Archive 4 checkpoints (1-4)

paper-releases/pbisc-diffusion/updates/20261008T060837Z-deck-new-figure-split-time/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/fig3_time_days.csv

@@ -0,0 +1,6 @@
+cohort,cv_scsemi_gpu_h,cv_lora_gpu_h,cv_scsemi_running_h,cv_scsemi_gpu_waits_excluded_h,draws_scsemi_gpu_h,draws_lora_gpu_h,cv_ratio
+immicro,23.73,1.225,27.22,20.66,11.72,0.6962,19.38
+ye2026_sepsis,38.47,0.618,40.76,11.43,20.32,0.5589,62.26
+imperial_eskd,30.89,0.5158,33.24,0,19.12,0.8999,59.89
+giroux2022,14.44,0.2961,15.1,38.47,12.97,0.2208,48.77
+ji2025_chili,7.397,0.3697,8.25,0,4.756,0.2829,20.01

paper-releases/pbisc-diffusion/updates/20261008T060837Z-deck-new-figure-split-time/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/fig3_time_days_slide_text.json

@@ -0,0 +1,33 @@
+{
+ "section": "fig3-time",
+ "scripts": [
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/tools/fig3_time_days.py"
+ ],
+ "generated_from": "tools/fig3_time_days.py (rerun to refresh; do not edit by hand)",
+ "slides": [
+  {
+   "title": "Time: days for scSemiProfiler, minutes for LoRA",
+   "layout": "full",
+   "figures": [
+    "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/presentations/2026-09-30--paper-summary--r00/figures/fig3_time_days.png"
+   ],
+   "keywords": [
+    "Healthy 5-fold: 1.0 GPU-day vs 1.2 h",
+    "Up to 1.6 days per cohort (Sepsis)",
+    "All runs: 7.7 days vs 22 h with pretraining"
+   ],
+   "footnote_short": "One GB10 GPU, job alone on it. GPU time from receipts; waits for the GPU excluded.",
+   "footnote": "GPU time on one GB10 GPU with every step alone on the GPU, from each run's receipts. One 5–fold experiment = the five fold models, every usable sample tested once by the model that never trained on it. scSemiProfiler = pretraining of every representative (two stages) + one model per test sample (upstream c2aea02, default settings); PBISC LoRA = fine–tune + sampling 1,500 cells per test sample. Ratio per cohort: 19–62x. scSemiProfiler running time including its CPU steps (waits for the GPU excluded): ImMicro (healthy) 27 h; Sepsis 41 h; Imperial 33 h; Giroux 15 h; Ji (whole blood) 8 h. All runs = 5–fold + the 25 five–donor draws: scSemiProfiler 184 h, LoRA 5.7 h; PBISC's pretraining (15.9 h, 33.1 M cells) is done once and reused by every cohort. Not counted on either side: data preparation and scoring. The equal–pair–budget scSemiProfiler runs are not included (they reuse the 5–fold run's representative models). Sources: data/sec09_same_pairs.csv, data/sec09_time_to_apply.csv, data/sec06_ft_method.csv, scsemi/work/<cohort>/cv/fold<k>/events.jsonl.",
+   "numbers": [
+    {
+     "value": "ImMicro (healthy): 5-fold scSemiProfiler 23.7 GPU h (0.99 days, running 27.2 h) vs LoRA 1.22 h (19x); 5-donor draws 11.7 h vs 0.70 h; Sepsis: 5-fold scSemiProfiler 38.5 GPU h (1.60 days, running 40.8 h) vs LoRA 0.62 h (62x); 5-donor draws 20.3 h vs 0.56 h; Imperial: 5-fold scSemiProfiler 30.9 GPU h (1.29 days, running 33.2 h) vs LoRA 0.52 h (60x); 5-donor draws 19.1 h vs 0.90 h; Giroux: 5-fold scSemiProfiler 14.4 GPU h (0.60 days, running 15.1 h) vs LoRA 0.30 h (49x); 5-donor draws 13.0 h vs 0.22 h; Ji (whole blood): 5-fold scSemiProfiler 7.4 GPU h (0.31 days, running 8.3 h) vs LoRA 0.37 h (20x); 5-donor draws 4.8 h vs 0.28 h",
+     "source": "data/fig3_time_days.csv"
+    },
+    {
+     "value": "all runs: scSemiProfiler 183.8 GPU h (7.66 days); PBISC LoRA 5.68 h + pretraining 15.9 h once",
+     "source": "data/fig3_time_days.csv"
+    }
+   ]
+  }
+ ]
+}

paper-releases/pbisc-diffusion/updates/20261008T060837Z-deck-new-figure-split-time/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure_plan.json

