairs of 430. ImMicro: no identity flag, all used. Imperial, Giroux, Ji: several pairs per donor. Case vs control = samples in scores//pooled/summary.json. Giroux 'moderate' = GEO 'severe' (paper term); its moderate vs mild subjects are 4F/0M vs 0F/5M (scored), so its disease metrics exclude sex-linked genes. Labels: data//QC_REPORT.md; ImMicro: GEO GSE314922/GSE314416.", + "sources": "tools/sec09_paired_validation.py; data/sec09_cohorts.csv, data/sec09_quoted_numbers.csv, data/sec09_slide_text.json", + "numbers": [ + { + "value": "403 paired samples (sum of usable rows)", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data//fewshot_n5/DRAWS.json (n_usable_rows)" + }, + { + "value": "275 donors (sum of usable donors)", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data//fewshot_n5/DRAWS.json (n_usable_donors)" + }, + { + "value": "dropped 15 identity-fail, 11 borderline, 1 < 200 cells, of 430", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data//features.npz (identity_pass, identity_check); samples.csv (n_cells); /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/prepare_dataset.py (MIN_CELLS)" + }, + { + "value": "ImMicro 107 pairs / 107 donors", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/immicro/fewshot_n5/DRAWS.json" + }, + { + "value": "Sepsis 82 pairs / 82 donors; sepsis 61 vs control 21", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis/fewshot_n5/DRAWS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/pooled/summary.json" + }, + { + "value": "Imperial 157 pairs / 57 donors; COVID+ 113 vs COVID\u2212 34 (+10 recovery, neither)", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/imperial_eskd/fewshot_n5/DRAWS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/imperial_eskd/pooled/summary.json" + }, + { + "value": "Giroux 24 pairs / 9 donors; moderate 11 vs mild 13", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/giroux2022/fewshot_n5/DRAWS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/giroux2022/pooled/summary.json" + }, + { + "value": "Ji 33 pairs / 20 donors; ChILI 18 vs ICI control 15", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ji2025_chili/fewshot_n5/DRAWS.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/ji2025_chili/pooled/summary.json" + } + ], + "original_slide": 37, + "original_title": "Five paired cohorts", + "original_footnote": "Usable pair = passed identity check and \u2265 200 cells: 403 of 430 pairs; donor-level splits.", + "figures_before_palette": [ + "figures/sec09_cohorts.png" + ], + "source_page_131": 62, + "deck_group": "Supp S07", + "previous_section_label": "Fig. 4a", + "parent_page": 81, + "parent_enriched_page": 90, + "figures_before_final_palette": [ + "figures/sec09_cohorts.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 92, + "parent_154_page": 104, + "parent_155_page": 105 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_strict_sepsis_quality", + "title": "Sepsis: exact-five reconstruction across five draws", + "section_label": "Supp. S07", + "figures": [ + "figures/final_palette_20261013/strict_sepsis_quality.png" + ], + "keywords": [], + "footnote": "Completed exact-five Sepsis draws. Bars = means; dots = support draws. Five matched methods; no new generation.", + "footnote_full": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Precision, coverage and composition are read directly from completed score summaries.\nSources: plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_quality.csv\nEvidence source: plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_quality.csv", + "sources": "plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_quality.csv", + "notes": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Precision, coverage and composition are read directly from completed score summaries.\nSources: plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_quality.csv", + "after_original": 38, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_strict_sepsis_quality", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/strict_sepsis_quality.png" + ], + "source_page_131": 64, + "deck_group": "Supp S07", + "section": "Supp S07", + "previous_section_label": "Fig. 4 update", + "parent_page": 82, + "parent_enriched_page": 91, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/strict_sepsis_quality.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 93, + "parent_154_page": 105, + "parent_155_page": 106 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_strict_sepsis_realism", + "title": "Sepsis: realism beyond precision and coverage", + "section_label": "Supp. S07", + "figures": [ + "figures/final_palette_20261013/strict_sepsis_realism.png" + ], + "keywords": [], + "footnote": "Completed exact-five outputs; raw readout, fixed R1 reference. Real/real control shown. Dots are five support draws.", + "footnote_full": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. C2ST uses cached kNN AUROC; MMD uses bandwidth 1; E_N uses the saved PC subsample, squared distance and exact lower bound. No refitting.\nSources: plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/en.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_realism.csv\nEvidence source: plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/en.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_realism.csv", + "sources": "plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/en.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_realism.csv", + "notes": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. C2ST uses cached kNN AUROC; MMD uses bandwidth 1; E_N uses the saved PC subsample, squared distance and exact lower bound. No refitting.\nSources: plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/extras/ye2026_sepsis/strict/en.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/metrics.csv; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/extras/ye2026_sepsis/strict/en.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_sepsis_realism.csv", + "after_original": 38, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_strict_sepsis_realism", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/strict_sepsis_realism.png" + ], + "source_page_131": 65, + "deck_group": "Supp S07", + "section": "Supp S07", + "previous_section_label": "Fig. 4 update", + "parent_page": 83, + "parent_enriched_page": 92, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/strict_sepsis_realism.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 94, + "parent_154_page": 106, + "parent_155_page": 107 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_strict_imperial_quality", + "title": "Imperial: first completed exact-five comparison", + "section_label": "Supp. S07", + "figures": [ + "figures/final_palette_20261013/strict_imperial_draw0_quality.png" + ], + "keywords": [], + "footnote": "One completed support draw (draw 0). Exact five paired samples; all five methods scored. Not the final five-draw comparison.", + "footnote_full": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Single-draw evidence is shown separately from the complete Sepsis set. Repeated held-out visits remain linked to actual donor IDs.\nSources: plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/protocol/splits/imperial_eskd/disease/draw0/support.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_draw0_quality.csv\nEvidence source: plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/protocol/splits/imperial_eskd/disease/draw0/support.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_draw0_quality.csv", + "sources": "plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/protocol/splits/imperial_eskd/disease/draw0/support.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_draw0_quality.csv", + "notes": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Single-draw evidence is shown separately from the complete Sepsis set. Repeated held-out visits remain linked to actual donor IDs.\nSources: plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/protocol/splits/imperial_eskd/disease/draw0/support.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_draw0_quality.csv", + "after_original": 38, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_strict_imperial_quality", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/strict_imperial_draw0_quality.png" + ], + "source_page_131": 66, + "deck_group": "Supp S07", + "section": "Supp S07", + "previous_section_label": "Fig. 4 update", + "parent_page": 84, + "parent_enriched_page": 93, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/strict_imperial_draw0_quality.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 95, + "parent_154_page": 107, + "parent_155_page": 108 + }, + { + "kind": "content", + "section": "Supp S07", + "section_label": "Supp. S07", + "sec09_index": 1, + "title": "5 donors: LoRA vs scSemiProfiler \u2014 cells", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_fewpair_cells.png" + ], + "keywords": [ + "LoRA-5 > scSemi-5: coverage, precision 5/5", + "Beats zero-shot + calibration: coverage 5/5", + "Lower L1 than calibration: ImMicro only" + ], + "footnote": "Completed five-donor comparisons. Multi-visit cohorts used more than five paired samples.", + "footnote_full": "Current scope: Completed five-donor comparisons. Multi-visit cohorts used more than five paired samples.\n\nOriginal slide 58: Base f01 6k\u00b71x\u00b7E16 (EMA, DDIM 50) + LoRA r=16, 3,000 steps. LoRA-5 / scSemi-5 = LoRA / scSemiProfiler trained on the same 5 donors (5\u201315 pairs); test = all other usable pairs. Zero-shot = base model without fine-tuning, fed each test pair's real bulk as measured; zero-shot + calibration = the same model fed that bulk after gene-wise calibration, alpha fitted on the same 5 training donors' pairs (both in the same call as LoRA-5). scSemi-5 scored in a separate call, same test set and reference (sha-checked); 25/25 draws. Verdict = better in >= 80% of shared draws. LoRA-5 lower L1 than scSemi-5 in 4/5 (Giroux split 3/5). LoRA-5 > zero-shot + calibration: coverage 5/5, precision 3/5, depth-matched 5/5 and 3/5 cohorts; > zero-shot (real bulk): coverage 5/5, precision 5/5, depth-matched 5/5 and 4/5. Depth-matched P/C vs scSemi-5: same ranking, coverage LoRA-5 0.83\u20130.97 vs scSemi-5 0.15\u20130.28. Composition L1, LoRA-5 minus zero-shot + calibration, per-donor Wilcoxon (gain = p < 0.05): ImMicro -0.03 (p 2e-10), Sepsis +0.02 (p 0.79), Imperial +0.02 (p 0.38), Giroux -0.06 (p 0.38), Ji -0.01 (p 0.70); minus zero-shot (real bulk): ImMicro -0.27 (p 2e-18), Sepsis -0.04 (p 0.15), Imperial -0.35 (p 2e-08), Giroux -0.11 (p 0.04), Ji -0.23 (p 0.02). scANVI labels; P/C = PCA-50, k=10 balls. Scores: scores//fewshot_n5/.", + "sources": "tools/sec09_paired_validation.py; data/sec09_fewpair_cells.csv, data/sec09_scsemi_status.csv, data/sec09_slide_text.json", + "numbers": [ + { + "value": "LoRA-5 vs scSemi-5 coverage (shared-draw means; LoRA-better draws/n; verdict): ImMicro 0.792 vs 0.036 (5/5, lora); Sepsis 0.694 vs 0.080 (5/5, lora); Imperial 0.737 vs 0.059 (5/5, lora); Giroux 0.961 vs 0.607 (5/5, lora); Ji 0.707 vs 0.145 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json and draw_scsemi5/summary.json" + }, + { + "value": "LoRA-5 vs scSemi-5 precision (shared-draw means; LoRA-better draws/n; verdict): ImMicro 0.952 vs 0.518 (5/5, lora); Sepsis 0.906 vs 0.728 (5/5, lora); Imperial 0.853 vs 0.330 (5/5, lora); Giroux 1.000 vs 0.992 (5/5, lora); Ji 0.978 vs 0.840 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json and draw_scsemi5/summary.json" + }, + { + "value": "LoRA-5 vs scSemi-5 comp_l1 (shared-draw means; LoRA-better draws/n; verdict): ImMicro 0.176 vs 0.233 (5/5, lora); Sepsis 0.466 vs 0.612 (5/5, lora); Imperial 0.503 vs 0.602 (5/5, lora); Giroux 0.256 vs 0.270 (3/5, split); Ji 0.513 vs 0.592 (4/5, lora) -> LoRA verdicts 4/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json and draw_scsemi5/summary.json" + }, + { + "value": "LoRA-5 vs scSemi-5 coverage_dm (shared-draw means; LoRA-better draws/n; verdict): ImMicro 0.966 vs 0.284 (5/5, lora); Sepsis 0.834 vs 0.158 (5/5, lora); Imperial 0.863 vs 0.152 (5/5, lora); Giroux 0.974 vs 0.238 (5/5, lora); Ji 0.908 vs 0.150 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json and draw_scsemi5/summary.json" + }, + { + "value": "LoRA-5 vs scSemi-5 precision_dm (shared-draw means; LoRA-better draws/n; verdict): ImMicro 0.975 vs 0.829 (5/5, lora); Sepsis 0.914 vs 0.504 (5/5, lora); Imperial 0.881 vs 0.429 (5/5, lora); Giroux 0.995 vs 0.426 (5/5, lora); Ji 0.974 vs 0.564 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json and draw_scsemi5/summary.json" + }, + { + "value": "LoRA-5 vs zero-shot + calibration coverage (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.792 vs 0.535 (5/5, lora); Sepsis 0.694 vs 0.428 (5/5, lora); Imperial 0.737 vs 0.634 (5/5, lora); Giroux 0.961 vs 0.944 (5/5, lora); Ji 0.707 vs 0.114 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_cal)" + }, + { + "value": "LoRA-5 vs zero-shot + calibration precision (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.952 vs 0.904 (5/5, lora); Sepsis 0.906 vs 0.899 (3/5, split); Imperial 0.853 vs 0.845 (3/5, split); Giroux 1.000 vs 0.996 (4/5, lora); Ji 0.978 vs 0.782 (5/5, lora) -> LoRA verdicts 3/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_cal)" + }, + { + "value": "LoRA-5 vs zero-shot + calibration comp_l1 (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.176 vs 0.203 (5/5, lora); Sepsis 0.466 vs 0.446 (2/5, split); Imperial 0.503 vs 0.506 (2/5, split); Giroux 0.256 vs 0.288 (2/5, split); Ji 0.513 vs 0.514 (3/5, split) -> LoRA verdicts 1/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_cal)" + }, + { + "value": "LoRA-5 vs zero-shot + calibration coverage_dm (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.966 vs 0.780 (5/5, lora); Sepsis 0.834 vs 0.658 (5/5, lora); Imperial 0.863 vs 0.681 (5/5, lora); Giroux 0.974 vs 0.937 (5/5, lora); Ji 0.908 vs 0.357 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_cal)" + }, + { + "value": "LoRA-5 vs zero-shot + calibration precision_dm (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.975 vs 0.940 (5/5, lora); Sepsis 0.914 vs 0.909 (3/5, split); Imperial 0.881 vs 0.868 (3/5, split); Giroux 0.995 vs 0.989 (5/5, lora); Ji 0.974 vs 0.658 (5/5, lora) -> LoRA verdicts 3/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_cal)" + }, + { + "value": "LoRA-5 vs zero-shot (real bulk) coverage (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.792 vs 0.190 (5/5, lora); Sepsis 0.694 vs 0.076 (5/5, lora); Imperial 0.737 vs 0.114 (5/5, lora); Giroux 0.961 vs 0.472 (5/5, lora); Ji 0.707 vs 0.066 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_raw)" + }, + { + "value": "LoRA-5 vs zero-shot (real bulk) precision (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.952 vs 0.677 (5/5, lora); Sepsis 0.906 vs 0.625 (5/5, lora); Imperial 0.853 vs 0.500 (5/5, lora); Giroux 1.000 vs 0.980 (5/5, lora); Ji 0.978 vs 0.799 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_raw)" + }, + { + "value": "LoRA-5 vs zero-shot (real bulk) comp_l1 (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.176 vs 0.445 (5/5, lora); Sepsis 0.466 vs 0.510 (5/5, lora); Imperial 0.503 vs 0.806 (5/5, lora); Giroux 0.256 vs 0.351 (5/5, lora); Ji 0.513 vs 0.721 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_raw)" + }, + { + "value": "LoRA-5 vs zero-shot (real bulk) coverage_dm (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.966 vs 0.213 (5/5, lora); Sepsis 0.834 vs 0.073 (5/5, lora); Imperial 0.863 vs 0.151 (5/5, lora); Giroux 0.974 vs 0.549 (5/5, lora); Ji 0.908 vs 0.199 (5/5, lora) -> LoRA verdicts 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_raw)" + }, + { + "value": "LoRA-5 vs zero-shot (real bulk) precision_dm (same-draw means; LoRA-better draws/n; verdict): ImMicro 0.975 vs 0.671 (5/5, lora); Sepsis 0.914 vs 0.595 (5/5, lora); Imperial 0.881 vs 0.561 (5/5, lora); Giroux 0.995 vs 0.996 (2/5, split); Ji 0.974 vs 0.480 (5/5, lora) -> LoRA verdicts 4/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json (arms lora5, zs_raw)" + }, + { + "value": "coverage cohort means over all draws: LoRA-5 ImMicro 0.792, Sepsis 0.694, Imperial 0.737, Giroux 0.961, Ji 0.707; scSemi-5 ImMicro 0.036, Sepsis 0.080, Imperial 0.059, Giroux 0.607, Ji 0.145; zero-shot + calibration ImMicro 0.535, Sepsis 0.428, Imperial 0.634, Giroux 0.944, Ji 0.114; zero-shot ImMicro 0.190, Sepsis 0.076, Imperial 0.114, Giroux 0.472, Ji 0.066", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/summary.json and draw_scsemi5/summary.json" + }, + { + "value": "LoRA-5 minus zero-shot + calibration composition L1, per-donor Wilcoxon (draw-averaged): ImMicro -0.027, n=107, p=2.2e-10; Sepsis +0.018, n=82, p=0.79; Imperial +0.016, n=57, p=0.38; Giroux -0.065, n=8, p=0.38; Ji -0.012, n=20, p=0.7 -> gain in 1/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/per_donor.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data//folds.csv" + }, + { + "value": "LoRA-5 minus zero-shot (real bulk) composition L1, per-donor Wilcoxon (draw-averaged): ImMicro -0.270, n=107, p=2.2e-18; Sepsis -0.044, n=82, p=0.15; Imperial -0.348, n=57, p=1.6e-08; Giroux -0.107, n=8, p=0.039; Ji -0.232, n=20, p=0.017 -> gain in 4/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores//fewshot_n5/draw/per_donor.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data//folds.csv" + }, + { + "value": "scSemiProfiler draws used / LoRA draws: ImMicro 5/5, Sepsis 5/5, Imperial 5/5, Giroux 5/5, Ji 5/5", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec09_scsemi_status.csv" + }, + { + "value": "training pairs per draw: ImMicro 5, Sepsis 5, Imperial 12\u201315, Giroux 12\u201314, Ji 7\u20139; test pairs: ImMicro 102, Sepsis 77, Imperial 142\u2013145, Giroux 10\u201312, Ji 24\u201326", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data//fewshot_n5/DRAWS.json (n_train_rows, n_test_rows)" + }, + { + "value": "LoRA r = 16; 3000 fixed steps; base runs/f01_6k_1x_seed20260627/checkpoint_epoch_016.pt; sampling: test donors only, 1,500 cells each, seed 20260927, DDIM 50, EMA, Bernoulli read-out", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/runs/immicro/fewshot_n5/draw0/lora/RUN.json; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/README.md" + }, + { + "value": "P/C definition: precision/coverage/density (PCA-50 + k = 10 balls fitted on the scored donors' observed R1 cells); PCA components in every summary.json = 50", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/score_ft.py" + } + ], + "refs": [ + "Kynk\u00e4\u00e4nniemi 2019, NeurIPS", + "Naeem 2020, ICML", + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Kynk\u00e4\u00e4nniemi et al. Improved precision and recall metric for assessing generative models. NeurIPS 2019", + "Naeem et al. Reliable fidelity and diversity metrics for generative models. ICML 2020", + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "original_slide": 58, + "original_title": "5 donors: LoRA vs scSemiProfiler \u2014 cells", + "original_footnote": "5 draws per cohort; test = all other usable pairs. Calibration \u03b1 from the same 5 donors. scANVI labels.", + "figures_before_palette": [ + "figures/sec09_fewpair_cells.png" + ], + "source_page_131": 110, + "deck_group": "Supp S07", + "previous_section_label": "Supplementary", + "parent_page": 85, + "parent_enriched_page": 94, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_fewpair_cells.