@@ -0,0 +1,369 @@
+{
+ "source": "Nature Style Manuscript Writing.pdf (paper-release-20260928/reference-papers): six main figures, up to 17 supplementary. User decisions of 2026-10-06: no PBISC-VAE; scSemiProfiler, a VAE-based method, is the comparison. User decisions of 2026-10-08 (supersede the earlier figure split): Fig. 2 shows no disease metrics; Fig. 3 compares with scSemiProfiler on healthy samples, with time as a main result; Figs 4-6 use the disease cohorts: Fig. 4 disease benchmarks (PBISC versions, scSemiProfiler, and BayesPrism / CIBERSORTx labelled as cell-type deconvolution), Fig. 5 rare-cell fractions and detection, Fig. 6 new findings from the bulk-only sepsis and COVID samples (K = 1,000 and 4,000).",
+ "status_legend": {
+  "Ready": "slide exists and fits the panel",
+  "Adapt": "slide exists; change its scope or style",
+  "Draft": "new draft exists, not yet approved",
+  "Build": "results exist; no slide yet",
+  "Run": "needs a new analysis or experiment",
+  "Plan": "not planned yet"
+ },
+ "figures": [
+  {
+   "n": 1,
+   "title": "Design and performance of PBISC",
+   "claim": "Diffusion on each gene's value and detection, conditioned on bulk, writes realistic blood cells.",
+   "panels": [
+    {
+     "p": "a",
+     "content": "Backbones compared: scDiffusion, SDD, PBISC-Diffusion",
+     "evidence": [
+      "Backbone: SDD → PBISC-Diffusion"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "b",
+     "content": "Overview: pretraining, diffusion, generation",
+     "evidence": [
+      "Overview draft (Nature style)"
+     ],
+     "status": "Draft"
+    },
+    {
+     "p": "c",
+     "content": "What the bulk condition contains",
+     "evidence": [
+      "Bulk conditioning"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "d",
+     "content": "Pretraining corpus: assays, conditions, cell types",
+     "evidence": [
+      "Pretraining corpus"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "e",
+     "content": "Width and context: lower loss, same biology",
+     "evidence": [
+      "Width & context ablation: loss saturation",
+      "Width & context ablation: benchmark"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "f",
+     "content": "Gene panel: 2k, 3k, 6k, 15k",
+     "evidence": [
+      "Gene context: 2k / 3k / 6k / 15k"
+     ],
+     "status": "Ready"
+    }
+   ],
+   "candidates": [],
+   "details": [
+    "PBISC-Diffusion: bulk in, cells out",
+    "Denoiser: joint value + support diffusion"
+   ]
+  },
+  {
+   "n": 2,
+   "title": "Fine-tuning PBISC with LoRA",
+   "claim": "On a healthy cohort, a few pairs and a LoRA of minutes give cells close to the real ones. No disease metrics.",
+   "panels": [
+    {
+     "p": "a",
+     "content": "Real bulk differs from the training input",
+     "evidence": [
+      "Real bulk ≠ pseudobulk (model input)",
+      "Same shift in three paired cohorts"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "b",
+     "content": "Without fine-tuning, cells fall off the real manifold",
+     "evidence": [
+      "Consequence: cells off-manifold"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "c",
+     "content": "LoRA on paired bulk and cells; how it is tested",
+     "evidence": [
+      "Fine-tuning with paired bulk + sc",
+      "Evaluation design"
+     ],
+     "status": "Adapt",
+     "note": "design example: ImMicro instead of sepsis"
+    },
+    {
+     "p": "d",
+     "content": "How generated cells are scored (cells and composition)",
+     "evidence": [
+      "How we score generated cells"