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 96, + "parent_154_page": 108, + "parent_155_page": 109 + }, + { + "kind": "content", + "section": "Supp S07", + "section_label": "Supp. S07", + "text_source": { + "file": "data/sec09_extra_slide_text.json", + "index": 0 + }, + "title": "5 donors: realism beyond precision/coverage", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_extra_fewpair.png" + ], + "keywords": [ + "Classifier AUC: LoRA-5 best in 25/25 draws", + "E_N: LoRA-5 below scSemi-5 in 25/25 draws", + "Still separable: LoRA-5 AUC 0.89\u20130.95" + ], + "footnote": "E_N: scSemiProfiler's own error metric. Misses only vs calibration, e.g. Imperial E_N 1.28 vs 1.22.", + "footnote_full": "Original slide 59: Same 5-donor draws as the previous slides; every metric on exactly the scored cells (scores_extra/, tools/extra_metrics.py). Real side for AUC and MMD\u00b2 = held-out real cells of the same samples (R2), projected like the generated cells; real vs real = R2 vs R3 (AUC 0.49\u20130.50). Classifier AUC: 5-fold CV ROC AUC of a kNN classifier (k = 5) separating real from generated cells, 0.5 = indistinguishable; bars start at 0.5. MMD\u00b2: unbiased, Gaussian kernel, bandwidth = median distance among real cells. E_N: scSemiProfiler's normalized error (Wang et al. 2024, eq. 17\u201318), their code's squared L2 on all cells, per-arm PCA-100; real vs real = within-sample half split (0.07\u20130.38); every generated arm is above 1. LoRA-5 best of the 4 arms in 25/25 draws on kNN AUC, 25/25 on RF AUC, 25/25 on 1-NN accuracy, 23/25 on MMD\u00b2, 22/25 on E_N. LoRA-5 better than scSemi-5, zero-shot and zero-shot + calibration on both AUCs, 1-NN accuracy and coverage at k = 3, 5, 10 in every draw. E_N scSemi-5 vs LoRA-5 (cohort means): ImMicro 3.65 vs 1.55; Sepsis 3.83 vs 1.46; Imperial 2.04 vs 1.28; Giroux 3.79 vs 1.99; Ji 4.27 vs 2.60; LoRA-5 lower in 25/25 draws. Draws where LoRA-5 is worse (cohort means in brackets): vs zero-shot + calibration: Sepsis precision k5 1/5 (0.778 vs 0.752), precision k10 2/5 (0.906 vs 0.899); Imperial MMD\u00b2 (\u03c3 = m) 1/5 (0.0156 vs 0.0218), MMD\u00b2 (\u03c3 sum) 1/5 (0.0443 vs 0.0673), E_N (all cells) 3/5 (1.28 vs 1.22), precision k3 2/5 (0.610 vs 0.598), precision k5 2/5 (0.726 vs 0.719), precision k10 2/5 (0.853 vs 0.845); Giroux MMD\u00b2 (\u03c3 = m) 1/5 (0.0178 vs 0.0257), MMD\u00b2 (\u03c3 sum) 1/5 (0.0584 vs 0.0838), precision k5 1/5 (0.995 vs 0.981), precision k10 1/5 (1.000 vs 0.996). LoRA-5 AUC per draw 0.88\u20130.97 (kNN), RF cohort means 0.90\u20130.96: a classifier still tells generated from real cells. Scores: scores_extra//fewshot_n5/draw/.", + "sources": "tools/sec09_extra_metrics.py; data/sec09_extra_fewpair.csv, data/sec09_extra_quoted_numbers.csv; plans/paired_finetune_intradataset_20260927/scores_extra/ (README.md, TABLES.md)", + "numbers": [ + { + "value": "LoRA-5 best of 4 arms (draws): kNN AUC 25/25, RF AUC 25/25, 1-NN 25/25, MMD\u00b2 23/25, E_N 22/25", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + }, + { + "value": "E_N LoRA-5 < scSemi-5 in 25/25 draws; cohort means scSemi-5 vs LoRA-5: ImMicro 3.65 vs 1.55; Sepsis 3.83 vs 1.46; Imperial 2.04 vs 1.28; Giroux 3.79 vs 1.99; Ji 4.27 vs 2.60", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/en.csv (cellset all, distance sq)" + }, + { + "value": "LoRA-5 kNN AUC cohort means: ImMicro 0.943, Sepsis 0.936, Imperial 0.918, Giroux 0.893, Ji 0.955; per draw 0.878\u20130.975", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + }, + { + "value": "real vs real: kNN AUC ImMicro 0.502, Sepsis 0.501, Imperial 0.498, Giroux 0.495, Ji 0.491; E_N ImMicro 0.382, Sepsis 0.100, Imperial 0.260, Giroux 0.145, Ji 0.073", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + }, + { + "value": "LoRA-5 worse than a comparator (draws): vs zero-shot + calibration: Sepsis precision k5 1/5 (0.778 vs 0.752), precision k10 2/5 (0.906 vs 0.899); Imperial MMD\u00b2 (\u03c3 = m) 1/5 (0.0156 vs 0.0218), MMD\u00b2 (\u03c3 sum) 1/5 (0.0443 vs 0.0673), E_N (all cells) 3/5 (1.28 vs 1.22), precision k3 2/5 (0.610 vs 0.598), precision k5 2/5 (0.726 vs 0.719), precision k10 2/5 (0.853 vs 0.845); Giroux MMD\u00b2 (\u03c3 = m) 1/5 (0.0178 vs 0.0257), MMD\u00b2 (\u03c3 sum) 1/5 (0.0584 vs 0.0838), precision k5 1/5 (0.995 vs 0.981), precision k10 1/5 (1.000 vs 0.996)", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + }, + { + "value": "cross-check: 720 cohort \u00d7 arm \u00d7 metric values equal scores_extra/{fewshot,pooled}_headline.csv", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra/fewshot_headline.csv; /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra/pooled_headline.csv" + } + ], + "refs": [ + "Lopez-Paz 2017, ICLR", + "Gretton 2012, JMLR", + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Lopez-Paz & Oquab. Revisiting classifier two-sample tests. ICLR 2017", + "Gretton et al. A kernel two-sample test. JMLR 13:723-773 (2012)", + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "original_slide": 59, + "original_title": "5 donors: realism beyond precision/coverage", + "original_footnote": "E_N: scSemiProfiler's own error metric. Misses only vs calibration, e.g. Imperial E_N 1.28 vs 1.22.", + "figures_before_palette": [ + "figures/sec09_extra_fewpair.png" + ], + "source_page_131": 111, + "deck_group": "Supp S07", + "previous_section_label": "Supplementary", + "parent_page": 86, + "parent_enriched_page": 95, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_extra_fewpair.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 97, + "parent_154_page": 109, + "parent_155_page": 110 + }, + { + "kind": "content", + "section": "Supp S07", + "section_label": "Supp. S07", + "text_source": { + "file": "data/sec09_extra_slide_text.json", + "index": 2 + }, + "title": "5 donors: precision/coverage at k = 3, 5, 10", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_extra_ksens.png" + ], + "keywords": [ + "Smaller k lowers every set, real included", + "Coverage: LoRA-5 first at all k, 25/25", + "Precision: LoRA-5 first in 20\u201323/25 draws" + ], + "footnote": "Completed sensitivity at k=3,5,10; no new sensitivity campaign required.", + "footnote_full": "Current scope: Completed sensitivity at k=3,5,10; no new sensitivity campaign required.\n\nOriginal slide 60: Precision and coverage recomputed with k = 3, 5, 10 nearest-neighbour balls in the scorer's PCA-50 (k = 10 reproduces the main-slide values exactly). Means over the 5 draws per cohort; error bars = SD; grey = real vs real (R1 vs R2, same samples). Both metrics fall with smaller k for every set including the real control (real coverage at k = 3: 0.92\u20130.99), so compare arms with the real row, not with 1. Expected coverage of a perfect generator (Naeem et al. 2020, eq. 9) at k = 3/5/10: 0.88/0.97/1.00 in every cohort. LoRA-5 first on coverage at every k in 25/25 draws; on precision in 23/21/20 of 25 draws (k = 3/5/10); not first: Sepsis 0/1/2, Imperial 2/2/2, Giroux 0/1/1 draws at k = 3/5/10, beaten by zero-shot + calibration. 5-fold pooled design (not drawn): coverage ranking of the 5 arms across k unchanged; LoRA first at every k in 5/5 cohorts. Scores: scores_extra//fewshot_n5/draw/metrics.csv.", + "sources": "tools/sec09_extra_metrics.py; data/sec09_extra_ksens.csv; plans/paired_finetune_intradataset_20260927/scores_extra/", + "numbers": [ + { + "value": "LoRA-5 first on coverage at k = 3, 5 and 10: 25/25 draws", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + }, + { + "value": "LoRA-5 first on precision (draws) at k = 3/5/10: 23/21/20 of 25", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + }, + { + "value": "real vs real coverage at k = 3: ImMicro 0.932, Sepsis 0.925, Imperial 0.919, Giroux 0.993, Ji 0.980", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + }, + { + "value": "every set falls with smaller k (cohort means): True", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//fewshot_n5/draw/{metrics,en}.csv" + } + ], + "refs": [ + "Naeem 2020, ICML", + "Khayatkhoei 2023, ICML" + ], + "refs_full": [ + "Naeem et al. Reliable fidelity and diversity metrics for generative models. ICML 2020", + "Khayatkhoei & AbdAlmageed. Emergent asymmetry of precision and recall for measuring fidelity and diversity of generative models in high dimensions. ICML 2023" + ], + "original_slide": 60, + "original_title": "5 donors: precision/coverage at k = 3, 5, 10", + "original_footnote": "5 donors, 5 draws per cohort; error bars = SD over draws. k = 10 is the value on the main slides.", + "figures_before_palette": [ + "figures/sec09_extra_ksens.png" + ], + "source_page_131": 112, + "deck_group": "Supp S07", + "previous_section_label": "Supplementary", + "parent_page": 87, + "parent_enriched_page": 96, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_extra_ksens.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 98, + "parent_154_page": 110, + "parent_155_page": 111 + }, + { + "kind": "content", + "section": "Supp S07", + "section_label": "Supp. S07", + "text_source": { + "file": "data/sec09_extra_slide_text.json", + "index": 1 + }, + "title": "5-fold: realism beyond precision/coverage", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_extra_fivefold.png" + ], + "keywords": [ + "Every metric: LoRA first in 5/5 cohorts", + "Ji whole blood: E_N 1.9 vs 4.1\u201310.1", + "Still separable: LoRA AUC 0.87\u20130.94" + ], + "footnote": "Donor-level 5-fold CV, one pooled call per cohort. E_N: scSemiProfiler's own error metric.", + "footnote_full": "Original slide 61: Donor-level 5-fold CV; each pair scored once by the fold model that never trained on its donor (one pooled call per cohort, all cross-fitted pairs). Arms as on the 5-fold slide. Metric definitions as on the 5-donor realism slide. LoRA first among the 5 arms in all 5 cohorts on 12/12 headline metrics (kNN and RF AUC, 1-NN accuracy, MMD\u00b2 at \u03c3 = m and summed, E_N all cells, precision and coverage at k = 3, 5, 10). LoRA kNN AUC ImMicro 0.941, Sepsis 0.886, Imperial 0.866, Giroux 0.881, Ji 0.908; real vs real ImMicro 0.505, Sepsis 0.506, Imperial 0.496, Giroux 0.501, Ji 0.494. E_N LoRA vs other arms: ImMicro 1.56 vs 1.77\u20134.83; Sepsis 1.20 vs 1.40\u20135.29; Imperial 1.13 vs 1.19\u20134.05; Giroux 1.70 vs 2.24\u20134.48; Ji 1.95 vs 4.13\u201310.13. scSemiProfiler, 5-fold CV (every training donor of the fold, inner-val included, as representatives; arm scsemi__raw) drawn for ImMicro, Sepsis, Imperial, Giroux, Ji; sources: ImMicro /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra/immicro/scsemi_cv_20261002/pooled; Sepsis /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra/ye2026_sepsis/scsemi_cv_20261002/pooled; Imperial /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra/imperial_eskd/scsemi_cv_20261002/pooled; Giroux /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra/giroux2022/scsemi_cv_20261002/pooled; Ji /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra/ji2025_chili/scsemi_cv_20261002/pooled. Scores: scores_extra//pooled/.", + "sources": "tools/sec09_extra_metrics.py; data/sec09_extra_fivefold.csv, data/sec09_extra_scsemi_status.csv; plans/paired_finetune_intradataset_20260927/scores_extra/", + "numbers": [ + { + "value": "LoRA first among 5 arms in all 5 cohorts on 12/12 metrics", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//pooled/{metrics,en}.csv" + }, + { + "value": "Ji E_N: LoRA 1.948; Zero-shot, real bulk 10.125, Zero-shot + calibration 7.076, sc-only fine-tune (+ calibration) 7.215, scSemiProfiler, 5-fold CV (other 4 folds) 4.132", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//pooled/{metrics,en}.csv" + }, + { + "value": "LoRA kNN AUC (pooled): ImMicro 0.941, Sepsis 0.886, Imperial 0.866, Giroux 0.881, Ji 0.908", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores_extra//pooled/{metrics,en}.csv" + }, + { + "value": "scSemiProfiler, 5-fold CV: ImMicro, Sepsis, Imperial, Giroux, Ji", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/data/sec09_extra_scsemi_status.csv" + } + ], + "refs": [ + "Lopez-Paz 2017, ICLR", + "Gretton 2012, JMLR" + ], + "refs_full": [ + "Lopez-Paz & Oquab. Revisiting classifier two-sample tests. ICLR 2017", + "Gretton et al. A kernel two-sample test. JMLR 13:723-773 (2012)" + ], + "original_slide": 61, + "original_title": "5-fold: realism beyond precision/coverage", + "original_footnote": "Donor-level 5-fold CV, one pooled call per cohort. E_N: scSemiProfiler's own error metric.", + "figures_before_palette": [ + "figures/sec09_extra_fivefold.png" + ], + "source_page_131": 113, + "deck_group": "Supp S07", + "previous_section_label": "Supplementary", + "parent_page": 88, + "parent_enriched_page": 97, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_extra_fivefold.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 99, + "parent_154_page": 111, + "parent_155_page": 112 + }, + { + "kind": "content", + "section": "Supp S07", + "section_label": "Supp. S07", + "text_source": { + "file": "data/sec09_same_pairs_slide_text.json", + "index": 0 + }, + "title": "Same training pairs: LoRA vs scSemiProfiler", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_same_pairs.png" + ], + "keywords": [ + "Coverage, precision: LoRA ahead in 5/5", + "LoRA ahead in composition 4/5, logFC r 3/4", + "19\u201362\u00d7 less GPU time per fold" + ], + "footnote": "Completed original five-fold LoRA/scSemi comparison; not the new total-n curve.", + "sources": "tools/sec09_same_pairs.py; data/sec09_same_pairs.csv; plans/paired_finetune_intradataset_20260927/scsemi/analysis/cv_budget_20261005/", + "say": "Here scSemiProfiler gets exactly the pairs LoRA trains on, every training donor of the fold. LoRA is still ahead on every metric and about 20 times cheaper. Sepsis, Imperial and Giroux are still running.", + "footnote_full": "Current scope: Completed original five-fold LoRA/scSemi comparison; not the new total-n curve.\n\nOriginal slide 62: 5-fold CV: for every fold, scSemiProfiler's representatives are exactly LoRA's training + inner-val pairs (the other four folds; checked per fold), test = the fold's samples from their real bulk only. Pooled = one scoring call per method over every usable sample, same donors and reference (sha-checked). scSemiProfiler = upstream c2aea02, default settings, official output. GPU time per fold on one GB10: LoRA = fine-tune + sampling 1,500 cells per test sample; scSemiProfiler = pretraining of every representative + one model per test sample (receipts); data preparation and scoring not counted on either side. scSemiProfiler ahead on composition: Giroux. scSemiProfiler ahead on logFC r: Giroux. Equal pair budget, ImMicro (n = 5, 10, 20, 40 pairs): scSemiProfiler L1 0.218\u20130.227 and coverage 0.04\u20130.04 at every n; LoRA L1 0.171\u20130.188, coverage 0.76\u20130.84 (drawn on the 'Pairs needed: composition, coverage, training steps' slide). Equal pair budget, Sepsis (n = 5, 10, 20, 40 pairs): scSemiProfiler L1 0.516\u20130.536 and coverage 0.10\u20130.10 at every n; LoRA L1 0.365\u20130.428, coverage 0.77\u20130.82 (drawn on the 'Pairs needed: composition, coverage, training steps' slide). Source: plans/paired_finetune_intradataset_20260927/scores//{pooled,scsemi_cv_20261002/pooled}/ and scsemi/analysis/cv_budget_20261005/.", + "numbers": [ + { + "value": "ImMicro: L1 LoRA 0.162 / scSemi 0.217; coverage 0.784 / 0.038; precision 0.924 / 0.505; logFC r nan / nan; top-50 DEG None / None; GPU per fold LoRA 14.7 min / scSemi 4.75 h", + "source": "data/sec09_same_pairs.csv" + }, + { + "value": "Sepsis: L1 LoRA 0.362 / scSemi 0.502; coverage 0.829 / 0.096; precision 0.926 / 0.760; logFC r 0.527 / 0.435; top-50 DEG 23 / 6; GPU per fold LoRA 7.4 min / scSemi 7.69 h", + "source": "data/sec09_same_pairs.csv" + }, + { + "value": "Imperial: L1 LoRA 0.376 / scSemi 0.520; coverage 0.850 / 0.082; precision 0.847 / 0.348; logFC r 0.819 / 0.705; top-50 DEG 143 / 87; GPU per fold LoRA 6.2 min / scSemi 6.18 h", + "source": "data/sec09_same_pairs.csv" + }, + { + "value": "Giroux: L1 LoRA 0.258 / scSemi 0.243; coverage 0.928 / 0.354; precision 0.998 / 0.872; logFC r 0.461 / 0.537; top-50 DEG 5 / 0; GPU per fold LoRA 3.6 min / scSemi 2.89 h", + "source": "data/sec09_same_pairs.csv" + }, + { + "value": "Ji: L1 LoRA 0.504 / scSemi 0.575; coverage 0.857 / 0.167; precision 0.971 / 0.858; logFC r 0.479 / 0.273; top-50 DEG 15 / 0; GPU per fold LoRA 4.4 min / scSemi 1.48 h", + "source": "data/sec09_same_pairs.csv" + }, + { + "value": "Equal pair budget, ImMicro (n = 5, 10, 20, 40 pairs): scSemiProfiler L1 0.218\u20130.227 and coverage 0.04\u20130.04 at every n; LoRA L1 0.171\u20130.188, coverage 0.76\u20130.84 (drawn on the 'Pairs needed: composition, coverage, training steps' slide). Equal pair budget, Sepsis (n = 5, 10, 20, 40 pairs): scSemiProfiler L1 0.516\u20130.536 and coverage 0.10\u20130.10 at every n; LoRA L1 0.365\u20130.428, coverage 0.77\u20130.82 (drawn on the 'Pairs needed: composition, coverage, training steps' slide).", + "source": "data/sec09_same_pairs_budget.csv" + } + ], + "original_slide": 62, + "original_title": "Same training pairs: LoRA vs scSemiProfiler", + "original_footnote": "5-fold CV; scSemiProfiler representatives = LoRA's training pairs. ", + "figures_before_palette": [ + "figures/sec09_same_pairs.png" + ], + "source_page_131": 114, + "deck_group": "Supp S07", + "previous_section_label": "Supplementary", + "parent_page": 89, + "parent_enriched_page": 98, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_same_pairs.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 100, + "parent_154_page": 112, + "parent_155_page": 113 + }, + { + "kind": "content", + "section": "Supp S07", + "section_label": "Supp. S07", + "text_source": { + "file": "data/sec09_scsemi_representative_slide_text.json", + "index": 0 + }, + "title": "scSemiProfiler and the representative-copy baseline", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec09_scsemi_representative.png" + ], + "keywords": [ + "scSemi-5 never beats copying its representative: 5/5", + "Its error tracks the representative's: rho 0.57\u20130.93", + "LoRA-5 below the copy error in 4/5 cohorts" + ], + "footnote": "Copy = composition of the one training sample scSemiProfiler assigns to each test sample.", + "sources": "plans/paired_finetune_intradataset_20260927/scsemi/tools/representative_dependence.py; tools/sec09_scsemi_representative.py; data/sec09_scsemi_representative.csv", + "footnote_full": "Original slide 63: Same 25 five-donor draws. scSemiProfiler gives every test sample exactly one representative, the nearest training sample in its bulk PCA (assignment.json), and runs its encoder on that representative's cells. Copy = composition L1 between the test sample's and its representative's observed cells (same frozen scANVI labels as the scorer). Bars: mean over draws of the per-draw mean; dots = draws. scSemi vs copy: paired Wilcoxon over test samples (ImMicro p 1e-33, Sepsis p 0.014, Imperial p 0.91, Giroux p 0.0073, Ji p 0.46). Upstream's own representative choice (KMeans on the same bulk PCA, the sample nearest each centre; proxy, not a rerun) gives a lower copy error than our 5-donor draws in Imperial, Giroux and Ji, where our draws hold all visits of 5 donors and KMeans spreads the same number of samples over as many donors; equal or higher in ImMicro and sepsis. Copy error: ImMicro (5 samples) ours 0.19, KMeans 0.18, random samples 0.19; Sepsis (5 samples) ours 0.58, KMeans 0.61, random samples 0.58; Imperial (14 samples) ours 0.61, KMeans 0.53, random samples 0.56; Giroux (13 samples) ours 0.25, KMeans 0.16, random samples 0.23; Ji (8 samples) ours 0.57, KMeans 0.46, random samples 0.51. Source: plans/paired_finetune_intradataset_20260927/scsemi/analysis/representative_dependence_20261002/.", + "numbers": [ + { + "value": "ImMicro: copy 0.187, scSemi-5 0.233 (Wilcoxon vs copy p 1e-33), LoRA-5 0.176; rho scSemi 0.57, LoRA 0.35", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/representative_dependence_20261002/summary.csv" + }, + { + "value": "Sepsis: copy 0.580, scSemi-5 0.612 (Wilcoxon vs copy p 0.014), LoRA-5 0.466; rho scSemi 0.88, LoRA 0.56", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/representative_dependence_20261002/summary.csv" + }, + { + "value": "Imperial: copy 0.608, scSemi-5 0.602 (Wilcoxon vs copy p 0.91), LoRA-5 0.503; rho scSemi 0.93, LoRA 0.55", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/representative_dependence_20261002/summary.csv" + }, + { + "value": "Giroux: copy 0.253, scSemi-5 0.267 (Wilcoxon vs copy p 0.0073), LoRA-5 0.258; rho scSemi 0.81, LoRA 0.08", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/representative_dependence_20261002/summary.csv" + }, + { + "value": "Ji: copy 0.570, scSemi-5 0.591 (Wilcoxon vs copy p 0.46), LoRA-5 0.511; rho scSemi 0.72, LoRA 0.39", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/representative_dependence_20261002/summary.csv" + }, + { + "value": "KMeans proxy, copy error: ImMicro (5 samples) ours 0.19, KMeans 0.18, random samples 0.19; Sepsis (5 samples) ours 0.58, KMeans 0.61, random samples 0.58; Imperial (14 samples) ours 0.61, KMeans 0.53, random samples 0.56; Giroux (13 samples) ours 0.25, KMeans 0.16, random samples 0.23; Ji (8 samples) ours 0.57, KMeans 0.46, random samples 0.51", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/representative_dependence_20261002/kmeans_proxy.csv" + } + ], + "refs": [ + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "original_slide": 63, + "original_title": "scSemiProfiler copies its representative", + "original_footnote": "Copy = composition of the one training sample scSemiProfiler assigns to each test sample.", + "figures_before_palette": [ + "figures/sec09_scsemi_representative.png" + ], + "source_page_131": 115, + "deck_group": "Supp S07", + "previous_section_label": "Supplementary", + "parent_page": 90, + "parent_enriched_page": 99, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec09_scsemi_representative.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 101, + "parent_154_page": 113, + "parent_155_page": 114 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_strict_ye2026_sepsis_deg", + "title": "Sepsis: patient-aware disease-gene agreement", + "section_label": "Supp. S08", + "figures": [ + "figures/expanded_master_20261013/disease/strict_ye2026_sepsis_deg.png" + ], + "keywords": [], + "footnote": "Completed corrected count-based DEG fits; mean of five support draws.", + "footnote_full": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Values are saved FDR-filtered top-50 list overlap counts, not correlations or cell classification accuracy. The recorded method is edgeR/TMM/limma-voom, with donor duplicateCorrelation for Imperial. A donor can contribute to both disease conditions; group donor counts must not be summed as independent people. Historical sample-level Welch metrics are different endpoints.\nSources: plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_ye2026_sepsis_deg.csv\nEvidence source: plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_ye2026_sepsis_deg.csv", + "sources": "plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_ye2026_sepsis_deg.csv", + "notes": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Values are saved FDR-filtered top-50 list overlap counts, not correlations or cell classification accuracy. The recorded method is edgeR/TMM/limma-voom, with donor duplicateCorrelation for Imperial. A donor can contribute to both disease conditions; group donor counts must not be summed as independent people. Historical sample-level Welch metrics are different endpoints.\nSources: plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw0/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw1-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw1/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw2-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw2/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw3-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw3/statistics/status.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/ye2026_sepsis-disease-draw4-score.json; plans/paper_figures_strictpairs_20261012/scores/ye2026_sepsis/disease/draw4/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_ye2026_sepsis_deg.csv", + "after_original": 40, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_strict_ye2026_sepsis_deg", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/strict_ye2026_sepsis_deg.png" + ], + "source_page_131": 69, + "deck_group": "Supp S08", + "section": "Supp S08", + "previous_section_label": "Fig. 4 update", + "parent_page": 91, + "parent_enriched_page": 100, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/strict_ye2026_sepsis_deg.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 102, + "parent_154_page": 114, + "parent_155_page": 115 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_strict_imperial_eskd_deg", + "title": "Imperial: patient-aware disease-gene agreement", + "section_label": "Supp. S08", + "figures": [ + "figures/expanded_master_20261013/disease/strict_imperial_eskd_deg.png" + ], + "keywords": [], + "footnote": "Single completed exact-five draw 0; patient-aware DEG fits. Other draws are outside this frozen snapshot.", + "footnote_full": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Values are saved FDR-filtered top-50 list overlap counts, not correlations or cell classification accuracy. The recorded method is edgeR/TMM/limma-voom, with donor duplicateCorrelation for Imperial. A donor can contribute to both disease conditions; group donor counts must not be summed as independent people. Historical sample-level Welch metrics are different endpoints.\nSources: plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_eskd_deg.csv\nEvidence source: plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_eskd_deg.csv", + "sources": "plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_eskd_deg.csv", + "notes": "Exactly one paired sample from each of five donors, no extra validation, fixed 3000-step LoRA. Only completed score receipts are used. Raw, external-reference calibration, same-pair calibration, scSemiProfiler and LoRA share the recorded held-out targets. Dots are support draws, not independent population replicates. These slides do not certify the unfinished four-cohort campaign. Values are saved FDR-filtered top-50 list overlap counts, not correlations or cell classification accuracy. The recorded method is edgeR/TMM/limma-voom, with donor duplicateCorrelation for Imperial. A donor can contribute to both disease conditions; group donor counts must not be summed as independent people. Historical sample-level Welch metrics are different endpoints.\nSources: plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/donor_deg/donor_deg_summary.csv; plans/paper_figures_strictpairs_20261012/receipts/imperial_eskd-disease-draw0-score.json; plans/paper_figures_strictpairs_20261012/scores/imperial_eskd/disease/draw0/statistics/status.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/strict_imperial_eskd_deg.csv", + "after_original": 40, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_strict_imperial_eskd_deg", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/strict_imperial_eskd_deg.png" + ], + "source_page_131": 70, + "deck_group": "Supp S08", + "section": "Supp S08", + "previous_section_label": "Fig. 4 update", + "parent_page": 92, + "parent_enriched_page": 101, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/strict_imperial_eskd_deg.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 103, + "parent_154_page": 115, + "parent_155_page": 116 + }, + { + "kind": "content", + "section": "Supp S08", + "section_label": "Supp. S08", + "text_source": { + "file": "data/fig4_disease_slide_text.json", + "index": 2 + }, + "title": "Sepsis: fine-tuning variants", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig4_disease_variants.png" + ], + "keywords": [ + "LoRA r16 \u2248 full fine-tune: r 0.53 vs 0.53", + "Rank 64: r 0.55, DEG 26 vs 23", + "Multi-K and recipes: no gain" + ], + "footnote": "Sepsis, 82 donors, pooled 5-fold. LoRA r = 16 is the model used throughout.", + "sources": "tools/fig4_disease.py; data/fig4_disease.csv", + "refs": [ + "Hu 2022, ICLR" + ], + "refs_full": [ + "Hu et al. LoRA: Low-Rank Adaptation of Large Language Models. ICLR 2022" + ], + "say": "On sepsis, LoRA at rank 16 matches full fine-tuning with 42 times fewer trainable weights; rank 64 is slightly higher, and multi-K training or recipes do not help.", + "footnote_full": "Original slide 44: Sepsis, 82 donors, pooled 5\u2013fold CV, same scorer and donors for every variant. LoRA r = 16 is the model used throughout the deck. Full fine\u2013tune trains all 54.2 M weights, LoRA r16 1.30 M. Multi\u2013K: LoRA trained on K = 250\u20134,000 conditions; recipes: half of each batch is pseudobulk from the training donors' cells. LoRA r = 4: r 0.476, partial r 0.416, DEG 18; LoRA r = 16 (used): r 0.527, partial r 0.469, DEG 23; LoRA r = 64: r 0.553, partial r 0.492, DEG 26; Full fine\u2013tune: r 0.534, partial r 0.468, DEG 24; LoRA, multi\u2013K: r 0.509, partial r 0.447, DEG 23; LoRA + recipes: r 0.496, partial r 0.431, DEG 23. Source: data/fig4_disease.csv (scores/ye2026_sepsis/{pooled_deg, lora_rank_20261001/pooled}).", + "numbers": [ + { + "value": "LoRA r = 4: r 0.476, partial r 0.416, DEG 18; LoRA r = 16 (used): r 0.527, partial r 0.469, DEG 23; LoRA r = 64: r 0.553, partial r 0.492, DEG 26; Full fine\u2013tune: r 0.534, partial r 0.468, DEG 24; LoRA, multi\u2013K: r 0.509, partial r 0.447, DEG 23; LoRA + recipes: r 0.496, partial r 0.431, DEG 23", + "source": "data/fig4_disease.csv" + } + ], + "original_slide": 44, + "original_title": "Sepsis: fine-tuning variants", + "original_footnote": "Sepsis, 82 donors, pooled 5-fold. LoRA r = 16 is the model used throughout.", + "figures_before_palette": [ + "figures/fig4_disease_variants.png" + ], + "source_page_131": 82, + "deck_group": "Supp S08", + "previous_section_label": "Fig. 4 details", + "parent_page": 93, + "parent_enriched_page": 102, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig4_disease_variants.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 104, + "parent_154_page": 116, + "parent_155_page": 117 + }, + { + "kind": "content", + "section": "Supp S08", + "section_label": "Supp. S08", + "title": "Pairs needed: donor and disease signal", + "layout": "figure_keywords", + "figures": [ + "figures/final_palette_20261013/sec08_pairs_signal.png" + ], + "keywords": [ + "No pairs or calibration only: logFC r \u2264 0.12", + "LoRA logFC r 0.29 \u2192 0.53, DEG 5 \u2192 23", + "scSemi at 5\u201340 pairs: logFC r 0.52\u20130.58, DEG 2\u20136" + ], + "footnote": "Sepsis: 61 case vs 21 control. logFC r = mean over T / B / Mono / NK. 0 = no pairs.", + "footnote_full": "Original slide 45: Sepsis 61 case vs 21 control, pooled 5-fold scoring. Disease logFC r: as measured \u22120.01, length-corrected 0.11, calibration from the same n pairs 0.08\u20130.12, LoRA 0.29 / 0.42 / 0.37 / 0.48 / 0.53 at 5 / 10 / 20 / 40 / 58\u201359 pairs; zero-shot from sc pseudobulk 0.47. Top-50 DEG (Welch, sum over 5 types): as measured 9, length-corrected 11, calibration 7\u201312, LoRA 5 / 8 / 18 / 22 / 23. ImMicro donor NK fraction r: as measured 0.49, length-corrected 0.49, calibration 0.50\u20130.53, LoRA 0.20 \u2192 0.70. Diamonds: scSemiProfiler with exactly LoRA's n training + inner-val pairs as representatives (scores//scsemi_budget_20261002/pooled; last point = every training donor, scsemi_cv_20261002/pooled): sepsis logFC r 0.53 / 0.56 / 0.58 / 0.52 / 0.44 and top-50 DEG 2 / 3 / 2 / 6 / 6 at 5 / 10 / 20 / 40 / 58\u201359 pairs; ImMicro NK fraction r 0.27 / 0.33 / 0.48 / 0.49 / 0.42.", + "sources": "tools/sec08_pairs_needed.py; data/sec08_pairs_signal.csv, data/sec08_slide_numbers.csv", + "refs": [ + "Roohani 2024, Nat Biotechnol", + "Li 2026, Sci Adv" + ], + "refs_full": [ + "Roohani, Huang & Leskovec. Predicting transcriptional outcomes of novel multigene perturbations with GEARS. Nat Biotechnol 42:927-935 (2024). doi:10.1038/s41587-023-01905-6", + "Li et al. A systematic comparison of single-cell perturbation response prediction models. Sci Adv 12:eaed3414 (2026). doi:10.1126/sciadv.aed3414" + ], + "original_slide": 45, + "original_title": "Pairs needed: donor and disease signal", + "original_footnote": "Sepsis: 61 case vs 21 control. logFC r = mean over T / B / Mono / NK. 0 = no pairs.", + "figures_before_palette": [ + "figures/sec08_pairs_signal.png" + ], + "source_page_131": 83, + "deck_group": "Supp S08", + "previous_section_label": "Fig. 4 details", + "parent_page": 94, + "parent_enriched_page": 103, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec08_pairs_signal.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 105, + "parent_154_page": 117, + "parent_155_page": 118 + }, + { + "kind": "content", + "section": "Supp S08", + "section_label": "Supp. S08", + "text_source": { + "file": "data/secS_deconvolution_slide_text.json", + "index": 0 + }, + "title": "Cell-type deconvolution, same pairs: composition", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/secS_deconv_composition.png" + ], + "keywords": [ + "PBISC closer than BayesPrism in 22/25 draws, than CIBERSORTx in 24/25", + "Composition compared with held-out observed cells", + "Five-fold: PBISC lowest error in 4 of 5 cohorts" + ], + "footnote": "Completed deconvolution outputs already exist. Verified identical Sepsis supports use adoption instead of method reruns.", + "sources": "tools/secS_deconvolution.py; data/secS_deconvolution.csv; plans/bayesprism_paired_20261006/results/{RESULTS_full.md, RESULTS_cibersortx.md, TABLE_*_full.csv, TABLE_*_cibersortx.csv}", + "refs": [ + "Chu 2022, Nat Cancer", + "Newman 2019, Nat Biotechnol", + "Wang 2024, Nat Commun" + ], + "refs_full": [ + "Chu et al. Cell type and gene expression deconvolution with BayesPrism enables Bayesian integrative analysis across bulk and single-cell RNA sequencing in oncology. Nat Cancer 3:505-517 (2022). doi:10.1038/s43018-022-00356-3", + "Newman et al. Determining cell type abundance and expression from bulk tissues with digital cytometry. Nat Biotechnol 37:773-782 (2019). doi:10.1038/s41587-019-0114-2", + "Wang et al. scSemiProfiler: advancing large-scale single-cell studies through semi-profiling with deep generative models and active learning. Nat Commun 15:5989 (2024). doi:10.1038/s41467-024-50150-1" + ], + "say": "For reference only. Could standard bulk deconvolution do the same job with the same paired donors? BayesPrism and CIBERSORTx used the paired donors' own cells as the reference and the test samples' real bulk as the mixture, scored with the same truth and code as PBISC. PBISC estimates the cell-type mix more closely in almost every draw. On pseudobulk BayesPrism recovers the right mix, so the gap comes from real bulk, which PBISC learns from the pairs. Deconvolution also returns proportions, not cells.", + "footnote_full": "Current scope: Completed deconvolution outputs already exist. Verified identical Sepsis supports use adoption instead of method reruns.\n\nOriginal slide 46: Cell-type deconvolution (cell-type fractions and per-type expression, not single cells) given the same paired donors as PBISC LoRA, on the paired benchmark's own splits, observed truth and scorer: 5-donor draws (reference = the draw's 5 training donors' observed cells, mixture = the test samples' real bulk) and 5-fold CV (reference = each fold's LoRA training + inner-validation pairs; pooled). BayesPrism: recommended gene set (protein-coding minus chrX/Y/M and ribosomal genes, 15-19k), reference states = gate | AIFI L2; its better theta (initial or final) in cell units is shown. CIBERSORTx: official Docker images (x86, run under qemu emulation on server_1), single-cell reference mode with S-mode batch correction (up to 400 reference cells per type), fractions in cell units (divided by each type's mean UMI per reference cell). CIBERSORTx HiRes gives per-sample cell-type expression for the genes it imputes per type (others are written as 1 and count as zero here); it imputed no gene of the four contrast types in 1 Ji ChILI and 2 Giroux draws, which therefore have no CIBERSORTx disease score. Composition L1 over the six-type gate, lower is better. 5-fold: PBISC has the lowest L1 except in Giroux (scSemiProfiler). Pipeline check: on pseudobulk mixtures BayesPrism recovers the observed mRNA shares (e.g. T 0.39 vs 0.38), so the error comes from real bulk. Source: plans/bayesprism_paired_20261006/results/RESULTS_full.md, RESULTS_cibersortx.md.", + "numbers": [ + { + "value": "ImMicro: 5 donors L1 PBISC 0.176 / BayesPrism 0.537 / CIBERSORTx 0.491; 5-fold PBISC 0.162 / BayesPrism 0.575 / CIBERSORTx 0.507 / scSemiProfiler 0.217; Sepsis: 5 donors L1 PBISC 0.466 / BayesPrism 0.446 / CIBERSORTx 0.582; 5-fold PBISC 0.362 / BayesPrism 0.475 / CIBERSORTx 0.607 / scSemiProfiler 0.502; Ji ChILI: 5 donors L1 PBISC 0.513 / BayesPrism 0.893 / CIBERSORTx 0.827; 5-fold PBISC 0.504 / BayesPrism 0.883 / CIBERSORTx 0.914 / scSemiProfiler 0.575; Imperial: 5 donors L1 PBISC 0.503 / BayesPrism 0.845 / CIBERSORTx 0.646; 5-fold PBISC 0.376 / BayesPrism 0.887 / CIBERSORTx 0.592 / scSemiProfiler 0.520; Giroux: 5 donors L1 PBISC 0.256 / BayesPrism 1.455 / CIBERSORTx 0.669; 5-fold PBISC 0.258 / BayesPrism 0.878 / CIBERSORTx 0.591 / scSemiProfiler 0.243", + "source": "data/secS_deconvolution.csv" + } + ], + "original_slide": 46, + "original_title": "Cell-type deconvolution, same pairs: composition", + "original_footnote": "Same donors, truth and scorer as PBISC. Reference = the paired donors' cells.", + "figures_before_palette": [ + "figures/secS_deconv_composition.png" + ], + "source_page_131": 84, + "deck_group": "Supp S08", + "previous_section_label": "Fig. 4 details", + "parent_page": 95, + "parent_enriched_page": 104, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/secS_deconv_composition.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 106, + "parent_154_page": 118, + "parent_155_page": 119 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_deconv_expression_coverage", + "title": "Deconvolution: expression coverage limits the comparison", + "section_label": "Supp. S08", + "figures": [ + "figures/final_palette_20261013/deconv_expression_coverage.png" + ], + "keywords": [], + "footnote": "Completed historical CIBERSORTx five-donor runs. Missing disease correlations remain unavailable; deconvolution does not generate individual cells.", + "footnote_full": "Historical five-donor training can include multiple paired visits in Imperial, Giroux and Ji. It is not the strict five-sample protocol. Existing predictions and frozen marker/label rules are reused. Coverage is the saved HiRes covered-gene count; some genes outside the four contrast cell types can be covered even when the disease correlation is unavailable. Two Giroux draws and one Ji draw lack that score. The internal summary field named bayesprism is reused by the old scorer for CIBERSORTx; these values are from CIBERSORTx files. Existing Sepsis method adoption does not mean new corrected strict deconvolution statistics are already complete.\nSources: plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw4_cibersortx/summary.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/deconv_expression_coverage.csv; plans/bayesprism_paired_20261006/results/RESULTS_cibersortx.md\nEvidence source: plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw4_cibersortx/summary.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/deconv_expression_coverage.csv; plans/bayesprism_paired_20261006/results/RESULTS_cibersortx.md", + "sources": "plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw4_cibersortx/summary.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/deconv_expression_coverage.csv; plans/bayesprism_paired_20261006/results/RESULTS_cibersortx.md", + "notes": "Historical five-donor training can include multiple paired visits in Imperial, Giroux and Ji. It is not the strict five-sample protocol. Existing predictions and frozen marker/label rules are reused. Coverage is the saved HiRes covered-gene count; some genes outside the four contrast cell types can be covered even when the disease correlation is unavailable. Two Giroux draws and one Ji draw lack that score. The internal summary field named bayesprism is reused by the old scorer for CIBERSORTx; these values are from CIBERSORTx files. Existing Sepsis method adoption does not mean new corrected strict deconvolution statistics are already complete.\nSources: plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ye2026_sepsis/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/imperial_eskd/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/giroux2022/fewshot_draw4_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw0_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw1_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw2_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw3_cibersortx/summary.json; plans/bayesprism_paired_20261006/results/ji2025_chili/fewshot_draw4_cibersortx/summary.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/deconv_expression_coverage.csv; plans/bayesprism_paired_20261006/results/RESULTS_cibersortx.md", + "after_original": 46, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_deconv_expression_coverage", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/deconv_expression_coverage.png" + ], + "source_page_131": 85, + "deck_group": "Supp S08", + "section": "Supp S08", + "previous_section_label": "Deconvolution detail", + "parent_page": 96, + "parent_enriched_page": 105, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/deconv_expression_coverage.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 107, + "parent_154_page": 119, + "parent_155_page": 120 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_k4000_quality", + "title": "K=4,000 recipe: paired reconstruction quality", + "section_label": "Supp. S08", + "figures": [ + "figures/final_palette_20261013/k4000_quality.png" + ], + "keywords": [], + "footnote": "Existing paired benchmark, five support draws per cohort and 1,500 cells/test sample. Bars = means; dots = draws. No new runs.", + "footnote_full": "The stored K1000 re-score exactly matched 25 original metrics in every draw. K4000 uses a changed condition recipe and draw-specific seeds; it is not a perfectly isolated one-factor K test. Imperial training uses all visits from five donors. Disease logFC/DEG are the historical sample-level estimators, not the new patient-aware strict metrics. No donor-draw p-values are used here; overlapping test donors are not independent replicates.\nSources: plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_quality.csv; plans/bulkonly_lora5_k4000_20261006/README.md\nEvidence source: plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_quality.csv; plans/bulkonly_lora5_k4000_20261006/README.md", + "sources": "plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_quality.csv; plans/bulkonly_lora5_k4000_20261006/README.md", + "notes": "The stored K1000 re-score exactly matched 25 original metrics in every draw. K4000 uses a changed condition recipe and draw-specific seeds; it is not a perfectly isolated one-factor K test. Imperial training uses all visits from five donors. Disease logFC/DEG are the historical sample-level estimators, not the new patient-aware strict metrics. No donor-draw p-values are used here; overlapping test donors are not independent replicates.\nSources: plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_quality.csv; plans/bulkonly_lora5_k4000_20261006/README.md", + "after_original": 67, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_k4000_quality", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/k4000_quality.png" + ], + "source_page_131": 120, + "deck_group": "Supp S08", + "section": "Supp S08", + "previous_section_label": "Recipe sensitivity", + "parent_page": 97, + "parent_enriched_page": 106, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/k4000_quality.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 108, + "parent_154_page": 120, + "parent_155_page": 121 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_k4000_disease", + "title": "K=4,000 recipe: disease signal does not improve uniformly", + "section_label": "Supp. S08", + "figures": [ + "figures/final_palette_20261013/k4000_disease.png" + ], + "keywords": [], + "footnote": "Existing paired benchmark, five support draws per cohort and 1,500 cells/test sample. Bars = means; dots = draws. No new runs.", + "footnote_full": "The stored K1000 re-score exactly matched 25 original metrics in every draw. K4000 uses a changed condition recipe and draw-specific seeds; it is not a perfectly isolated one-factor K test. Imperial training uses all visits from five donors. Disease logFC/DEG are the historical sample-level estimators, not the new patient-aware strict metrics. No donor-draw p-values are used here; overlapping test donors are not independent replicates.\nSources: plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_disease.csv; plans/bulkonly_lora5_k4000_20261006/README.md\nEvidence source: plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_disease.csv; plans/bulkonly_lora5_k4000_20261006/README.md", + "sources": "plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_disease.csv; plans/bulkonly_lora5_k4000_20261006/README.md", + "notes": "The stored K1000 re-score exactly matched 25 original metrics in every draw. K4000 uses a changed condition recipe and draw-specific seeds; it is not a perfectly isolated one-factor K test. Imperial training uses all visits from five donors. Disease logFC/DEG are the historical sample-level estimators, not the new patient-aware strict metrics. No donor-draw p-values are used here; overlapping test donors are not independent replicates.\nSources: plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/ye2026_sepsis/AGGREGATE.md; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/per_draw.csv; plans/bulkonly_lora5_k4000_20261006/analysis/benchmark/imperial_eskd/AGGREGATE.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_disease.csv; plans/bulkonly_lora5_k4000_20261006/README.md", + "after_original": 67, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_k4000_disease", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/k4000_disease.png" + ], + "source_page_131": 121, + "deck_group": "Supp S08", + "section": "Supp S08", + "previous_section_label": "Recipe sensitivity", + "parent_page": 98, + "parent_enriched_page": 107, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/k4000_disease.