+     ],
+     "status": "Adapt",
+     "note": "disease rows move to Fig. 4"
+    },
+    {
+     "p": "e",
+     "content": "Healthy cohort: precision, coverage, composition",
+     "evidence": [
+      "5-fold CV: train on the other 4 folds"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only; drop the disease columns"
+    },
+    {
+     "p": "f",
+     "content": "Healthy cohort: classifier AUROC, MMD, E_N",
+     "evidence": [
+      "5-fold: realism beyond precision/coverage"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only"
+    },
+    {
+     "p": "g",
+     "content": "Full, LoRA and condition-encoder fine-tuning",
+     "evidence": [
+      "Full vs LoRA vs cond.-encoder-only"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only; no logFC"
+    },
+    {
+     "p": "h",
+     "content": "How many pairs are needed",
+     "evidence": [
+      "Pairs needed: composition, coverage, training steps"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only; scSemiProfiler diamonds go to Fig. 3"
+    },
+    {
+     "p": "i",
+     "content": "LoRA rank and K choices",
+     "evidence": [
+      "LoRA rank: r = 4, 16, 64",
+      "K diversification: no gain"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only; supplementary candidate"
+    }
+   ],
+   "candidates": [],
+   "details": []
+  },
+  {
+   "n": 3,
+   "title": "PBISC vs scSemiProfiler on healthy donors",
+   "claim": "On healthy donors, LoRA cells are closer to the real ones than scSemiProfiler's, in minutes instead of days.",
+   "panels": [
+    {
+     "p": "a",
+     "content": "Five donors: generated cells and realism",
+     "evidence": [
+      "5 donors: LoRA vs scSemiProfiler — cells",
+      "5 donors: realism beyond precision/coverage",
+      "5 donors: precision/coverage at k = 3, 5, 10"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only"
+    },
+    {
+     "p": "b",
+     "content": "Same training pairs (5-fold) and equal pair budget (5-40)",
+     "evidence": [
+      "Same training pairs: LoRA vs scSemiProfiler"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only; add the n = 5-40 budget runs"
+    },
+    {
+     "p": "c",
+     "content": "scSemiProfiler copies its representative",
+     "evidence": [
+      "scSemiProfiler copies its representative"
+     ],
+     "status": "Adapt",
+     "note": "ImMicro only"
+    },
+    {
+     "p": "d",
+     "content": "Time: GPU hours and days per experiment",
+     "evidence": [
+      "Time: days for scSemiProfiler, minutes for LoRA",
+      "5 donors: time to apply"
+     ],
+     "status": "Ready",
+     "note": "all cohorts shown: time is a property of the method"
+    }
+   ],
+   "candidates": [],
+   "details": []
+  },
+  {
+   "n": 4,
+   "title": "Disease signal on paired cohorts",
+   "claim": "LoRA cells recover the paired single cells' disease signal; compared with other PBISC versions, scSemiProfiler and deconvolution.",
+   "panels": [
+    {
+     "p": "a",
+     "content": "Five paired cohorts",
+     "evidence": [
+      "Five paired cohorts"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "b",
+     "content": "PBISC versions on the disease benchmarks: LoRA best",
+     "evidence": [],
+     "status": "Build",
+     "note": "zero-shot, calibration, sc-only, full FT, ranks, multi-K; data exist"
+    },
+    {
+     "p": "c",
+     "content": "LoRA vs scSemiProfiler: disease logFC, DEG",
+     "evidence": [
+      "5 donors: LoRA vs scSemiProfiler — disease signal",
+      "Pairs needed: donor and disease signal"
+     ],
+     "status": "Adapt",