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 109, + "parent_154_page": 121, + "parent_155_page": 122 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_k4000_calibration", + "title": "K=4,000 recipe: calibration gains are selective", + "section_label": "Supp. S08", + "figures": [ + "figures/final_palette_20261013/k4000_calibration.png" + ], + "keywords": [], + "footnote": "Saved paired-test gates: 50 features \u00d7 5 draws per cohort. Descriptive counts, not 250 independent biological replicates.", + "footnote_full": "Level gate: generated/observed ratio 0.5\u20132 and Spearman at least 0.3. Contrast gate: same sign and generated magnitude 0.5\u20132 times observed. Disease contrasts shown are adult sepsis/HC and COVID+/COVID\u2212. Imperial severity gates are omitted here because historical severity grouping required separate correction. The K4000 bulk-only discovery workflow was not completed and is not inferred from calibration.\nSources: plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/SUMMARY.md; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/SUMMARY.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_calibration.csv\nEvidence source: plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/SUMMARY.md; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/SUMMARY.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_calibration.csv", + "sources": "plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/SUMMARY.md; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/SUMMARY.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_calibration.csv", + "notes": "Level gate: generated/observed ratio 0.5\u20132 and Spearman at least 0.3. Contrast gate: same sign and generated magnitude 0.5\u20132 times observed. Disease contrasts shown are adult sepsis/HC and COVID+/COVID\u2212. Imperial severity gates are omitted here because historical severity grouping required separate correction. The K4000 bulk-only discovery workflow was not completed and is not inferred from calibration.\nSources: plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/ye2026_sepsis/SUMMARY.md; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/per_feature.csv; plans/bulkonly_lora5_k4000_20261006/analysis/compare_k1000/imperial_eskd/SUMMARY.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/k4000_calibration.csv", + "after_original": 67, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_k4000_calibration", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/k4000_calibration.png" + ], + "source_page_131": 122, + "deck_group": "Supp S08", + "section": "Supp S08", + "previous_section_label": "Recipe sensitivity", + "parent_page": 99, + "parent_enriched_page": 108, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/k4000_calibration.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 110, + "parent_154_page": 122, + "parent_155_page": 123 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_influenza_response", + "title": "Influenza: response order versus response genes", + "section_label": "Supp. S09", + "figures": [ + "figures/final_palette_20261013/fig_randolph_celltype_response.png" + ], + "keywords": [], + "footnote": "Historical zero-shot influenza test. Response-strength ordering is recovered; gene-wise agreement remains below the matched bulk broadcast.", + "footnote_full": "Generated cells rank monocytes as strongest and T cells as weakest responders, but fail the preregistered gene-wise beyond-broadcast rule. Different arms can have different eligible donor counts; each broadcast matches its arm's donors. Observed-PBMC broadcast is an oracle. No new fitting or cell generation.\nSources: plans/utility_studies_final6k_20260925/randolph/RESULTS.md; plans/utility_studies_final6k_20260925/randolph/MANIFEST.json; plans/utility_studies_final6k_20260925/randolph/out/extended/extended_analysis.json\nEvidence source: plans/utility_studies_final6k_20260925/randolph/RESULTS.md; plans/utility_studies_final6k_20260925/randolph/MANIFEST.json; plans/utility_studies_final6k_20260925/randolph/out/extended/extended_analysis.json", + "sources": "plans/utility_studies_final6k_20260925/randolph/RESULTS.md; plans/utility_studies_final6k_20260925/randolph/MANIFEST.json; plans/utility_studies_final6k_20260925/randolph/out/extended/extended_analysis.json", + "notes": "Generated cells rank monocytes as strongest and T cells as weakest responders, but fail the preregistered gene-wise beyond-broadcast rule. Different arms can have different eligible donor counts; each broadcast matches its arm's donors. Observed-PBMC broadcast is an oracle. No new fitting or cell generation.\nSources: plans/utility_studies_final6k_20260925/randolph/RESULTS.md; plans/utility_studies_final6k_20260925/randolph/MANIFEST.json; plans/utility_studies_final6k_20260925/randolph/out/extended/extended_analysis.json", + "after_original": 43, + "palette": "biological", + "new_experiment": false, + "addition_key": "disease_influenza_response", + "figures_before_palette": [ + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/randolph/out/fig_randolph_celltype_response.png" + ], + "source_page_131": 76, + "deck_group": "Supp S09", + "section": "Supp S09", + "previous_section_label": "External diagnostic", + "parent_page": 100, + "parent_enriched_page": 109, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig_randolph_celltype_response.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 111, + "parent_154_page": 123, + "parent_155_page": 124 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_covid_map", + "title": "COVID: calibration on a shared real-cell map", + "section_label": "Supp. S09", + "figures": [ + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/figures/fig1_umap_seven_panels_by_label_depth1000.png" + ], + "keywords": [], + "footnote": "Historical 7 COVID / 5 HC samples, common depth. High-resolution reference overview; following detail panels provide readable quantitative results.", + "footnote_full": "This is an existing zero-shot input/calibration study, not LoRA/scSemiProfiler. Observed cells define the reference map. Use saved quantitative endpoints rather than interpreting overlap as a performance score. Biological cell-type colors are preserved.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/prec_cov.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/composition_full_chain.csv\nEvidence source: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/prec_cov.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/composition_full_chain.csv", + "sources": "plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/prec_cov.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/composition_full_chain.csv", + "notes": "This is an existing zero-shot input/calibration study, not LoRA/scSemiProfiler. Observed cells define the reference map. Use saved quantitative endpoints rather than interpreting overlap as a performance score. Biological cell-type colors are preserved.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/prec_cov.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/composition_full_chain.csv", + "after_original": 43, + "palette": "biological", + "new_experiment": false, + "addition_key": "disease_covid_map", + "figures_before_palette": [ + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/figures/fig1_umap_seven_panels_by_label_depth1000.png" + ], + "source_page_131": 77, + "deck_group": "Supp S09", + "section": "Supp S09", + "previous_section_label": "External diagnostic", + "parent_page": 101, + "parent_enriched_page": 110, + "figures_before_final_palette": [ + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/figures/fig1_umap_seven_panels_by_label_depth1000.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 112, + "parent_154_page": 124, + "parent_155_page": 125 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_covid_direction", + "title": "COVID monocytes: direction after depth matching", + "section_label": "Supp. S09", + "figures": [ + "figures/final_palette_20261013/fig2_covid_hc_monocyte_arrows_depth1000.png" + ], + "keywords": [], + "footnote": "Historical 7 COVID / 5 HC samples, common depth. High-resolution reference overview; following panels separate direction, magnitude and relabeling null.", + "footnote_full": "Native-depth and depth-matched results differ. Relabelling-null shares are descriptive; not every directional cosine exceeds its null. No claim of independent significance or completed Figure 6 reproduction.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv\nEvidence source: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv", + "sources": "plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv", + "notes": "Native-depth and depth-matched results differ. Relabelling-null shares are descriptive; not every directional cosine exceeds its null. No claim of independent significance or completed Figure 6 reproduction.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/MANIFEST.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/receipts/receipt_checks.json; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv", + "after_original": 43, + "palette": "biological", + "new_experiment": false, + "addition_key": "disease_covid_direction", + "figures_before_palette": [ + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/figures/fig2_covid_hc_monocyte_arrows_depth1000.png" + ], + "source_page_131": 78, + "deck_group": "Supp S09", + "section": "Supp S09", + "previous_section_label": "External diagnostic", + "parent_page": 102, + "parent_enriched_page": 111, + "figures_before_final_palette": [ + "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/figures/fig2_covid_hc_monocyte_arrows_depth1000.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 113, + "parent_154_page": 125, + "parent_155_page": 126 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_covid_monocyte_direction_detail", + "title": "COVID monocytes: direction, magnitude and null", + "section_label": "Supp. S09", + "figures": [ + "figures/expanded_master_20261013/disease/covid_monocyte_direction_detail.png" + ], + "keywords": [], + "footnote": "Readable detail of the retained COVID overview. Existing 7 COVID / 5 HC paired samples; native and depth-matched results stay separate.", + "footnote_full": "Every saved arm is retained, including the held-out real-cell reference, raw, TPM, both reference calibrations, paired calibration and own-sc-pseudobulk conditions. Cosine and length ratio use the recorded PCA-50 disease displacement; they are not read from the 2D UMAP. Relabeling-null share is the existing descriptive fraction of label permutations whose cosine is at least the observed one, not a newly fitted p-value or proof of biological reproduction. The full original depth-matched spatial overview remains in the deck. Only native UMAP coordinates were persisted in embedding_native_cells.parquet; these detail panels use saved numeric displacement tables instead of refitting depth-matched UMAP.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_monocyte_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md\nEvidence source: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_monocyte_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md", + "sources": "plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_monocyte_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md", + "notes": "Every saved arm is retained, including the held-out real-cell reference, raw, TPM, both reference calibrations, paired calibration and own-sc-pseudobulk conditions. Cosine and length ratio use the recorded PCA-50 disease displacement; they are not read from the 2D UMAP. Relabeling-null share is the existing descriptive fraction of label permutations whose cosine is at least the observed one, not a newly fitted p-value or proof of biological reproduction. The full original depth-matched spatial overview remains in the deck. Only native UMAP coordinates were persisted in embedding_native_cells.parquet; these detail panels use saved numeric displacement tables instead of refitting depth-matched UMAP.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_monocyte_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md", + "after_original": 43, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_covid_monocyte_direction_detail", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/covid_monocyte_direction_detail.png" + ], + "source_page_131": 79, + "deck_group": "Supp S09", + "section": "Supp S09", + "previous_section_label": "External diagnostic", + "parent_page": 103, + "parent_enriched_page": 112, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/covid_monocyte_direction_detail.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 114, + "parent_154_page": 126, + "parent_155_page": 127 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_covid_cd14_direction_detail", + "title": "COVID CD14 monocytes: direction, magnitude and null", + "section_label": "Supp. S09", + "figures": [ + "figures/expanded_master_20261013/disease/covid_cd14_direction_detail.png" + ], + "keywords": [], + "footnote": "Readable detail of the retained COVID overview. Existing 7 COVID / 5 HC paired samples; native and depth-matched results stay separate.", + "footnote_full": "Every saved arm is retained, including the held-out real-cell reference, raw, TPM, both reference calibrations, paired calibration and own-sc-pseudobulk conditions. Cosine and length ratio use the recorded PCA-50 disease displacement; they are not read from the 2D UMAP. Relabeling-null share is the existing descriptive fraction of label permutations whose cosine is at least the observed one, not a newly fitted p-value or proof of biological reproduction. The full original depth-matched spatial overview remains in the deck. Only native UMAP coordinates were persisted in embedding_native_cells.parquet; these detail panels use saved numeric displacement tables instead of refitting depth-matched UMAP.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_cd14_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md\nEvidence source: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_cd14_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md", + "sources": "plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_cd14_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md", + "notes": "Every saved arm is retained, including the held-out real-cell reference, raw, TPM, both reference calibrations, paired calibration and own-sc-pseudobulk conditions. Cosine and length ratio use the recorded PCA-50 disease displacement; they are not read from the 2D UMAP. Relabeling-null share is the existing descriptive fraction of label permutations whose cosine is at least the observed one, not a newly fitted p-value or proof of biological reproduction. The full original depth-matched spatial overview remains in the deck. Only native UMAP coordinates were persisted in embedding_native_cells.parquet; these detail panels use saved numeric displacement tables instead of refitting depth-matched UMAP.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_cd14_direction_detail.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/placement/tables/displacement_relabel_null.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md", + "after_original": 43, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_covid_cd14_direction_detail", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/covid_cd14_direction_detail.png" + ], + "source_page_131": 80, + "deck_group": "Supp S09", + "section": "Supp S09", + "previous_section_label": "External diagnostic", + "parent_page": 104, + "parent_enriched_page": 113, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/covid_cd14_direction_detail.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 115, + "parent_154_page": 127, + "parent_155_page": 128 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_covid_endpoints", + "title": "COVID: biological endpoints give a mixed verdict", + "section_label": "Supp. S09", + "figures": [ + "figures/expanded_master_20261013/disease/covid_endpoints.png" + ], + "keywords": [], + "footnote": "Historical 7 COVID / 5 HC samples. Saved effect ratios and donor ranks; no new fits. Reference calibration is not uniformly best.", + "footnote_full": "Four predefined endpoints are all retained. Reference calibration improves some plasmablast/composition endpoints but loses HLA-DR donor ranking and overshoots the ISG effect. All generated arms fail the original preregistered gene-wide beyond-broadcast criterion; the alternate pipeline definition is not substituted.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/effects_final/endpoints_paired12.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_endpoints.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/global_r/global_logfc_r_two_definitions.csv\nEvidence source: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/effects_final/endpoints_paired12.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_endpoints.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/global_r/global_logfc_r_two_definitions.csv", + "sources": "plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/effects_final/endpoints_paired12.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_endpoints.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/global_r/global_logfc_r_two_definitions.csv", + "notes": "Four predefined endpoints are all retained. Reference calibration improves some plasmablast/composition endpoints but loses HLA-DR donor ranking and overshoots the ISG effect. All generated arms fail the original preregistered gene-wide beyond-broadcast criterion; the alternate pipeline definition is not substituted.\nSources: plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/effects_final/endpoints_paired12.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/covid_endpoints.csv; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/RESULTS.md; plans/utility_studies_final6k_20260925/refcal_scoring/arunachalam/global_r/global_logfc_r_two_definitions.csv", + "after_original": 43, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_covid_endpoints", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/covid_endpoints.png" + ], + "source_page_131": 81, + "deck_group": "Supp S09", + "section": "Supp S09", + "previous_section_label": "External diagnostic", + "parent_page": 105, + "parent_enriched_page": 114, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/covid_endpoints.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 116, + "parent_154_page": 128, + "parent_155_page": 129 + }, + { + "kind": "content", + "title": "Rare-state frequencies in observed cells", + "section_label": "Supp. S10", + "layout": "full", + "figures": [ + "figures/review_style_20261013_rare_observed.png" + ], + "keywords": [], + "footnote": "Original five-donor outputs. Median across draws; state share within its parent population. Controls included.", + "footnote_full": "Additional view from the audited results review. Original five-donor outputs. Median across draws; state share within its parent population. Controls included.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/review_style_20261013_rare_observed.pdf", + "addition_key": "review:22", + "after_original": 49, + "figures_before_palette": [ + "figures/review_style_20261013_rare_observed.png" + ], + "source_page_131": 89, + "deck_group": "Supp S10", + "section": "Supp S10", + "previous_section_label": "Fig. 5a", + "parent_page": 106, + "parent_enriched_page": 115, + "figures_before_final_palette": [ + "figures/review_style_20261013_rare_observed.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 117, + "parent_154_page": 129, + "parent_155_page": 130 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_rare_misses_1", + "title": "Rare-state misses: rules 1\u20138", + "section_label": "Supp. S10", + "figures": [ + "figures/expanded_master_20261013/disease/rare_misses_1.png" + ], + "keywords": [], + "footnote": "Saved zero-aware diagnostic; all 15 predefined rules are shown across two slides. NA means not evaluable, not zero error.", + "footnote_full": "Historical five-donor training can include multiple paired visits in Imperial, Giroux and Ji. It is not the strict five-sample protocol. Existing predictions and frozen marker/label rules are reused. Each percentage is the saved mean across support draws of missed observed-positive donor fraction. Actual donor IDs, zero draws and common depth/cell budget are retained by the existing analysis. Mandatory IGLL5 is absent for the Ji plasmablast marker rule, which stays unavailable.\nSources: plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_1.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md\nEvidence source: plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_1.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md", + "sources": "plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_1.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md", + "notes": "Historical five-donor training can include multiple paired visits in Imperial, Giroux and Ji. It is not the strict five-sample protocol. Existing predictions and frozen marker/label rules are reused. Each percentage is the saved mean across support draws of missed observed-positive donor fraction. Actual donor IDs, zero draws and common depth/cell budget are retained by the existing analysis. Mandatory IGLL5 is absent for the Ji plasmablast marker rule, which stays unavailable.\nSources: plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_1.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md", + "after_original": 50, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_rare_misses_1", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/rare_misses_1.png" + ], + "source_page_131": 91, + "deck_group": "Supp S10", + "section": "Supp S10", + "previous_section_label": "Fig. 5 detail", + "parent_page": 107, + "parent_enriched_page": 116, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/rare_misses_1.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 118, + "parent_154_page": 130, + "parent_155_page": 131 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_rare_misses_2", + "title": "Rare-state misses: rules 9\u201315", + "section_label": "Supp. S10", + "figures": [ + "figures/expanded_master_20261013/disease/rare_misses_2.png" + ], + "keywords": [], + "footnote": "Saved zero-aware diagnostic; all 15 predefined rules are shown across two slides. NA means not evaluable, not zero error.", + "footnote_full": "Historical five-donor training can include multiple paired visits in Imperial, Giroux and Ji. It is not the strict five-sample protocol. Existing predictions and frozen marker/label rules are reused. Each percentage is the saved mean across support draws of missed observed-positive donor fraction. Actual donor IDs, zero draws and common depth/cell budget are retained by the existing analysis. Mandatory IGLL5 is absent for the Ji plasmablast marker rule, which stays unavailable.\nSources: plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_2.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md\nEvidence source: plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_2.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md", + "sources": "plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_2.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md", + "notes": "Historical five-donor training can include multiple paired visits in Imperial, Giroux and Ji. It is not the strict five-sample protocol. Existing predictions and frozen marker/label rules are reused. Each percentage is the saved mean across support draws of missed observed-positive donor fraction. Actual donor IDs, zero draws and common depth/cell budget are retained by the existing analysis. Mandatory IGLL5 is absent for the Ji plasmablast marker rule, which stays unavailable.\nSources: plans/paper_figures_strictpairs_20261012/source_data/rare_all_states_aggregate.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/rare_misses_2.csv; plans/paper_figures_strictpairs_20261012/reports/rare_provenance.json; plans/paper_figures_strictpairs_20261012/reports/rare_readiness.md", + "after_original": 50, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_rare_misses_2", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/rare_misses_2.png" + ], + "source_page_131": 92, + "deck_group": "Supp S10", + "section": "Supp S10", + "previous_section_label": "Fig. 5 detail", + "parent_page": 108, + "parent_enriched_page": 117, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/rare_misses_2.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 119, + "parent_154_page": 131, + "parent_155_page": 132 + }, + { + "kind": "content", + "title": "Rare-state fidelity: zero-aware reanalysis", + "section_label": "Supp. S10", + "layout": "full", + "figures": [ + "figures/review_style_20261013_rare_corrected_comparison.png" + ], + "keywords": [], + "footnote": "Original five-donor outputs; zero draws retained and actual donor IDs. 59 state/cohort combinations, including controls.", + "footnote_full": "Additional view from the audited results review. Original five-donor outputs; zero draws retained and actual donor IDs. 59 state/cohort combinations, including controls.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/review_style_20261013_rare_corrected_comparison.pdf", + "addition_key": "review:23", + "after_original": 50, + "figures_before_palette": [ + "figures/review_style_20261013_rare_corrected_comparison.png" + ], + "source_page_131": 93, + "deck_group": "Supp S10", + "section": "Supp S10", + "previous_section_label": "Fig. 5b", + "parent_page": 109, + "parent_enriched_page": 118, + "figures_before_final_palette": [ + "figures/review_style_20261013_rare_corrected_comparison.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 120, + "parent_154_page": 132, + "parent_155_page": 133 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_sepsis_rare_support", + "title": "Sepsis rare states across paired-data budgets", + "section_label": "Supp. S10", + "figures": [ + "figures/expanded_master_20261013/disease/sepsis_rare_support.png" + ], + "keywords": [], + "footnote": "Historical cross-fitted learning curve: n training pairs plus validation, not strict total n. All 15 rules; arm-specific depth matching.", + "footnote_full": "Existing pooled predictions for 82 donors and 1500 generated cells per sample. Each donor was held out from its fold model. Raw and pseudobulk are zero-shot controls; numbered/full columns are historical LoRA budgets. Observed cells were thinned separately to each arm's depth, so compare gen/obs within an arm. No new generation or rare-state scoring; the full 15-rule family, including controls, is preserved.\nSources: plans/rare_states_screen_20261004/out/sepsis_summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_support.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json\nEvidence source: plans/rare_states_screen_20261004/out/sepsis_summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_support.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json", + "sources": "plans/rare_states_screen_20261004/out/sepsis_summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_support.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json", + "notes": "Existing pooled predictions for 82 donors and 1500 generated cells per sample. Each donor was held out from its fold model. Raw and pseudobulk are zero-shot controls; numbered/full columns are historical LoRA budgets. Observed cells were thinned separately to each arm's depth, so compare gen/obs within an arm. No new generation or rare-state scoring; the full 15-rule family, including controls, is preserved.\nSources: plans/rare_states_screen_20261004/out/sepsis_summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_support.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json", + "after_original": 51, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_sepsis_rare_support", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/sepsis_rare_support.png" + ], + "source_page_131": 95, + "deck_group": "Supp S10", + "section": "Supp S10", + "previous_section_label": "Fig. 5 detail", + "parent_page": 110, + "parent_enriched_page": 119, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/sepsis_rare_support.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 121, + "parent_154_page": 133, + "parent_155_page": 134 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_sepsis_rare_direction", + "title": "Sepsis rare-state direction across support budgets", + "section_label": "Supp. S10", + "figures": [ + "figures/expanded_master_20261013/disease/sepsis_rare_direction.png" + ], + "keywords": [], + "footnote": "Saved sign-agreement counts across all 15 rules, including controls. Historical n + validation; directions are not effect-size accuracy.", + "footnote_full": "Counts summarize existing same_direction booleans; no thresholds or p-values are recomputed. All rule-family members are retained, including weak or failed states and positive/negative controls. A matching sign does not establish that a rare state's abundance or marker co-expression is reproduced.\nSources: plans/rare_states_screen_20261004/out/sepsis_contrasts.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_direction.csv; plans/rare_states_screen_20261004/README.md\nEvidence source: plans/rare_states_screen_20261004/out/sepsis_contrasts.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_direction.csv; plans/rare_states_screen_20261004/README.md", + "sources": "plans/rare_states_screen_20261004/out/sepsis_contrasts.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_direction.csv; plans/rare_states_screen_20261004/README.md", + "notes": "Counts summarize existing same_direction booleans; no thresholds or p-values are recomputed. All rule-family members are retained, including weak or failed states and positive/negative controls. A matching sign does not establish that a rare state's abundance or marker co-expression is reproduced.\nSources: plans/rare_states_screen_20261004/out/sepsis_contrasts.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/sepsis_rare_direction.csv; plans/rare_states_screen_20261004/README.md", + "after_original": 51, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_sepsis_rare_direction", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/sepsis_rare_direction.png" + ], + "source_page_131": 96, + "deck_group": "Supp S10", + "section": "Supp S10", + "previous_section_label": "Fig. 5 detail", + "parent_page": 111, + "parent_enriched_page": 120, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/sepsis_rare_direction.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 122, + "parent_154_page": 134, + "parent_155_page": 135 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_external_rare_screen", + "title": "Rare-state screen in DRESS, Kawasaki and SLE", + "section_label": "Supp. S11", + "figures": [ + "figures/expanded_master_20261013/disease/external_rare_screen.png" + ], + "keywords": [], + "footnote": "Historical one-seed zero-shot screen conditioned on each sample's own sc pseudobulk. DRESS was in training; real-bulk validation is separate.", + "footnote_full": "All 15 predefined states are shown. Samples' own sc pseudobulk conditions remove the real-bulk shift. Observed marker states are thinned to generated depth. DRESS includes cultured/stimulated samples and was seen during training. This is descriptive capability evidence, not an unseen-disease paired validation or Figure 6 reproduction.\nSources: plans/rare_states_screen_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/external_rare_screen.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json\nEvidence source: plans/rare_states_screen_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/external_rare_screen.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json", + "sources": "plans/rare_states_screen_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/external_rare_screen.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json", + "notes": "All 15 predefined states are shown. Samples' own sc pseudobulk conditions remove the real-bulk shift. Observed marker states are thinned to generated depth. DRESS includes cultured/stimulated samples and was seen during training. This is descriptive capability evidence, not an unseen-disease paired validation or Figure 6 reproduction.\nSources: plans/rare_states_screen_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/external_rare_screen.csv; plans/rare_states_screen_20261004/README.md; plans/rare_states_screen_20261004/tools/states.json", + "after_original": 51, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_external_rare_screen", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/external_rare_screen.png" + ], + "source_page_131": 97, + "deck_group": "Supp S11", + "section": "Supp S11", + "previous_section_label": "Rare-state appendix", + "parent_page": 112, + "parent_enriched_page": 121, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/external_rare_screen.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 123, + "parent_154_page": 135, + "parent_155_page": 136 + }, + { + "kind": "content", + "section": "Supp S11", + "section_label": "Supp. S11", + "title": "Consequence: disease direction flips", + "layout": "figure_keywords", + "figures": [ + "figures/final_palette_20261013/sec05_behcet_direction.png" + ], + "keywords": [ + "CD16 gap: raw \u22123.97 vs observed +4.94", + "C1Q flip vanishes at common depth", + "sc pseudobulk: right sign in 5/5 seeds" + ], + "footnote": "Historical Behcet diagnostic; reference is unmatched and includes training donors. This is not independent patient-level subtype validation.", + "say": "Raw bulk gives the wrong sign for CD16 and C1Q in 5/5 seeds. At one common monocyte depth the C1Q gap is \u22120.08 \u00b1 0.09 pp (1/5 seeds positive), so the C1Q flip is a depth effect; the CD16 flip stays at \u22123.59. sc pseudobulk of the 8 sc donors gives the observed sign in 5/5 seeds (CD16 +5.48, C1Q +1.82). Observed p is descriptive (4 vs 4). No paired data in this cohort, so no fine-tune here.", + "sources": "tools/sec05_behcet_direction.py; data/sec05_behcet_direction.csv", + "footnote_full": "Original slide 52: Beh\u00e7et, final model. Bulk GSE198533 (BD 9, HC 10), 5,000 cells per sample; observed = sc GSE198616 (BD 4, HC 4, different people; descriptive p). At common depth, raw CD16 is \u22123.59 (seed 0) and raw C1Q is \u22120.08 \u00b1 0.09 pp.", + "original_slide": 52, + "original_title": "Consequence: disease direction flips", + "original_footnote": "Beh\u00e7et. Bulk GSE198533 (9 BD / 10 HC); sc GSE198616 (4 / 4, different people) gives observed and pseudobulk.", + "figures_before_palette": [ + "figures/sec05_behcet_direction.png" + ], + "source_page_131": 98, + "deck_group": "Supp S11", + "previous_section_label": "Fig. 5d", + "parent_page": 113, + "parent_enriched_page": 122, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec05_behcet_direction.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 124, + "parent_154_page": 136, + "parent_155_page": 137 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_c1q_cross_disease", + "title": "C1Q: single-gene detection versus co-detection", + "section_label": "Supp. S11", + "figures": [ + "figures/expanded_master_20261013/disease/c1q_cross_disease.png" + ], + "keywords": [], + "footnote": "Existing pseudobulk-conditioned comparison at matched depth. Multi-gene C1Q rates remain compressed, including in-corpus DRESS.", + "footnote_full": "One sampling seed. DRESS has seven samples, including four cultured/stimulated samples; Kawasaki has 15 and SLE 56. Corpus inclusion is not randomized and DRESS samples were seen in training. Retain the distinction between any-gene detection and at least two C1Q genes in the same cell; no causal claim about pretraining exposure.\nSources: plans/c1q_crossdisease_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/c1q_cross_disease.csv; plans/c1q_crossdisease_20261004/README.md\nEvidence source: plans/c1q_crossdisease_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/c1q_cross_disease.csv; plans/c1q_crossdisease_20261004/README.md", + "sources": "plans/c1q_crossdisease_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/c1q_cross_disease.csv; plans/c1q_crossdisease_20261004/README.md", + "notes": "One sampling seed. DRESS has seven samples, including four cultured/stimulated samples; Kawasaki has 15 and SLE 56. Corpus inclusion is not randomized and DRESS samples were seen in training. Retain the distinction between any-gene detection and at least two C1Q genes in the same cell; no causal claim about pretraining exposure.\nSources: plans/c1q_crossdisease_20261004/out/summary.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/c1q_cross_disease.csv; plans/c1q_crossdisease_20261004/README.md", + "after_original": 52, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_c1q_cross_disease", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/c1q_cross_disease.png" + ], + "source_page_131": 99, + "deck_group": "Supp S11", + "section": "Supp S11", + "previous_section_label": "Rare-state appendix", + "parent_page": 114, + "parent_enriched_page": 123, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/c1q_cross_disease.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 125, + "parent_154_page": 137, + "parent_155_page": 138 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_behcet_k4000", + "title": "Beh\u00e7et: K=4,000 changes direction more than identity", + "section_label": "Supp. S11", + "figures": [ + "figures/final_palette_20261013/behcet_k4000.png" + ], + "keywords": [], + "footnote": "Two existing sampling seeds; 6 BD / 10 HC held-out bulks. K is condition size; each bulk still generates 5,000 cells. Descriptive.", + "footnote_full": "The real reference is 4 BD and 4 HC sc donors, mostly unmatched and including training donors. The three BD-only training candidates include two primary identity passes and one uncertain candidate. K4000 strengthens the any-C1Q group direction, but \u22652-gene cells remain far below observed. Same-pair calibration can also recover direction; this plot does not claim LoRA uniquely does so. No p-values are promoted.\nSources: plans/behcet_c1q_lora3_20261004/out/group_table.csv; plans/behcet_c1q_lora3_20261004/out/group_table_k4000.csv; plans/behcet_c1q_lora3_20261004/out/capability.csv; plans/behcet_c1q_lora3_20261004/out/capability_k4000.csv; plans/behcet_c1q_lora3_20261004/README.md; plans/paper_figures_strictpairs_20261012/reports/rare_behcet_provenance.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/behcet_k4000.csv\nEvidence source: plans/behcet_c1q_lora3_20261004/out/group_table.csv; plans/behcet_c1q_lora3_20261004/out/group_table_k4000.csv; plans/behcet_c1q_lora3_20261004/out/capability.csv; plans/behcet_c1q_lora3_20261004/out/capability_k4000.csv; plans/behcet_c1q_lora3_20261004/README.md; plans/paper_figures_strictpairs_20261012/reports/rare_behcet_provenance.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/behcet_k4000.csv", + "sources": "plans/behcet_c1q_lora3_20261004/out/group_table.csv; plans/behcet_c1q_lora3_20261004/out/group_table_k4000.csv; plans/behcet_c1q_lora3_20261004/out/capability.csv; plans/behcet_c1q_lora3_20261004/out/capability_k4000.csv; plans/behcet_c1q_lora3_20261004/README.md; plans/paper_figures_strictpairs_20261012/reports/rare_behcet_provenance.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/behcet_k4000.csv", + "notes": "The real reference is 4 BD and 4 HC sc donors, mostly unmatched and including training donors. The three BD-only training candidates include two primary identity passes and one uncertain candidate. K4000 strengthens the any-C1Q group direction, but \u22652-gene cells remain far below observed. Same-pair calibration can also recover direction; this plot does not claim LoRA uniquely does so. No p-values are promoted.\nSources: plans/behcet_c1q_lora3_20261004/out/group_table.csv; plans/behcet_c1q_lora3_20261004/out/group_table_k4000.csv; plans/behcet_c1q_lora3_20261004/out/capability.csv; plans/behcet_c1q_lora3_20261004/out/capability_k4000.csv; plans/behcet_c1q_lora3_20261004/README.md; plans/paper_figures_strictpairs_20261012/reports/rare_behcet_provenance.json; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/behcet_k4000.csv", + "after_original": 52, + "palette": "native", + "new_experiment": false, + "addition_key": "disease_behcet_k4000", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/behcet_k4000.png" + ], + "source_page_131": 100, + "deck_group": "Supp S11", + "section": "Supp S11", + "previous_section_label": "Beh\u00e7et diagnostic", + "parent_page": 115, + "parent_enriched_page": 124, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/behcet_k4000.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 126, + "parent_154_page": 138, + "parent_155_page": 139 + }, + { + "kind": "content", + "title": "Behcet: one-gene and multi-gene C1Q rules", + "section_label": "Supp. S11", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/review_style_20261013_behcet.png" + ], + "keywords": [], + "footnote": "Descriptive: unmatched reference includes training donors. Same three pairs; two identity-pass and one uncertain.", + "footnote_full": "Additional view from the audited results review. Descriptive: unmatched reference includes training donors. Same three pairs; two identity-pass and one uncertain.", + "sources": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/review_style_20261013_behcet.pdf", + "addition_key": "review:24", + "after_original": 52, + "figures_before_palette": [ + "figures/review_style_20261013_behcet.png" + ], + "source_page_131": 101, + "deck_group": "Supp S11", + "section": "Supp S11", + "previous_section_label": "Fig. 5c", + "parent_page": 116, + "parent_enriched_page": 125, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/review_style_20261013_behcet.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 127, + "parent_154_page": 139, + "parent_155_page": 140 + }, + { + "kind": "content", + "section": "Supp S12", + "section_label": "Supp. S12", + "layout": "full", + "sources": "tools/sec08_cross_cohort.py; data/sec08_cross_cohort.csv, data/sec08_cross_cohort_steps.csv; plans/paired_crosscohort_20261002/ (README, scores/aggregate.csv)", + "text_source": { + "file": "data/sec08_cross_cohort_slide_text.json", + "index": 0 + }, + "title": "Pairs from other cohorts do not transfer", + "figures": [ + "figures/final_palette_20261013/sec08_cross_cohort.png" + ], + "keywords": [ + "Train on the other cohorts, test the held-out one", + "Calibration ties or beats them on composition in 4/5", + "Own-cohort pairs best on coverage 5/5, disease r 4/4" + ], + "footnote": "Completed leave-one-cohort-out test at K=1000; target-cohort sc was excluded from training.", + "say": "Why one LoRA per cohort: a model trained on the other PBMC cohorts' pairs (three, or four for Ji, plus Liu 2021) rarely beats a zero-pair gene-wise calibration on composition, and no arm without the cohort's own pairs matches the own-cohort LoRA on coverage or disease signal.", + "footnote_full": "Current scope: Completed leave-one-cohort-out test at K=1000; target-cohort sc was excluded from training.