+     "note": "add the 5-fold same-pairs disease columns"
+    },
+    {
+     "p": "d",
+     "content": "Disease genes at every list length (CAT)",
+     "evidence": [
+      "Disease genes at every list length (CAT)"
+     ],
+     "status": "Ready"
+    },
+    {
+     "p": "e",
+     "content": "vs BayesPrism and CIBERSORTx (cell-type deconvolution)",
+     "evidence": [
+      "Deconvolution with the same pairs: disease signal"
+     ],
+     "status": "Adapt",
+     "note": "label as cell-type deconvolution in the title"
+    },
+    {
+     "p": "f",
+     "content": "Lupus benchmark (sc-pseudobulk input) and the bulk broadcast",
+     "evidence": [
+      "SLE benchmark: PBISC vs scSemiProfiler",
+      "SLE benchmark: per cell type",
+      "Bulk broadcast baseline"
+     ],
+     "status": "Ready",
+     "note": "moved from Fig. 3"
+    }
+   ],
+   "candidates": [],
+   "details": [
+    "Deconvolution with the same pairs: composition"
+   ]
+  },
+  {
+   "n": 5,
+   "title": "Rare cell states in disease cohorts",
+   "claim": "LoRA cells carry rare disease states near their observed fractions, in the donors that have them.",
+   "panels": [
+    {
+     "p": "a",
+     "content": "Rare-state fractions, generated vs observed, four paired cohorts",
+     "evidence": [],
+     "status": "Run",
+     "note": "LoRA, scSemiProfiler, zero-shot; BayesPrism states where defined"
+    },
+    {
+     "p": "b",
+     "content": "Finding rare states: per-donor detection",
+     "evidence": [],
+     "status": "Run",
+     "note": "donors with vs without the state"
+    },
+    {
+     "p": "c",
+     "content": "Behçet: C1Q-high monocytes from real bulk",
+     "evidence": [
+      "Consequence: disease direction flips"
+     ],
+     "status": "Adapt",
+     "note": "add the LoRA-3 result (C1Q-high BD > HC)"
+    },
+    {
+     "p": "d",
+     "content": "Diseases absent from the corpus: Kawasaki MS1, SLE ISG-high T",
+     "evidence": [],
+     "status": "Build",
+     "note": "rare_states_screen_20261004 results"
+    }
+   ],
+   "candidates": [],
+   "details": []
+  },
+  {
+   "n": 6,
+   "title": "Discovery from public bulk-only data",
+   "claim": "New findings from LoRA cells of bulk-only sepsis and COVID dialysis samples (K = 1,000 and 4,000).",
+   "panels": [
+    {
+     "p": "a",
+     "content": "Bulk-only samples generated: sepsis 122, COVID dialysis 164",
+     "evidence": [
+      "Next: cohorts with a paired subset (planned)"
+     ],
+     "status": "Adapt",
+     "note": "update: K = 1,000 and 4,000 done"
+    },
+    {
+     "p": "b",
+     "content": "New findings",
+     "evidence": [],
+     "status": "Run",
+     "note": "earlier screen: 0 confirmed beyond the bulk"
+    }
+   ],
+   "candidates": [],
+   "details": []
+  }
+ ],
+ "supplementary": {
+  "slides": [
+   "Pairs from other cohorts do not transfer",
+   "Longer training on other cohorts: gap stays",
+   "No pretraining: same cohort, from scratch",
+   "No pretraining: other cohorts, from scratch",
+   "Rare states: from scratch vs pretrained"
+  ],
+  "planned": [
+   {
+    "p": "S",
+    "content": "Influenza exposure from raw clinical bulk (Randolph)",
+    "evidence": [],
+    "status": "Build",
+    "note": "manuscript result; placement open"
+   }
+  ],
+  "notes": {
+   "Pairs from other cohorts do not transfer": "moved from Fig. 5; placement open",
+   "Longer training on other cohorts: gap stays": "moved from Fig. 5; placement open"
+  }
+ },
+ "appendix": [
+  "Appendix: scSemiProfiler without output threshold"
+ ],
+ "discussion": [
+  "Next: other tissues from a few dozen pairs"
+ ]
+}