\n\nOriginal slide 68: Each held-out cohort is scored on all its usable samples against the same frozen reference as its within-cohort benchmark. Zero-shot = base model on real bulk. Calibration from other cohorts = gene-wise alpha (mean pseudobulk CPM / mean bulk CPM) fitted on the training cohorts' pairs, cohort-balanced; 0 target pairs. LoRA r16 / r64 / full fine-tune = trained on the other PBMC cohorts' pairs plus Liu 2021 (training only: its cells are in the pretraining corpus; 44 pairs after excluding 14 hashtag-negative-heavy rows), cohort-balanced batches, inner-validation on 10 % of each training cohort's donors, step-saturation protocol (20k-step blocks, stop when a block lowers the inner-validation loss by < 1 % of the total), value at the last milestone (number in the bar = steps). Ji's bulk is whole blood and is never trained on. Own-cohort LoRA = LoRA r16 on the cohort's own pairs, 5-fold pooled. Giroux's logFC r excludes 196 sex-linked genes. Best on composition among zero-target-pair arms: ImMicro Calibration from other cohorts, Sepsis Calibration from other cohorts, Imperial Full fine-tune, other cohorts, Giroux Calibration from other cohorts, Ji (whole blood) Calibration from other cohorts.", + "numbers": [ + { + "value": "ImMicro: L1 zs 0.446 / xcal 0.172 / lora16sat 0.175 / lora64sat 0.199 / fullsat 0.213 / own 0.162; coverage zs 0.19 / xcal 0.57 / lora16sat 0.45 / lora64sat 0.38 / fullsat 0.35 / own 0.78", + "source": "data/sec08_cross_cohort.csv" + }, + { + "value": "Sepsis: L1 zs 0.509 / xcal 0.444 / lora16sat 0.491 / lora64sat 0.570 / fullsat 0.585 / own 0.362; coverage zs 0.07 / xcal 0.46 / lora16sat 0.34 / lora64sat 0.29 / fullsat 0.35 / own 0.83; logFC r zs -0.02 / xcal 0.10 / lora16sat 0.10 / lora64sat 0.23 / fullsat 0.07 / own 0.53", + "source": "data/sec08_cross_cohort.csv" + }, + { + "value": "Imperial: L1 zs 0.803 / xcal 0.651 / lora16sat 0.635 / lora64sat 0.594 / fullsat 0.580 / own 0.376; coverage zs 0.11 / xcal 0.64 / lora16sat 0.68 / lora64sat 0.72 / fullsat 0.61 / own 0.85; logFC r zs 0.64 / xcal 0.69 / lora16sat 0.49 / lora64sat 0.58 / fullsat 0.46 / own 0.82", + "source": "data/sec08_cross_cohort.csv" + }, + { + "value": "Giroux: L1 zs 0.337 / xcal 0.202 / lora16sat 0.431 / lora64sat 0.410 / fullsat 0.465 / own 0.258; coverage zs 0.29 / xcal 0.89 / lora16sat 0.82 / lora64sat 0.81 / fullsat 0.78 / own 0.93; logFC r zs 0.34 / xcal 0.24 / lora16sat 0.30 / lora64sat 0.28 / fullsat 0.20 / own 0.46", + "source": "data/sec08_cross_cohort.csv" + }, + { + "value": "Ji (whole blood): L1 zs 0.725 / xcal 0.565 / lora16sat 0.740 / lora64sat 0.620 / fullsat 1.030 / own 0.504; coverage zs 0.06 / xcal 0.07 / lora16sat 0.34 / lora64sat 0.40 / fullsat 0.08 / own 0.86; logFC r zs 0.07 / xcal 0.13 / lora16sat 0.05 / lora64sat 0.16 / fullsat 0.14 / own 0.48", + "source": "data/sec08_cross_cohort.csv" + } + ], + "original_slide": 68, + "original_title": "Pairs from other cohorts do not transfer", + "original_footnote": "Leave-one-cohort-out, K = 1000; train on the other cohorts (+ Liu 2021), test the held-out one.", + "figures_before_palette": [ + "figures/sec08_cross_cohort.png" + ], + "source_page_131": 123, + "deck_group": "Supp S12", + "previous_section_label": "Supplementary", + "parent_page": 117, + "parent_enriched_page": 126, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec08_cross_cohort.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 128, + "parent_154_page": 140, + "parent_155_page": 141 + }, + { + "kind": "content", + "section": "Supp S12", + "section_label": "Supp. S12", + "layout": "full", + "sources": "tools/sec08_cross_cohort.py; data/sec08_cross_cohort.csv (minutes = RUN.json wall_seconds / 60), data/sec08_cross_cohort_steps.csv; plans/paired_crosscohort_20261002/ (README, scores/aggregate.csv, runs/hoL_/K1000//RUN.json)", + "text_source": { + "file": "data/sec08_cross_cohort_slide_text.json", + "index": 1 + }, + "title": "Longer training on other cohorts: gap stays", + "figures": [ + "figures/final_palette_20261013/sec08_cross_cohort_steps.png" + ], + "keywords": [ + "To saturation: r16 20\u201331 min, full 59\u2013107 min", + "Full fine-tune drifts in Sepsis and Ji (Ji L1 0.84 \u2192 1.03)", + "So PBISC fits LoRA on a few pairs per cohort" + ], + "footnote": "Existing saturation experiment; no additional saturation run introduced here.", + "say": "Each run trained until its inner-validation loss saturated. More steps on other cohorts never closed the coverage gap to the cohort's own pairs: Imperial and Giroux improved a little, and full fine-tuning drifted further in Sepsis and Ji (whole-blood bulk).", + "footnote_full": "Current scope: Existing saturation experiment; no additional saturation run introduced here.\n\nOriginal slide 69: Same runs as the previous slide. Each point = the milestone checkpoint at the end of a 20k-step block (or at the early stop), sampled and scored on the held-out cohort. Dashed = calibration from other cohorts; dotted = LoRA on the cohort's own pairs. Training stopped (steps): LoRA r16 ImMicro 28k, Sepsis 26k, Imperial 40k, Giroux 35k, Ji 40k; LoRA r64 ImMicro 39k, Sepsis 59k, Imperial 69k, Giroux 69k, Ji 51k; Full fine-tune ImMicro 65k, Sepsis 108k, Imperial 93k, Giroux 120k, Ji 86k. Saturation steps across cohorts: r16 up to 40k, r64 up to 69k, full up to 120k. Training time to saturation (RUN.json wall seconds, one GB10): LoRA r16 ImMicro 21 min, Sepsis 20 min, Imperial 30 min, Giroux 28 min, Ji 31 min; LoRA r64 ImMicro 27 min, Sepsis 44 min, Imperial 54 min, Giroux 53 min, Ji 41 min; Full fine-tune ImMicro 59 min, Sepsis 93 min, Imperial 82 min, Giroux 107 min, Ji 77 min. Gap = no milestone of any mode reaches the own-cohort LoRA's coverage (true in all five cohorts); full fine-tune composition L1 from the first to the last milestone: ImMicro +0.005, Sepsis +0.087, Imperial -0.052, Giroux -0.078, Ji +0.192 (drift >= 0.05 in Sepsis and Ji; improved by >= 0.05 in Imperial and Giroux).", + "numbers": [ + { + "value": "ImMicro LoRA r16: 20k L1 0.163 cov 0.45, 28k L1 0.175 cov 0.45", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "ImMicro LoRA r64: 20k L1 0.196 cov 0.44, 39k L1 0.199 cov 0.38", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "ImMicro Full fine-tune: 20k L1 0.208 cov 0.39, 40k L1 0.211 cov 0.37, 60k L1 0.223 cov 0.35, 65k L1 0.213 cov 0.35", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Sepsis LoRA r16: 20k L1 0.500 cov 0.35, 26k L1 0.491 cov 0.34", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Sepsis LoRA r64: 20k L1 0.550 cov 0.29, 40k L1 0.564 cov 0.29, 59k L1 0.570 cov 0.29", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Sepsis Full fine-tune: 20k L1 0.498 cov 0.33, 40k L1 0.530 cov 0.33, 60k L1 0.546 cov 0.34, 80k L1 0.576 cov 0.34, 100k L1 0.577 cov 0.35, 108k L1 0.585 cov 0.35", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Imperial LoRA r16: 20k L1 0.657 cov 0.68, 40k L1 0.635 cov 0.68", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Imperial LoRA r64: 20k L1 0.622 cov 0.68, 40k L1 0.603 cov 0.71, 60k L1 0.577 cov 0.72, 69k L1 0.594 cov 0.72", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Imperial Full fine-tune: 20k L1 0.632 cov 0.60, 40k L1 0.609 cov 0.60, 60k L1 0.565 cov 0.61, 80k L1 0.576 cov 0.60, 93k L1 0.580 cov 0.61", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Giroux LoRA r16: 20k L1 0.416 cov 0.79, 35k L1 0.431 cov 0.82", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Giroux LoRA r64: 20k L1 0.491 cov 0.80, 40k L1 0.469 cov 0.82, 60k L1 0.430 cov 0.82, 69k L1 0.410 cov 0.81", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Giroux Full fine-tune: 20k L1 0.543 cov 0.75, 40k L1 0.532 cov 0.75, 60k L1 0.499 cov 0.76, 80k L1 0.466 cov 0.76, 100k L1 0.490 cov 0.76, 120k L1 0.465 cov 0.78", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Ji (whole blood) LoRA r16: 20k L1 0.757 cov 0.27, 40k L1 0.740 cov 0.34", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Ji (whole blood) LoRA r64: 20k L1 0.639 cov 0.38, 40k L1 0.668 cov 0.39, 51k L1 0.620 cov 0.40", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Ji (whole blood) Full fine-tune: 20k L1 0.838 cov 0.17, 40k L1 1.006 cov 0.11, 60k L1 1.032 cov 0.08, 80k L1 1.052 cov 0.07, 86k L1 1.030 cov 0.08", + "source": "data/sec08_cross_cohort_steps.csv" + }, + { + "value": "Training time to saturation, LoRA r16: ImMicro 21 min, Sepsis 20 min, Imperial 30 min, Giroux 28 min, Ji 31 min", + "source": "data/sec08_cross_cohort.csv (minutes = RUN.json wall_seconds / 60)" + }, + { + "value": "Training time to saturation, LoRA r64: ImMicro 27 min, Sepsis 44 min, Imperial 54 min, Giroux 53 min, Ji 41 min", + "source": "data/sec08_cross_cohort.csv (minutes = RUN.json wall_seconds / 60)" + }, + { + "value": "Training time to saturation, Full fine-tune: ImMicro 59 min, Sepsis 93 min, Imperial 82 min, Giroux 107 min, Ji 77 min", + "source": "data/sec08_cross_cohort.csv (minutes = RUN.json wall_seconds / 60)" + } + ], + "original_slide": 69, + "original_title": "Longer training on other cohorts: gap stays", + "original_footnote": "Test metrics at every 20k-step milestone (best inner-validation state so far). K = 1000.", + "figures_before_palette": [ + "figures/sec08_cross_cohort_steps.png" + ], + "source_page_131": 124, + "deck_group": "Supp S12", + "previous_section_label": "Supplementary", + "parent_page": 118, + "parent_enriched_page": 127, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec08_cross_cohort_steps.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 129, + "parent_154_page": 141, + "parent_155_page": 142 + }, + { + "kind": "content", + "section": "Supp S12", + "section_label": "Supp. S12", + "text_source": { + "file": "data/sec07_no_pretraining_slide_text.json", + "index": 0 + }, + "title": "No pretraining: same cohort, from scratch", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec07_no_pretraining.png" + ], + "keywords": [ + "Same realism; 5 pairs: donor r 0.70 vs 0.17", + "58 pairs: scratch logFC r 0.58 vs 0.49", + "Scratch 61\u201386 min, LoRA 3\u20135 min" + ], + "footnote": "Completed Sepsis fold0 pilot, not all folds. Pretraining supports speed/few-pair adaptation, not universal superiority.", + "sources": "tools/sec07_no_pretraining.py; data/sec07_no_pretraining.csv; plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/fold0_scratch_20261005/", + "say": "Pilot on one sepsis fold. Training the same model from scratch on the cohort's pairs gives realistic cells, and with 58 pairs a higher disease logFC r, but with few pairs it cannot place donors: what pretraining buys is the donor-specific composition and minutes instead of an hour per cohort. Folds 1-4 are needed before any claim.", + "footnote_full": "Current scope: Completed Sepsis fold0 pilot, not all folds. Pretraining supports speed/few-pair adaptation, not universal superiority.\n\nOriginal slide 70: Pilot on one fold: sepsis fold 0, 17 test donors (12 sepsis, 5 controls); folds 1-4 not run, so no claim yet. From scratch = the same denoiser (54.2 M weights) re-initialised at random (tools/ft_sdd_aug.py --init random; seed 20260627) and trained in full on the cohort's own pairs at the pretraining lr 2e-4, batch 256, EMA 0.999, inner-val selection every 1,000 steps, patience 10, cap 100k steps; the checkpoint supplies only the architecture, 6k panel, value scaler and bulk z-score constants. Pretrained + LoRA = f01 6k 1x E16 + LoRA r16 (lr 5e-4). Same training pairs (5 or 58, + 7-8 inner-val), real-bulk conditions K = 1,000, test donors and sampling seed; recipes = K-diversified pseudobulk augmentation. The 58-pair scratch arms were still improving at the 100k cap. Coverage = PCA-50 k = 10 balls (precision, C2ST, MMD, E_N agree; scores_extra/ye2026_sepsis/fold0_scratch_20261005/). Donor composition r = per cell type (T, B, Mono, NK) Pearson r over the 17 donors of generated vs observed fraction, mean of the four. Dashed = zero-shot. Training time = RUN.json wall seconds on one GB10 (LoRA also needs the one-time 15.9 h pretraining, which every cohort reuses). Source: plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/fold0_scratch_20261005/ and README 'No-pretraining ablation'.", + "numbers": [ + { + "value": "5 pairs: pretrained+LoRA L1 0.363 / scratch 0.448; coverage 0.81 / 0.80; logFC r 0.23 / 0.18; donor composition r 0.70 / 0.17; time 3.2 / 61 min; steps (selected/run) 2750/4750 vs 66000/76000", + "source": "data/sec07_no_pretraining.csv" + }, + { + "value": "58 pairs: pretrained+LoRA L1 0.373 / scratch 0.398; coverage 0.86 / 0.86; logFC r 0.49 / 0.58; donor composition r 0.77 / 0.68; time 5.4 / 79 min; steps (selected/run) 6500/8500 vs 100000/100000", + "source": "data/sec07_no_pretraining.csv" + }, + { + "value": "58 + recipes pairs: pretrained+LoRA L1 0.385 / scratch 0.385; coverage 0.87 / 0.88; logFC r 0.45 / 0.54; donor composition r 0.80 / 0.77; time 4.5 / 86 min; steps (selected/run) 3750/5750 vs 100000/100000", + "source": "data/sec07_no_pretraining.csv" + } + ], + "original_slide": 70, + "original_title": "No pretraining: same cohort, from scratch", + "original_footnote": "Pilot: sepsis fold 0, 17 test donors (12 sepsis, 5 controls). Folds 1\u20134 not run yet.", + "figures_before_palette": [ + "figures/sec07_no_pretraining.png" + ], + "source_page_131": 125, + "deck_group": "Supp S12", + "previous_section_label": "Supplementary", + "parent_page": 119, + "parent_enriched_page": 128, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec07_no_pretraining.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 130, + "parent_154_page": 142, + "parent_155_page": 143 + }, + { + "kind": "content", + "section": "Supp S12", + "section_label": "Supp. S12", + "text_source": { + "file": "data/sec08_cross_scratch_slide_text.json", + "index": 0 + }, + "title": "No pretraining: other cohorts, from scratch", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec08_cross_scratch.png" + ], + "keywords": [ + "Scratch \u2248 LoRA r16: coverage 0.33 vs 0.34", + "Own-cohort pairs far ahead: coverage 0.83", + "Scratch 127 min vs LoRA r16 20 min" + ], + "footnote": "Pilot: held-out sepsis, 82 samples. Genes no training cohort measured are zeroed in every arm.", + "sources": "tools/sec08_cross_scratch.py; data/sec08_cross_scratch.csv; plans/paired_crosscohort_20261002/scores_masked/ye2026_sepsis/, generated_masked/ye2026_sepsis/hoL_ye2026_sepsis/*.log, runs/hoL_ye2026_sepsis/K1000/*/RUN.json; plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis/cells_panel6k.npz (observed share)", + "say": "Held-out sepsis pilot. Training from scratch on the other cohorts' pairs ends where the pretrained arms end: no arm without sepsis's own pairs gets near the own-cohort LoRA. Genes that no training cohort measured are zeroed in every arm, because the scratch model never learned them.", + "footnote_full": "Original slide 71: Pilot, one held-out cohort: sepsis (82 samples), training pool ImMicro + Imperial + Giroux + Liu 2021, K = 1,000, same inner-val selection and sampling as the 'Pairs from other cohorts do not transfer' and 'Longer training on other cohorts' slides. From scratch = the same denoiser at random initialisation (--init random), full training at lr 2e-4, inner validation every 1,000 steps (patience 10), milestones every 50k steps without the 1% saturation rule; stopped at 162k (selected 152k; 127 min on one GB10). The pretrained arms: inner validation every 250 steps (patience 8), 20k-step blocks with the 1% rule; all four runs ended by patience early stopping. Every arm is scored with the 46 panel genes that no training cohort measured set to 0 in the generated cells (tools/mask_rescore.py): the scratch model never gets a loss on them and puts ~94% of its UMI there (91\u201394% across milestones; observed sepsis cells 0.014%); the pretrained arms barely change (e.g. LoRA r16 coverage 0.34 masked vs 0.34 unmasked). Calibration needs no training. Own-cohort LoRA = 5-fold LoRA on sepsis pairs (mean 6.3 min per fold). Scratch milestones: L1 0.492 / 0.497 / 0.510 / 0.499; coverage 0.339 / 0.343 / 0.330 / 0.328; logFC r 0.247 / 0.245 / 0.247 / 0.229. Source: plans/paired_crosscohort_20261002/scores_masked/ye2026_sepsis/ (table_masked.py), runs/hoL_ye2026_sepsis/K1000/.", + "numbers": [ + { + "value": "Calibration: L1 0.444, coverage 0.461, logFC r 0.096, top-50 DEG 10, time none", + "source": "data/sec08_cross_scratch.csv" + }, + { + "value": "LoRA r16: L1 0.491, coverage 0.345, logFC r 0.106, top-50 DEG 4, time 20 min", + "source": "data/sec08_cross_scratch.csv" + }, + { + "value": "LoRA r64: L1 0.568, coverage 0.284, logFC r 0.226, top-50 DEG 5, time 44 min", + "source": "data/sec08_cross_scratch.csv" + }, + { + "value": "Full FT: L1 0.585, coverage 0.340, logFC r 0.084, top-50 DEG 3, time 93 min", + "source": "data/sec08_cross_scratch.csv" + }, + { + "value": "From scratch: L1 0.499, coverage 0.328, logFC r 0.229, top-50 DEG 1, time 127 min", + "source": "data/sec08_cross_scratch.csv" + }, + { + "value": "Own-cohort LoRA: L1 0.362, coverage 0.828, logFC r 0.526, top-50 DEG 28, time 6 min", + "source": "data/sec08_cross_scratch.csv" + }, + { + "value": "Scratch milestones (masked) 50k / 100k / 150k / 162k steps: L1 0.492 / 0.497 / 0.510 / 0.499; coverage 0.339 / 0.343 / 0.330 / 0.328; logFC r 0.247 / 0.245 / 0.247 / 0.229", + "source": "data/sec08_cross_scratch.csv" + } + ], + "original_slide": 71, + "original_title": "No pretraining: other cohorts, from scratch", + "original_footnote": "Pilot: held-out sepsis, 82 samples. Genes no training cohort measured are zeroed in every arm.", + "figures_before_palette": [ + "figures/sec08_cross_scratch.png" + ], + "source_page_131": 126, + "deck_group": "Supp S12", + "previous_section_label": "Supplementary", + "parent_page": 120, + "parent_enriched_page": 129, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec08_cross_scratch.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 131, + "parent_154_page": 143, + "parent_155_page": 144 + }, + { + "kind": "content", + "section": "Supp S12", + "section_label": "Supp. S12", + "text_source": { + "file": "data/secS_scratch_rare_states_slide_text.json", + "index": 0 + }, + "title": "Rare states: from scratch vs pretrained", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/secS_scratch_rare_states.png" + ], + "keywords": [ + "Same pairs: no significant difference on average", + "Cycling T/NK: pretrained closer, 8 of 8", + "C1Q-high needs the cohort's own pairs" + ], + "footnote": "Pilot, one seed: sepsis fold 0 (17 donors), held-out sepsis (82, masked). 1 = observed rate; grey = within 2-fold.", + "say": "Supplementary pilot: one sepsis fold and one held-out cohort, one seed. With the same pairs, the model trained from scratch shows no significant difference from the pretrained model on average over 13 rare states, within about plus or minus 0.3 log2. Pretraining stays closer on cycling T and NK cells, in both panels; scratch is closer on DN2 B cells and where the pretrained model over-calls labels. In this fold, C1Q-high monocytes reach about half the observed rate or more only with the cohort's own pairs, with either initialisation.", + "sources": "plans/rare_states_screen_20261004/out/{sepsis_fold0,sepsis_loco_masked}_*.csv; plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis/folds.csv; plans/paired_finetune_intradataset_20260927/runs/ye2026_sepsis/fold0/scratch_*/RUN.json; plans/paired_crosscohort_20261002/generated_masked/ye2026_sepsis/hoL_ye2026_sepsis/*.log; tools/secS_scratch_rare_states.py; data/secS_scratch_rare_states{,_bootstrap}.csv", + "footnote_full": "Original slide 72: Supplementary. Rare-state screen (plans/rare_states_screen_20261004): 13 rare states plus two controls, marker rules from the literature (tools/states.json) or scANVI labels; per arm, generated state % of its parent / observed % (1 = same rate; grey band = within 2-fold). For marker states the observed cells are thinned to the arm's generated depth; the two label states (plasmablast, proliferating T) are not thinned. The figure shows 8 of the 13 states; every average uses all 13 (data/secS_scratch_rare_states_bootstrap.csv). Same cohort: sepsis fold 0, 17 held-out donors, the same training pairs, conditions and sampling seed for both initialisations; from scratch = random initialisation, full training (lr 2e-4, cap 100k steps). Other cohorts: all 82 sepsis donors held out, trained on ImMicro + Imperial + Giroux + Liu pairs; the 46 genes that no training cohort measured are zeroed in every arm (the scratch model is never trained on them). Donor bootstrap of D = |log2| scratch - |log2| pretrained, mean of the 13 rare states (D > 0 = pretrained closer): same cohort 5 pairs +0.22 [-0.08, +0.41]; 58 pairs +0.02 [-0.35, +0.25]; 58 pairs, full FT +0.14 [-0.14, +0.41]; 58 pairs + recipes -0.09 [-0.38, +0.18]; other cohorts, scratch final vs LoRA r16 -0.04 [-0.20, +0.18]; LoRA r64 +0.06 [-0.08, +0.22]; Full FT +0.29 [+0.07, +0.57]; Calibration +0.72 [+0.57, +0.95]; scratch 50k vs LoRA r16 -0.31 [-0.45, -0.11]; LoRA r64 -0.21 [-0.37, -0.02]; Full FT +0.02 [-0.19, +0.30]; Calibration +0.44 [+0.30, +0.66]. A non-significant average is not a shown equivalence (intervals about +/-0.3 log2). Cycling T/NK: pretrained closer in 4 of 4 same-cohort settings (1 significant) and 4 of 4 other-cohort comparisons (4 significant; scratch 0.08x). ISG-high T: pretrained closer in 4 of 4 same-cohort settings (1 significant) but scratch closer in 4 of 4 other-cohort comparisons (0 significant; scratch 1.10x vs pretrained 1.54\u20131.98x). Same-cohort state-level calls (13 states x 4 settings): pretrained closer 3, scratch closer 7 of 52; of the scratch calls 4 are label over-calls by the pretrained arms (Plasmablast, Proliferating T), 2 are states where both arms are more than 4-fold off (Granulocyte-like), and 1 is DN2 B (58 pairs + recipes -0.70 [-1.21, -0.08]); scratch is also closer on MS1 at 58 pairs (-0.97 [-1.44, +0.03], n.s.). C1Q-high: 0.47x or more with the cohort's own pairs (both initialisations), 0.10x or less without them. Fold 0 is the own-pair LoRA's best fold for C1Q-high and cycling T/NK (58 pairs: 1.03x and 1.00x vs 0.27\u20130.82x and 0.32\u20130.64x in folds 1\u20134); scratch has not been run on folds 1\u20134. One fold, one held-out cohort, one seed; no multiplicity correction. Full analysis: plans/rare_states_screen_20261004/README.md and out/scratch_vs_lora_20261005/REPORT.md (two verification passes on the same-cohort and unmasked analyses); the masked cross-cohort numbers here come from the REPORT addendum and masked_cross_bootstrap.csv, which supersede the unmasked cross-cohort readings in the REPORT body.", + "numbers": [ + { + "value": "C1Q-high mono: same cohort 5 pairs LoRA 1.07 / scratch 1.01, 58 pairs LoRA 1.05 / scratch 0.47, 58 pairs + recipes LoRA 0.63 / scratch 0.53; other cohorts Own-cohort LoRA (reference) 0.51, Calibration 0.08, LoRA r16 0.04, Full FT 0.08, From scratch 0.07", + "source": "data/secS_scratch_rare_states.csv" + }, + { + "value": "MS1 mono: same cohort 5 pairs LoRA 0.15 / scratch 0.10, 58 pairs LoRA 0.21 / scratch 0.40, 58 pairs + recipes LoRA 0.30 / scratch 0.46; other cohorts Own-cohort LoRA (reference) 0.72, Calibration 0.73, LoRA r16 0.12, Full FT 0.27, From scratch 0.16", + "source": "data/secS_scratch_rare_states.csv" + }, + { + "value": "Cycling mono: same cohort 5 pairs LoRA 0.11 / scratch 0.06, 58 pairs LoRA 0.09 / scratch 0.09, 58 pairs + recipes LoRA 0.10 / scratch 0.09; other cohorts Own-cohort LoRA (reference) 0.24, Calibration 0.26, LoRA r16 0.07, Full FT 0.12, From scratch 0.04", + "source": "data/secS_scratch_rare_states.csv" + }, + { + "value": "Cycling T/NK: same cohort 5 pairs LoRA 1.15 / scratch 0.51, 58 pairs LoRA 0.93 / scratch 0.42, 58 pairs + recipes LoRA 0.99 / scratch 0.48; other cohorts Own-cohort LoRA (reference) 0.52, Calibration 0.48, LoRA r16 0.13, Full FT 0.36, From scratch 0.08", + "source": "data/secS_scratch_rare_states.csv" + }, + { + "value": "ISG-high T: same cohort 5 pairs LoRA 0.46 / scratch 0.15, 58 pairs LoRA 0.37 / scratch 0.17, 58 pairs + recipes LoRA 0.53 / scratch 0.18; other cohorts Own-cohort LoRA (reference) 0.58, Calibration 1.54, LoRA r16 1.71, Full FT 1.98, From scratch 1.10", + "source": "data/secS_scratch_rare_states.csv" + }, + { + "value": "Plasmablast (label): same cohort 5 pairs LoRA 1.87 / scratch 1.12, 58 pairs LoRA 3.47 / scratch 1.24, 58 pairs + recipes LoRA 2.30 / scratch 1.18; other cohorts Own-cohort LoRA (reference) 0.76, Calibration 0.78, LoRA r16 1.38, Full FT 2.21, From scratch 0.87", + "source": "data/secS_scratch_rare_states.csv" + }, + { + "value": "DN2 B: same cohort 5 pairs LoRA 1.11 / scratch 1.63, 58 pairs LoRA 0.35 / scratch 0.63, 58 pairs + recipes LoRA 0.28 / scratch 0.46; other cohorts Own-cohort LoRA (reference) 0.29, Calibration 0.28, LoRA r16 0.14, Full FT 0.29, From scratch 0.27", + "source": "data/secS_scratch_rare_states.csv" + }, + { + "value": "Granulocyte-like: same cohort 5 pairs LoRA 0.05 / scratch 0.13, 58 pairs LoRA 0.04 / scratch 0.04, 58 pairs + recipes LoRA 0.03 / scratch 0.06; other cohorts Own-cohort LoRA (reference) 0.06, Calibration 0.07, LoRA r16 0.02, Full FT 0.01, From scratch 0.02", + "source": "data/secS_scratch_rare_states.csv" + } + ], + "original_slide": 72, + "original_title": "Rare states: from scratch vs pretrained", + "original_footnote": "Pilot, one seed: sepsis fold 0 (17 donors), held-out sepsis (82, masked). 1 = observed rate; grey = within 2-fold.", + "figures_before_palette": [ + "figures/secS_scratch_rare_states.png" + ], + "source_page_131": 127, + "deck_group": "Supp S12", + "previous_section_label": "Supplementary", + "parent_page": 121, + "parent_enriched_page": 130, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/secS_scratch_rare_states.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 132, + "parent_154_page": 144, + "parent_155_page": 145 + }, + { + "kind": "content", + "section": "Supp S13", + "section_label": "Supp. S13", + "text_source": { + "file": "data/secA_scsemi_readout_slide_text.json", + "index": 0 + }, + "title": "Appendix: scSemiProfiler without output threshold", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/secA_scsemi_readout.png" + ], + "keywords": [ + "Official output zeroes values \u2264 10 (fixed)", + "Removed: scSemi coverage up in 5/5 cohorts", + "No LoRA-5 comparison flips (23/23 same)" + ], + "footnote": "Completed modified-readout check; main results use official output. Models were not retrained for the sensitivity.", + "say": "Appendix only (sensitivity check). The threshold is part of scSemiProfiler's official output, so removing it is a modified run; the main comparison stays the official output. This slide shows that our conclusions do not depend on that read-out.", + "sources": "plans/paired_finetune_intradataset_20260927/scsemi/tools/{resample_output_variants,compare_output_variants}.py; tools/secA_scsemi_readout.py; data/secA_scsemi_readout.csv", + "footnote_full": "Current scope: Completed modified-readout check; main results use official output. Models were not retrained for the sensitivity.\n\nOriginal slide 74: Appendix, sensitivity only. scSemiProfiler's official output (upstream commit c2aea02) is the decoder's ZINB mean with every value <= 10 normalized units set to 0; the cut is fixed in the code (vaegan.myVAE.sample, bound = 10, and fast_semi: xsemi * (xsemi > 10)), not an option. At our cohorts' depth (about 1,000-2,000 UMI per cell on the panel) it removes most low- and mid-expressed genes. Here the same trained per-target models were read out again without that cut (CPU, no retraining, target seeds as in the run), exported and scored exactly as the official arm (same 25 five-donor draws, donors, reference; Giroux disease metrics without sex-linked genes). A control re-read with the cut reproduces the official output: pseudobulk r >= 0.996 over 1790 targets, coverage within 0.015. Because this changes the method's official read-out, the main slides keep the official output. Bars = mean of 5 draws, dots = draws. Source: plans/paired_finetune_intradataset_20260927/scsemi/analysis/output_variants_20261003/ (README section 'Read-out sensitivity').", + "numbers": [ + { + "value": "ImMicro: coverage LoRA-5 0.79, official 0.04, no threshold 0.23; precision 0.95 / 0.52 / 0.74; L1 0.176 / 0.233 / 0.199", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/output_variants_20261003/per_draw.csv" + }, + { + "value": "Sepsis: coverage LoRA-5 0.69, official 0.08, no threshold 0.18; precision 0.91 / 0.73 / 0.77; L1 0.466 / 0.612 / 0.591; logFC r 0.35 / 0.40 / 0.46; top-50 DEG 16.4 / 5.2 / 14.0", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/output_variants_20261003/per_draw.csv" + }, + { + "value": "Imperial: coverage LoRA-5 0.74, official 0.06, no threshold 0.42; precision 0.85 / 0.33 / 0.72; L1 0.503 / 0.602 / 0.603; logFC r 0.75 / 0.56 / 0.66; top-50 DEG 113.6 / 46.8 / 94.6", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/output_variants_20261003/per_draw.csv" + }, + { + "value": "Giroux: coverage LoRA-5 0.96, official 0.61, no threshold 0.83; precision 1.00 / 0.99 / 1.00; L1 0.256 / 0.270 / 0.273; logFC r 0.30 / 0.40 / 0.45; top-50 DEG 0.2 / 0.0 / 0.0", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/output_variants_20261003/per_draw.csv" + }, + { + "value": "Ji: coverage LoRA-5 0.71, official 0.15, no threshold 0.53; precision 0.98 / 0.84 / 0.92; L1 0.513 / 0.592 / 0.575; logFC r 0.33 / 0.15 / 0.19; top-50 DEG 0.2 / 0.0 / 0.0", + "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scsemi/analysis/output_variants_20261003/per_draw.csv" + } + ], + "original_slide": 74, + "original_title": "Appendix: scSemiProfiler without output threshold", + "original_footnote": "Modified read-out, not scSemiProfiler's official output. Main slides use the official one.", + "figures_before_palette": [ + "figures/secA_scsemi_readout.png" + ], + "source_page_131": 129, + "deck_group": "Supp S13", + "previous_section_label": "Appendix", + "parent_page": 122, + "parent_enriched_page": 131, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/secA_scsemi_readout.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 133, + "parent_154_page": 145, + "parent_155_page": 146 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_bulkonly_imperial_wave_shift", + "title": "Bulk-only exploration: calibration populations differ", + "section_label": "Supp. S14", + "figures": [ + "figures/expanded_master_20261013/disease/bulkonly_imperial_wave_shift.png" + ], + "keywords": [], + "footnote": "Historical Imperial sample composition. Paired and bulk-only sets differ in wave; repeated samples are not independent donors.", + "footnote_full": "Counts are read from the original synthesis table, including the recorded mixed w1|w2 category rather than silently reallocating it. This describes a calibration-transfer limitation and is not a causal wave effect or novel biological discovery. Figure 6 remains awaiting the user's new reproduction dataset.\nSources: plans/bulkonly_lora5_20261006/analysis/wf/synthesis_imperial_eskd/i00_samples_by_set_group_wave.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_imperial_wave_shift.csv; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md\nEvidence source: plans/bulkonly_lora5_20261006/analysis/wf/synthesis_imperial_eskd/i00_samples_by_set_group_wave.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_imperial_wave_shift.csv; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md", + "sources": "plans/bulkonly_lora5_20261006/analysis/wf/synthesis_imperial_eskd/i00_samples_by_set_group_wave.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_imperial_wave_shift.csv; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md", + "notes": "Counts are read from the original synthesis table, including the recorded mixed w1|w2 category rather than silently reallocating it. This describes a calibration-transfer limitation and is not a causal wave effect or novel biological discovery. Figure 6 remains awaiting the user's new reproduction dataset.\nSources: plans/bulkonly_lora5_20261006/analysis/wf/synthesis_imperial_eskd/i00_samples_by_set_group_wave.csv; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_imperial_wave_shift.csv; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md", + "after_original": 55, + "palette": "biological", + "new_experiment": false, + "addition_key": "disease_bulkonly_imperial_wave_shift", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/bulkonly_imperial_wave_shift.png" + ], + "source_page_131": 106, + "deck_group": "Supp S14", + "section": "Supp S14", + "previous_section_label": "Historical exploration", + "parent_page": 123, + "parent_enriched_page": 132, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/bulkonly_imperial_wave_shift.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 134, + "parent_154_page": 146, + "parent_155_page": 147 + }, + { + "kind": "divider", + "title": "Historical scoring and plans", + "original_slide": 73, + "original_title": "Appendix", + "original_footnote": "", + "source_page_131": 128, + "deck_group": "Historical archive", + "section": "Historical archive", + "previous_section_label": null, + "section_label": "Historical archive", + "parent_page": 124, + "parent_enriched_page": 133, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 135, + "parent_154_page": 147, + "parent_155_page": 148 + }, + { + "kind": "content", + "section": "Historical archive", + "section_label": "Historical only", + "text_source": { + "file": "data/fig5_rare_states_slide_text.json", + "index": 0 + }, + "title": "Rare-state fractions: original analysis", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig5_rare_fraction.png" + ], + "keywords": [ + "Original output retained for traceability", + "Use the corrected donor-level analysis for updated conclusions" + ], + "footnote": "Historical scorer; excluded from current main conclusions. Zero-draw and sample-unit limitations apply.", + "sources": "tools/fig5_rare_states.py; data/fig5_rare_states.csv; plans/rare_states_paired_20261008/out/", + "refs": [], + "refs_full": [], + "say": "Across four disease cohorts and 56 rare-state comparisons, LoRA-5 keeps rare states much closer to their observed fractions than scSemiProfiler with the same five donors, which often loses a state entirely.", + "footnote_full": "Historical scoring: zero-draw and sample-unit limitations apply. Corrected five-donor analyses follow; five-fold correction is separate.\n\nOriginal slide 49: Main comparison: 5 donors per draw (LoRA-5 vs scSemiProfiler with the same 5 donors as representatives, and zero-shot), every other usable sample tested, mean over 5 draws of log2(generated % / observed %) of the state within its parent type. Left: one point per state x cohort; below the diagonal = LoRA-5 closer; states a method never generates are drawn as 'not generated' (scSemiProfiler-5: 6 of 56; LoRA-5: 0). Zero-shot (no pairs) is closer than LoRA-5 on 28 of 56 fractions, LoRA-5 ranks donors better on 35 of 56. State \u00d7 cohort pairs with \u2265 30 observed state cells in \u2265 3 donors (native depth); negative control left out. Marker rules from the literature (plans/rare_states_screen_20261004/tools/states.json); observed parent cells thinned to the generated depth; labels = the benchmark's frozen scANVI calls; the benchmark's Poisson read-out reproduced cell by cell. Source: data/fig5_rare_states.csv (plans/rare_states_paired_20261008/out/summary.csv).", + "numbers": [ + { + "value": "{\"5 donors\": {\"n\": 56, \"frac_closer\": 51, \"within2_lora\": 19, \"within2_sc\": 5, \"not_generated_lora\": 0, \"not_generated_sc\": 6, \"rho_better\": 43, \"rho_median_lora\": 0.2381686657463106, \"rho_median_sc\": 0.17609850402537303, \"dir_lora\": 0.7321428571428571, \"dir_sc\": 0.6071428571428571}, \"5-fold\": {\"n\": 56, \"frac_closer\": 47, \"within2_lora\": 25, \"within2_sc\": 9, \"not_generated_lora\": 0, \"not_generated_sc\": 5, \"rho_better\": 38, \"rho_median_lora\": 0.31696815102004133, \"rho_median_sc\": 0.2431928042007824, \"dir_lora\": 0.8035714285714286, \"dir_sc\": 0.6785714285714286}}", + "source": "data/fig5_rare_states.csv" + } + ], + "original_slide": 49, + "original_title": "Rare states with 5 pairs: how many cells", + "original_footnote": "4 paired disease cohorts; 5 donors, mean of 5 draws. 0 = observed fraction; band = within 2-fold.", + "figures_before_palette": [ + "figures/fig5_rare_fraction.png" + ], + "source_page_131": 88, + "deck_group": "Historical archive", + "previous_section_label": "Historical scorer", + "parent_page": 125, + "parent_enriched_page": 134, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig5_rare_fraction.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 136, + "parent_154_page": 148, + "parent_155_page": 149 + }, + { + "kind": "content", + "section": "Historical archive", + "section_label": "Historical only", + "text_source": { + "file": "data/fig5_rare_states_slide_text.json", + "index": 1 + }, + "title": "Rare-state ranking: original analysis", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/fig5_rare_donors.png" + ], + "keywords": [ + "Original output retained for traceability", + "Use the corrected donor-level analysis for updated conclusions" + ], + "footnote": "Historical scorer; excluded from current main conclusions. Zero-draw and sample-unit limitations apply.", + "sources": "tools/fig5_rare_states.py; data/fig5_rare_states.csv; plans/rare_states_paired_20261008/out/", + "refs": [], + "refs_full": [], + "say": "LoRA-5 also puts the rare cells in the right donors more often, and keeps the disease direction in about three of four comparisons.", + "footnote_full": "Historical scoring: zero-draw and sample-unit limitations apply. Corrected five-donor analyses follow; five-fold correction is separate.\n\nOriginal slide 50: Main comparison (5 donors, 5 draws). Left: Spearman over the tested donors of the generated vs observed (thinned) state percentage, mean over draws. Right: share of draws where the case-vs-control difference of donor means has the observed sign (sepsis vs controls, COVID+ vs COVID-, severe vs mild, ChILI vs controls). State \u00d7 cohort pairs with \u2265 30 observed state cells in \u2265 3 donors (native depth); negative control left out. Marker rules from the literature (plans/rare_states_screen_20261004/tools/states.json); observed parent cells thinned to the generated depth; labels = the benchmark's frozen scANVI calls; the benchmark's Poisson read-out reproduced cell by cell. Source: data/fig5_rare_states.csv (out/summary.csv, out/contrasts.csv).", + "numbers": [ + { + "value": "{\"5 donors\": {\"n\": 56, \"frac_closer\": 51, \"within2_lora\": 19, \"within2_sc\": 5, \"not_generated_lora\": 0, \"not_generated_sc\": 6, \"rho_better\": 43, \"rho_median_lora\": 0.2381686657463106, \"rho_median_sc\": 0.17609850402537303, \"dir_lora\": 0.7321428571428571, \"dir_sc\": 0.6071428571428571}, \"5-fold\": {\"n\": 56, \"frac_closer\": 47, \"within2_lora\": 25, \"within2_sc\": 9, \"not_generated_lora\": 0, \"not_generated_sc\": 5, \"rho_better\": 38, \"rho_median_lora\": 0.31696815102004133, \"rho_median_sc\": 0.2431928042007824, \"dir_lora\": 0.8035714285714286, \"dir_sc\": 0.6785714285714286}}", + "source": "data/fig5_rare_states.csv" + } + ], + "original_slide": 50, + "original_title": "Rare states with 5 pairs: in the right donors", + "original_footnote": "Spearman over donors of generated vs observed state %; direction = case-vs-control sign.", + "figures_before_palette": [ + "figures/fig5_rare_donors.png" + ], + "source_page_131": 90, + "deck_group": "Historical archive", + "previous_section_label": "Historical scorer", + "parent_page": 126, + "parent_enriched_page": 135, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/legacy_rare/fig5_rare_donors.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 137, + "parent_154_page": 149, + "parent_155_page": 150 + }, + { + "kind": "content", + "section": "Historical archive", + "section_label": "Historical only", + "text_source": { + "file": "data/fig5_rare_states_slide_text.json", + "index": 2 + }, + "title": "Rare states: original five-fold analysis", + "layout": "full", + "figures": [ + "figures/expanded_master_20261013/palette/fig5_rare_fivefold.png" + ], + "keywords": [ + "Original output retained for traceability", + "Use the corrected donor-level analysis for updated conclusions" + ], + "footnote": "Historical scorer; excluded from current main conclusions. Zero-draw and sample-unit limitations apply.", + "sources": "tools/fig5_rare_states.py; data/fig5_rare_states.csv; plans/rare_states_paired_20261008/out/", + "refs": [], + "refs_full": [], + "say": "With all training pairs the picture is the same; this five-fold result is reported separately.", + "footnote_full": "Historical scoring: zero-draw and sample-unit limitations apply. Corrected five-donor analyses follow; five-fold correction is separate.\n\nOriginal slide 51: Reported separately: 5-fold CV, every sample generated by the fold model that never trained on it; scSemiProfiler with exactly LoRA's training pairs as representatives. State \u00d7 cohort pairs with \u2265 30 observed state cells in \u2265 3 donors (native depth); negative control left out. Marker rules from the literature (plans/rare_states_screen_20261004/tools/states.json); observed parent cells thinned to the generated depth; labels = the benchmark's frozen scANVI calls; the benchmark's Poisson read-out reproduced cell by cell. Source: data/fig5_rare_states.csv.", + "numbers": [ + { + "value": "{\"5 donors\": {\"n\": 56, \"frac_closer\": 51, \"within2_lora\": 19, \"within2_sc\": 5, \"not_generated_lora\": 0, \"not_generated_sc\": 6, \"rho_better\": 43, \"rho_median_lora\": 0.2381686657463106, \"rho_median_sc\": 0.17609850402537303, \"dir_lora\": 0.7321428571428571, \"dir_sc\": 0.6071428571428571}, \"5-fold\": {\"n\": 56, \"frac_closer\": 47, \"within2_lora\": 25, \"within2_sc\": 9, \"not_generated_lora\": 0, \"not_generated_sc\": 5, \"rho_better\": 38, \"rho_median_lora\": 0.31696815102004133, \"rho_median_sc\": 0.2431928042007824, \"dir_lora\": 0.8035714285714286, \"dir_sc\": 0.6785714285714286}}", + "source": "data/fig5_rare_states.csv" + } + ], + "original_slide": 51, + "original_title": "5-fold CV: rare states", + "original_footnote": "Donor-level 5-fold CV, pooled; reported separately from the 5-pair comparison.", + "figures_before_palette": [ + "figures/fig5_rare_fivefold.png" + ], + "source_page_131": 94, + "deck_group": "Historical archive", + "previous_section_label": "Historical scorer", + "parent_page": 127, + "parent_enriched_page": 136, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/fig5_rare_fivefold.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 138, + "parent_154_page": 150, + "parent_155_page": 151 + }, + { + "kind": "content", + "section": "Historical archive", + "section_label": "Historical archive", + "title": "Bulk-only cohort plan: original snapshot", + "layout": "figure_keywords", + "figures": [ + "figures/sec10_paired_subset_cohorts.png" + ], + "keywords": [ + "Original cohort/sample inventory", + "Later exploratory outputs follow", + "New-data reproduction is a separate step" + ], + "footnote": "Historical cohort counts. Existing exploratory outputs are reviewed separately; reproduction with the user-prepared new dataset remains pending.", + "say": "Historical planned cohort map. Later exploratory outputs exist; this slide does not claim new-data reproduction is complete.", + "sources": "tools/sec10_paired_subset_cohorts.py; data/sec10_cohort_counts.csv, data/sec10_bulk_only_groups.csv, data/sec10_quoted_numbers.csv", + "footnote_full": "Original slide 55: Counts from pairing manifests. QC = identity check (Imperial also >=200 cells); sex-flag = bulk sex differs from metadata or is ambiguous. Not shown: Asan (no donor map), Giroux (24 pairs, 17 bulk-only). Model: f01 6k 1x E16 + LoRA r16.", + "original_slide": 55, + "original_title": "Next: cohorts with a paired subset (planned)", + "original_footnote": "Counts from pairing manifests. Model: f01 6k 1x E16 + LoRA r16.", + "figures_before_palette": [ + "figures/sec10_paired_subset_cohorts.png" + ], + "source_page_131": 104, + "deck_group": "Historical archive", + "previous_section_label": "Historical plan", + "parent_page": 128, + "parent_enriched_page": 137, + "figures_before_final_palette": [ + "figures/sec10_paired_subset_cohorts.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 139, + "parent_154_page": 151, + "parent_155_page": 152 + }, + { + "kind": "content", + "layout": "full", + "key": "disease_bulkonly_verification_history", + "title": "Historical bulk-only search: no confirmed new finding", + "section_label": "Historical archive", + "figures": [ + "figures/expanded_master_20261013/disease/bulkonly_verification_history.png" + ], + "keywords": [], + "footnote": "Historical exploratory analysis of existing bulk-only generations. Review dispositions are not experimental validation. Figure 6 reproduction still awaits new data.", + "footnote_full": "The original K1000 workflow considered 18 Sepsis and 21 Imperial candidate findings. Both reports retain zero confirmed findings under their internal two-verifier grading rule. This is an account of prior exploration, not newly completed biology. Known weaknesses include shared generation noise across draws, calibration transfer, clinical confounding and some invalid verifier p-values. No novel cell population survived verification; the user-prepared Figure 6 reproduction is still pending.\nSources: plans/bulkonly_lora5_20261006/analysis/wf/findings_ye2026_sepsis.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_ye2026_sepsis.md; plans/bulkonly_lora5_20261006/analysis/wf/findings_imperial_eskd.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_verification_history.csv; plans/bulkonly_lora5_20261006/README.md\nEvidence source: plans/bulkonly_lora5_20261006/analysis/wf/findings_ye2026_sepsis.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_ye2026_sepsis.md; plans/bulkonly_lora5_20261006/analysis/wf/findings_imperial_eskd.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_verification_history.csv; plans/bulkonly_lora5_20261006/README.md", + "sources": "plans/bulkonly_lora5_20261006/analysis/wf/findings_ye2026_sepsis.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_ye2026_sepsis.md; plans/bulkonly_lora5_20261006/analysis/wf/findings_imperial_eskd.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_verification_history.csv; plans/bulkonly_lora5_20261006/README.md", + "notes": "The original K1000 workflow considered 18 Sepsis and 21 Imperial candidate findings. Both reports retain zero confirmed findings under their internal two-verifier grading rule. This is an account of prior exploration, not newly completed biology. Known weaknesses include shared generation noise across draws, calibration transfer, clinical confounding and some invalid verifier p-values. No novel cell population survived verification; the user-prepared Figure 6 reproduction is still pending.\nSources: plans/bulkonly_lora5_20261006/analysis/wf/findings_ye2026_sepsis.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_ye2026_sepsis.md; plans/bulkonly_lora5_20261006/analysis/wf/findings_imperial_eskd.json; plans/bulkonly_lora5_20261006/analysis/wf/REPORT_imperial_eskd.md; presentations/2026-09-30--paper-summary--r00/data/expanded_master_20261013/disease/bulkonly_verification_history.csv; plans/bulkonly_lora5_20261006/README.md", + "after_original": 55, + "palette": "biological", + "new_experiment": false, + "addition_key": "disease_bulkonly_verification_history", + "figures_before_palette": [ + "figures/expanded_master_20261013/disease/bulkonly_verification_history.png" + ], + "source_page_131": 105, + "deck_group": "Historical archive", + "section": "Historical archive", + "previous_section_label": "Historical exploration", + "parent_page": 129, + "parent_enriched_page": 138, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/disease/bulkonly_verification_history.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 140, + "parent_154_page": 152, + "parent_155_page": 153 + }, + { + "kind": "divider", + "title": "Discussion and outlook", + "original_slide": 75, + "original_title": "Discussion", + "original_footnote": "", + "source_page_131": 130, + "deck_group": "Discussion", + "section": "Discussion", + "previous_section_label": null, + "section_label": "Discussion", + "parent_page": 130, + "parent_enriched_page": 139, + "figures_before_final_palette": [], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 141, + "parent_154_page": 153, + "parent_155_page": 154 + }, + { + "kind": "content", + "section": "Discussion", + "section_label": "Discussion", + "text_source": { + "file": "data/sec10_other_tissues_slide_text.json", + "index": 0 + }, + "title": "Next: other tissues from a few dozen pairs", + "layout": "full", + "figures": [ + "figures/final_palette_20261013/sec10_other_tissues.png" + ], + "keywords": [ + "Pretrained: a 3-min LoRA works from 5 pairs", + "Scratch: 58 pairs + recipes come close (86 min)", + "Next: test this route on other tissues" + ], + "footnote": "Pilots: sepsis fold 0 (17 donors); last row held-out sepsis (82). Folds 1\u20134 and a second cohort pending.", + "say": "Pretraining is what makes PBISC practical: a LoRA of a few minutes on as few as five pairs. In a one-fold sepsis pilot, though, with enough compute, a few dozen paired bulk and single-cell samples plus pseudobulk recipes came close to the same results without corpus pretraining, and the outcome depends on the training protocol. This does not weaken the pretraining result. It suggests that a PBISC-like bulk-to-single-cell model could be built for other organs and tissues that have no large single-cell corpus yet. Folds 1 to 4 and a second cohort will test this first.", + "sources": "tools/sec10_other_tissues.py; data/sec10_other_tissues.csv (from sec07_no_pretraining.csv, sec08_cross_scratch.csv, secS_scratch_rare_states_bootstrap.csv, sec06_ft_method.csv, sec01_model_numbers.csv train_split_cells / corpus_cells); plans/paired_finetune_intradataset_20260927/runs/ye2026_sepsis/fold0/*/RUN.json; plans/paired_crosscohort_20261002/generated_masked/ye2026_sepsis/hoL_ye2026_sepsis/locoL_scratchsat_*.log", + "footnote_full": "Original slide 76: Discussion point; the from-scratch evidence is in the Supplementary. Same cohort, same pairs: sepsis fold 0, 17 held-out donors; every run also early-stops on the same 7 inner-validation pairs. From scratch = the same denoiser at random initialisation, full training (lr 2e-4, inner-val selection, cap 100k steps; the 58-pair and 58-pair + recipes runs were still improving at the cap). It still took the architecture, the diffusion settings, the 6k gene panel, the value scaler and the bulk z-score constants from the PBMC checkpoint, so a corpus-free setup (panel and scalers built from the pairs alone) is untested. Recipes = 50% of each batch is pseudobulk built from the training donors' own cells, K from the pretraining grid (250\u20134,000). Bold = higher by 0.05 or more (one fold, no interval), or less training time per cohort; the pretraining row is not scored. Rare states: mean |log2 gen/obs| over 13 states, scratch minus pretrained -0.09 [-0.38, +0.18] (donor bootstrap, n.s.); per-state calls = significant, marker states only, excluding states that both arms miss by more than 4-fold. Protocol dependence: recipes, the step budget, and the gene space - a scratch model learns only genes measured in its training pairs (trained on other cohorts, it put 91\u201394% of its UMI on the 46 genes that none of those cohorts measured). Pairs from other cohorts do not replace the target cohort's own: own-cohort LoRA coverage 0.83 on held-out sepsis. Pilot only: folds 1\u20134 and a second cohort are needed before the claim. Pretraining (15.9 h on 33.1 M train cells of the 41.4 M-cell corpus) is done once and reused by every cohort.", + "numbers": [ + { + "value": "Corpus pretraining: pretrained 33.1 M train cells, 15.9 h, once / scratch none", + "source": "data/sec01_model_numbers.csv train_split_cells; sec06_ft_method.csv pretrain_wall_hours" + }, + { + "value": "Training, sepsis fold 0 (one GPU): pretrained 3\u20135 min / scratch 61\u201386 min", + "source": "data/sec07_no_pretraining.csv minutes" + }, + { + "value": "5 pairs: donor composition r: pretrained 0.70 / scratch 0.17", + "source": "data/sec07_no_pretraining.csv comp_r" + }, + { + "value": "58 pairs + recipes: donor composition r: pretrained 0.80 / scratch 0.77", + "source": "data/sec07_no_pretraining.csv comp_r" + }, + { + "value": "58 pairs + recipes: coverage: pretrained 0.87 / scratch 0.88", + "source": "data/sec07_no_pretraining.csv coverage" + }, + { + "value": "58 pairs + recipes: disease logFC r: pretrained 0.45 / scratch 0.54", + "source": "data/sec07_no_pretraining.csv logfc_r" + }, + { + "value": "58 pairs + recipes: 13 rare states, n.s. on average: pretrained closer on ISG-high T / scratch closer on DN2 B", + "source": "data/secS_scratch_rare_states_bootstrap.csv" + }, + { + "value": "Other cohorts' pairs only (sepsis held out): coverage: pretrained 0.34 (LoRA r16) / scratch 0.33", + "source": "data/sec08_cross_scratch.csv coverage (masked)" + } + ], + "original_slide": 76, + "original_title": "Next: other tissues from a few dozen pairs", + "original_footnote": "Pilots: sepsis fold 0 (17 donors); last row held-out sepsis (82). Folds 1\u20134 and a second cohort pending.", + "figures_before_palette": [ + "figures/sec10_other_tissues.png" + ], + "source_page_131": 131, + "deck_group": "Discussion", + "previous_section_label": "Discussion", + "parent_page": 131, + "parent_enriched_page": 140, + "figures_before_final_palette": [ + "figures/expanded_master_20261013/palette/sec10_other_tissues.png" + ], + "method_palette": "PBISC-Diffusion_color_palette_20261013.md", + "parent_142_page": 142, + "parent_154_page": 154, + "parent_155_page": 155 + } + ] +} diff --git a/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure_assignment_20261013/STRENGTHENING_PROPOSAL.md b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure_assignment_20261013/STRENGTHENING_PROPOSAL.md new file mode 100644 index 0000000000000000000000000000000000000000..57fed12bbf55942f5a3b247d33e814d790c57b63 --- /dev/null +++ b/paper-releases/pbisc-diffusion/updates/20261015T100456Z-figure1-diffusion-readable-corpus/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure_assignment_20261013/STRENGTHENING_PROPOSAL.md @@ -0,0 +1,158 @@ +# Figure 1–6 / Supplementary: 근거 보강 계획과 반영 결과 + +2026-10-13 사용자 수정 지시를 반영했다. Figure 번호는 [AUTHOR_SCOPE.md](AUTHOR_SCOPE.md)를 따른다. 현재 덱의 쪽수는 [현재 페이지 대응표](../figure1_focus_20261015/page_map.csv), 변경·검수 결과는 [작업 보고서](../strengthening_20261013/README.md)를 기준으로 한다. 이전 논의용 제안서는 [수정 전 사본](../strengthening_20261013/STRENGTHENING_PROPOSAL_before_revision.md)에 보존했다. + +2026-10-15 Figure 1 보완: 1a에 noisy gene state → iterative denoising → sparse expression을 명시했고, 1b 주요 비율·범례를 확대했다. 작은 범주는 명시적으로 묶고 전체 수치/범주는 support 9쪽에 보존했다. [변경 보고서](../figure1_focus_20261015/README.md). + +2026-10-14 Figure 1 수정: main 1a는 참조 그림체의 벡터 도식으로 교체했고, main 1b는 실제 train 33,141,380개 기준 도넛으로 추가했다. Healthy는 train 기준 88.4241%이며, 아래 full-source 88.51%와 분모가 다르다. [Figure 1 변경 보고서](../figure1_revision_20261014/README.md). + +2026-10-13 작업은 저장된 결과 재시각화, 분할·원 논문 검증, observed marker의 CPU 분석이다. 학습·생성·deconvolution을 새로 실행하지 않았다. 기존 campaign의 미완료 단위와 새로운 분석을 구분한다. 과거 13:58 KST 상태를 현재 실행 상태로 인용하지 않는다. + +## 우선 결정 + +| 구분 | 이번 반영 | 남은 핵심 작업 | 새 모델 학습 필요 | +|---|---|---|---| +| Figure 1 | Healthy / Disease·at-risk → 상세 condition, cell type; corpus split와 K 모식도 | 사용 경계·donor 교차표의 최종 캡션 확인 | 없음 | +| Figure 2 | 기존 architecture/hint 도식, validation progression, 비용·parameter 비교 유지 | 동일 family 안에서 주 패널 선정; 미완료 run의 saturation 과장 금지 | 기본적으로 없음 | +| Figure 3 | 같은 107명 observed/raw/calibration/full FT/LoRA 직접 비교; 실제 split 도식 | 기존 strict total-support campaign 완료; 공통 cell-space 정성 패널은 저장 출력에서 후속 조립 | 중복 실행 없음 | +| Figure 4 | 누락 scSemi CV CAT/RBO 복원, 다중 지표 비교, strict Sepsis donor-aware DEG | 나머지 cohort의 기존 strict 분석 완료, 동일 donor를 보존하는 CI | 중복 실행 없음 | +| Figure 5 | 원 논문 target 대조; Ye paired observed marker 진단 | 원 저자 cluster identity 연결 후 observed/PBISC/scSemi 비교 | 먼저 CPU 분석 | +| Figure 6 | 기존 논문 결과 재현 계획 유지 | endpoint·cohort 고정 후 기존 bulk-only 출력 분석 | 기본적으로 없음 | + +지금 가장 먼저 보강할 일은 Figure 5의 정확한 population identity, Figure 4의 donor-aware 비교 완성이다. 그림 수나 유리한 지표의 수를 늘리는 것을 종료 기준으로 삼지 않는다. + +## Figure 1 — condition 비율과 corpus → recipe의 관계 + +### 완료: 두 단계 condition 도표 + +총 **41,395,184개 source cells**를 기준으로 Healthy **36,639,276개, 88.51%**, Disease / at-risk **4,755,908개, 11.49%**를 나눈다. 가운데 도넛에서 후자를 11개 원래 condition으로 세분하고 cell-type 도넛을 함께 둔다. + +- ACPA+ at-risk는 **1,792,129개**로, 확진 RA가 아니다. 이 때문에 상위 범주를 단순 Disease로 쓰지 않는다. +- At-risk를 제외한 disease/history annotation은 **2,963,779개, 전체의 7.16%**다. Post-COVID와 Cytopenia도 원래 annotation 의미를 유지한다. +- Healthy는 donor 상태다. In-vitro 자극, post-vaccination, CMV-seropositive healthy를 모두 resting untreated cell로 해석하지 않는다. +- **88.5%는 donor 비율이 아니라 cell 비율**이다. Condition 라벨은 corpus 설명용이며 최종 base model의 입력 label이 아니다. + +### 완료: 41.4M와 33.1M 차이는 split + +| Source allocation | Cells | 전체 중 비율 | +|---|---:|---:| +| Train | 33,141,380 | 80.06% | +| Validation | 4,361,032 | 10.54% | +| Test | 3,892,772 | 9.40% | + +따라서 약 **33.4M가 아니라 33.1M**이며, 차이 8,253,804개는 validation+test다. K recipe 때문에 제외된 세포 수가 아니다. 실제 group 크기가 달라 nominal 80/10/10이 cell 비율로 정확히 일치하지 않는다. + +### 완료: K와 target exposure 모식도 + +`train sample에서 서로 다른 K cells 선택 → counts 합으로 pseudobulk condition 생성 → 같은 condition으로 각 member cell을 target으로 학습`한다. Registered K는 250/500/1,000/2,000/4,000, schedule part당 8,192 recipes이며 16 parts를 사용한다. + +Regular grid로 31,145,450개 고유 train cells를 덮은 뒤, **3,658개 top-up recipes**로 남은 1,995,930개를 포함한다. Final contract는 **33,141,380개 train cells 전부의 coverage**를 확인한다. 총 target occurrences 195,639,930회는 반복 노출을 포함하므로 고유 세포 수와 다르다. Top-up의 실제 K는 1–4,000이다. + +`unique source cells`, `recipe 개수`, `K`, `target occurrences`, `N_generated`를 서로 바꿔 쓰지 않는다. Source condition 비율을 sampler exposure 비율로 해석하지 않는다. + +**보충자료:** 내부 corpus split에는 sample 중복은 없지만 biological donor 27명이 split을 가로지른 기록이 있다. 외부 SLE/paired test 누출을 입증하는 결과와 구분한다. AIFI reference의 pretraining 노출도 target donor 중복과 별개로 기록한다. + +**재사용/산출물:** [corpus bundle](../strengthening_20261013/corpus/result.json), `manuscript/source_data/corpus/`, final run `ALLCELL_DATA_CONTRACT.json`. 새 학습 없음. + +## Figure 2 — architecture / gene context에 집중 + +기존에 추가한 context/trunk, pred/oracle hint 경로 도식과 full/late validation curves, parameter/training-time 패널을 유지한다. 이 Figure에 real-bulk mismatch·paired adaptation을 넣지 않는다. + +**주 그림 조합:** 간략한 모델 구조 → width/context family의 validation progression → SLE pseudobulk 성능 → gene-panel family → parameter/time trade-off. Saved corpus-validation loss는 SLE held-out benchmark metric과 다른 자료임을 표시한다. + +**실제로 남는 보강:** 같은 평가 조건의 arm만 비교하고, 제한된 학습 기록을 saturation으로 부르지 않는다. 16× run은 E2까지만 있어 수렴 여부가 미확인이다. Architecture 1% detection floor와 gene-panel 0.1% floor의 차이도 보존한다. 6k의 5 seeds는 sampling 반복이며 독립 pretraining 5회가 아니다. + +**추가 실험 조건:** 특정 architecture의 안정적 우위를 논문의 필수 주장으로 삼는 경우에만 가장 가까운 경쟁 checkpoint의 동일 sampling 반복을 검토한다. 현재 시각화 보강 때문에 architecture 전체를 재학습할 필요는 없다. + +## Figure 3 — 입력 mismatch와 생성 결과 회복을 연결 + +### 완료: 같은 사람을 나란히 보여주는 직접 비교 + +Observed paired scRNA의 cell-type mixture와 raw zero-shot / input calibration / full FT / LoRA r16을 **동일 107명, 동일 순서**로 나란히 그렸다. Donor 순서는 observed monocyte 비율만으로 정했고, 모든 donor와 6개 type을 남겼다. Full FT/LoRA와 raw zero-shot은 같은 raw measured-bulk 입력을 사용한다. + +같은 사람을 잇는 보조 패널은 composition L1과 **generated-vs-observed pseudobulk gene-expression Pearson r**를 함께 보여준다. 후자는 같은 5,508 genes의 log1p-CPM에 대한 값이며 cell별 일대일 일치도가 아니다. + +- Mean composition L1: raw **0.4549**, calibration **0.1985**, full FT **0.1572**, LoRA r16 **0.1617**. +- Median expression r: raw **0.8921**, calibration **0.8802**, full FT **0.9141**, LoRA r16 **0.8994**. + +즉 output composition 회복은 분명하지만 모든 지표에서 LoRA가 가장 좋거나 모든 expression mismatch가 사라진다고 쓰지 않는다. Adaptation은 measured-bulk 값을 직접 바꾸는 연산이 아니다. 기존 input-calibration density/error 패널과 생성 output 패널의 역할을 구분한다. + +### 확인: LoRA all / LoRA20의 실제 test + +기존 pair-budget 그림은 **고정된 5-fold outer test**를 사용한다. + +| 표기 | Training | Validation | Test/fold | 사용하지 않는 나머지 | +|---|---:|---:|---:|---:| +| LoRA5 | 5 | 7 | 21–22 | 73–74 | +| LoRA20 | 20 | 7 | 21–22 | 58–59 | +| LoRA all | 78–79 | 7 | 21–22 | 0 | + +“All”은 **각 fold에서 사용할 수 있는 training pool 전체**다. 107명 전체를 학습해 같은 사람에게 test하지 않는다. 각 donor의 held-out prediction을 모아 pooled 107명 결과를 만든다. LoRA20의 나머지 87명이 모두 test가 되는 설계도 아니다. Test를 고정해야 support 양에 따른 차이를 같은 사람들에서 비교할 수 있다. + +Outer/inner 분할 seed는 **20260927**이다. Training subset은 fold별 **1000–1004** seed의 random permutation에서 prefix를 취한다. 따라서 `5 ⊂ 10 ⊂ 20 ⊂ 40 ⊂ all`인 **random nested selection**이다. 저장된 20개 subset과 outer/inner 분할을 replay해 정확히 일치함을 확인했다. + +**프로토콜 구분:** 기존 별도 total-five 5 draws는 5 support+0 validation, 나머지 102명 test/draw다. 새 strict healthy campaign은 5/10/20/40 total support와 고정 outer-test fold, 추가 validation 없음이다. 두 total-five 설계도 같은 실험으로 합치지 않는다. + +**남은 작업:** 기존 strict curve의 미완료 산출물을 마무리한다. 추가 정성 패널이 필요하면 저장된 observed/raw/LoRA cells를 observed-only fit의 동일 projection에 배치하고 동일 test·depth·N_gen을 유지한다. 이는 CPU figure 작업을 우선하며 새로운 학습 근거가 아니다. r4/r16/r64/full FT는 지표별 절충으로 보고 r16 보편 우위를 만들기 위한 합성점수는 사용하지 않는다. + +**근거:** [healthy bundle](../strengthening_20261013/healthy/result.json), [split audit](../strengthening_20261013/healthy/split_audit.csv). + +## Figure 4 — reference가 있는 여러 평가를 보강 + +### 완료: 기존 결과에서 추가한 세 패널 + +1. **Complete historical CV-CAT:** 누락됐던 네 cohort의 scSemi CV 결과를 포함한다. Raw/calibration/LoRA/scSemi/bulk-broadcast, 저장 결과가 있는 Sepsis full FT를 같은 조건에서 비교한다. +2. **LogFC r / partial r / RBO:** observed paired scRNA를 reference로 사용한다. Same-patient cell-repeat r는 내부 cell resampling control이며 독립 환자 generalization ceiling이 아니다. +3. **Strict five-support Sepsis:** 완료된 5 draws의 logFC r, partial r, CAT100, donor-aware DEG overlap을 함께 제시한다. 점은 overlapping support draws, 큰 점은 descriptive mean이다. Draw를 독립 환자로 간주한 CI를 붙이지 않는다. + +CAT는 [Irizarry et al., 2005](https://doi.org/10.1038/nmeth756), RBO는 [Webber et al., 2010](https://doi.org/10.1145/1852102.1852106)를 근거로 한다. Same-sign CAT는 현재 benchmark에서 방향 일치를 추가한 정의다. Donor-aware DEG의 필요성은 [Squair et al., 2021](https://www.nature.com/articles/s41467-021-25960-2)에 근거를 둔다. 구체적 구현·유전자 축·reference·출처는 [metric bundle](../strengthening_20261013/disease_metrics/result.json)에 묶었다. + +**주장 범위:** Strict Sepsis에서 LoRA/scSemi의 평균 logFC r는 **0.3421/0.4025**, CAT100은 **0.0768/0.0328**, donor-aware DEG overlap은 **17.6/11.4**다. 순위·DEG 회복에서의 장점과 전체 logFC correlation의 반례가 동시에 있다. ‘모든 질환·모든 지표에서 최고’라는 결론은 근거와 맞지 않는다. + +### 남은 필수 보강 + +- 나머지 질환의 기존 strict outputs와 patient-aware statistics를 끝까지 연결한다. Sepsis 기존 deconvolution adoption을 다시 실행하지 않는다. +- 같은 donor가 여러 visit/draw에 나타나는 구조를 보존해 **방법 간 차이의 donor CI**를 계산한다. 고정된 생성 결과의 CI는 재학습 불확실성까지 포함하지 않는다. +- Primary endpoint와 보조 endpoint를 고정한다. LogFC r, partial r, CAT/RBO, DEG가 서로 다른 질문에 답한다는 점을 설명한다. +- Cohort/cell type별 NA, 불리한 결과, support/HVG/test 접근 범위를 Supplementary에 보존한다. Giroux의 sex/severity confounding은 sex gene 제거만으로 없어지지 않는다. + +새 지표를 계속 찾거나 원하는 순위가 나올 때까지 실험을 반복하기보다, 이미 있는 reference-backed 지표들의 비교 조건과 불확실성을 완성한다. + +## Figure 5 — 각 paired 원 논문의 population을 먼저 검증 + +원 논문에서 직접 보고한 population과 일반적인 marker rule을 구분한다. 네 cohort에 억지로 rare population 하나씩을 배정하지 않는다. 상세 원문/figure/marker/시점/파일 가용성은 [문헌·observed 분석 bundle](../strengthening_20261013/rare_literature/result.json)을 따른다. + +| Cohort | 원 논문 기반 우선 대상 | 다음 단계 | +|---|---|---| +| Ye sepsis | Fig.5 **Mono_c06-CD14-MKI67**, pediatric pulmonary sepsis; MKI67/STMN1/MPO/ELANE | 원 저자의 fine-cluster identity와 감염 부위·연령 대조를 연결 | +| Imperial ESKD | Fig.5 **dexamethasone-associated monocytes**, wave2 | Steroid 투약 시점·wave2·세부 cluster를 매칭; wave1 음성 대조 | +| Giroux COVID | IgG 전환 전 CD14 monocyte clusters 2/4(mild), 7(moderate), RNA/regulon states | Fine labels·serostatus 연결; IRF1/KLF3는 저장된 6k panel 밖이므로 exact rare identity 주장 보류 | +| Ji CHILI | Cell-type DE/pathway 중심 | 기존 DE/pathway endpoint는 Figure4에 적합; rare target 근거가 있어야 Figure5에 채택 | + +### 이번 observed-truth 진단에서 배운 점 + +Ye의 identity-passed **82명, frozen AIFI monocyte 111,269개**에서 CD14/MKI67/STMN1/MPO/ELANE만 CPU로 추출했다. MKI67+STMN1 native-depth 동시검출의 donor 평균은 소아 폐렴군 **2.33%**, 소아 복부감염군 **4.10%**, 성인 폐렴군 **0.39%**다. + +이 결과는 cycling-marker 신호가 실제 paired sc에 있다는 진단이지 **원 논문의 Mono_c06 정체성 또는 폐렴 특이성 재현 완료가 아니다**. 검출률은 depth와 dropout에도 영향을 받는다. 73-marker cache의 일반 rule로 원 논문의 cluster를 대체하면 잘못된 결론을 만들 수 있다. + +### 실행 순서와 완료 기준 + +1. **로컬 재사용 조사는 완료:** Ye의 canonical H5AD obs schema 133개, raw archive member 665개, benchmark metadata를 조사했으나 저자 fine-cluster labels를 찾지 못했다. [조사 목록](../strengthening_20261013/rare_literature/ye_author_label_inventory.json)을 보존했다. 다음은 원 논문의 공개 annotation/분석 코드에서 author_cell_id와 연결되는 라벨을 찾는 일이다. 확보할 수 없을 때만 원 논문 marker 기반 reference-only classifier/rule을 정의한다. +2. **Observed paired sc에서 먼저** 연령·감염 부위 또는 steroid timing에 따른 population identity/abundance를 재현한다. 이 단계가 안 맞으면 generated 결과를 성공 사례로 제시하지 않는다. +3. 같은 held-out samples의 observed/PBISC/scSemi를 동일 projection·측정 gene·depth/N_gen 조건에서 비교한다. Original fine labels로 확인한 parent population과 rare cluster를 분리한다. +4. 정의·projection에 사용하지 않은 marker/program으로 identity를 추가 검증한다. 모델 panel에 없는 marker를 zero로 채우지 않는다. +5. 성공·과소생성·관측 음성 사례와 donor별 전체 결과를 함께 둔다. Observed zero를 진정한 부재로 단정하지 않는다. + +기존 paired K1000/K4000은 모두 **1,500 generated cells/sample**이다. K sensitivity와 generated-cell count sensitivity를 혼동하지 않는다. 지금 필요한 작업은 저장 자료의 정확한 annotation·CPU 재분석이며, 새 GPU 학습이 기본 선행 조건은 아니다. + +## Figure 6 — bulk-only 논문 결과 재현은 별도 TODO + +추천 후보는 [Stephenson et al., 2025](https://pmc.ncbi.nlm.nih.gov/articles/PMC12366655/) Fig.3C의 monocyte HLA-II 감소다. Sample-time severity와 peak severity를 구분해 endpoint를 고정한 뒤, 기존 Imperial bulk-only 출력에서 donor-aware 효과를 계산한다. + +Paired donor를 모두 제외한 후보 집합은 기존 audit상 97 samples/45 donors, wave1이다. 이 집합으로 wave2 steroid-associated population을 재현할 수는 없다. Figure5의 **paired wave2** 가능성과 Figure6의 **donor-disjoint bulk-only wave1**을 혼동하지 않는다. HLA-II 방향은 기존 workflow에서도 살펴본 후향적 재현이며 blind independent validation이 아니다. 이번 작업에서 Figure6 분석은 실행하지 않았다. + +## Supplementary와 추가 GPU의 경계 + +Supplementary에는 전체 cohort/type/draw, 부정적 결과, support·validation·HVG 접근량, source/exposure 구분, rank/full FT 절충, 평가법·readout 감도, rare marker/annotation 근거와 음성 사례를 둔다. Main Figure2는 SLE pseudobulk, Figure3–5는 measured paired라는 범위를 유지한다. + +새 GPU 작업은 (a) 필수적인 architecture/training-seed 안정성 주장에 근거가 부족하거나, (b) adapted model의 conditioning 기전을 확인하는 matched-noise wrong-bulk intervention이 꼭 필요할 때에만 별도로 정의한다. 두 항목 모두 현재 기본 필수 과제는 아니다. 실행 전 reuse registry와 checkpoint/support/test/mask/seed/readout 계약을 대조하고 바뀐 연산만 수행한다.