Archive 4 checkpoints (1-4)

paper-releases/pbisc-diffusion/updates/20261013T072655Z-enriched-140-slide-master/diffusion/dataset/server_1/model8/presentations/2026-09-30--paper-summary--r00/data/figure_enrichment_20261013/page_map.csv

@@ -0,0 +1,141 @@
+current_page,parent_reordered_page,original_expanded_page,original_76_page,group,panel,title,addition_key
+1,1,1,1,Introduction,Introduction,PBISC-Diffusion: manuscript evidence,
+2,2,2,2,Introduction,Introduction,Figures 1-6: the expanded evidence map,
+3,3,3,,Introduction,Introduction,How this expanded deck is organized,guide:scope
+4,4,4,3,Introduction,Current scope,Manuscript map: the six figures,
+5,5,6,4,Figure 1,Figure 1,Figure 1. Overview and corpus,
+6,6,7,5,Figure 1,Figure 1 plan,"Figure 1: overview, corpus and denoising",
+7,7,9,7,Figure 1,Fig. 1a,Experimental overview,
+8,8,11,9,Figure 1,Fig. 1b,Pretraining corpus,
+9,9,16,13,Figure 1,Fig. 1c,Corpus scale and generation workflow,
+10,,,,Figure 1,Fig. 1d,From Gaussian noise to generated cell states,enrichment:trajectory_four_pca_frames
+11,10,19,15,Figure 2,Figure 2,Figure 2. SLE model ablations,
+12,11,20,16,Figure 2,Figure 2 plan,Figure 2: architecture and gene context,
+13,,,,Figure 2,Fig. 2a,Where context width and trunk width change,enrichment:arch_context_vs_trunk
+14,,,,Figure 2,Fig. 2b,Predicted and oracle hints: one trained denoiser,enrichment:arch_hint_injection
+15,43,12,10,Figure 2,Fig. 2c,Architecture: full and late validation progression,
+16,12,13,11,Figure 2,Fig. 2d,Width & context ablation: benchmark,
+17,,,,Figure 2,Fig. 2 details,Architecture arms: dimensions and endpoints,enrichment:arch_width_arm_register
+18,13,14,12,Figure 2,Fig. 2e,Gene context: 2k / 3k / 6k / 15k,
+19,,,,Figure 2,Fig. 2f,Gene panels: progression relative to each endpoint,enrichment:arch_panel_loss_progression
+20,,,,Figure 2,Fig. 2 details,Gene-panel arms: exact sizes and training provenance,enrichment:arch_panel_arm_register
+21,,,,Figure 2,Fig. 2g,All ablation arms: parameters and training time,enrichment:arch_all_arm_compute
+22,,,,Figure 2,Fig. 2h,Real-cell repeat: a sampling reference,enrichment:real_cell_repeat
+23,14,15,,Figure 2,Fig. 2i,Native-pseudobulk model comparisons,review:6
+24,15,52,28,Figure 3,Figure 3,Figure 3. Real-bulk adaptation,
+25,16,53,29,Figure 3,Figure 3 plan,Figure 3: the paired-adaptation story,
+26,17,21,17,Figure 3,Fig. 3a,Bulk and pooled cells: different expression values,
+27,,,,Figure 3,Fig. 3a detail,Paired calibration reduces same-donor expression error,enrichment:mismatch_same_donor_errors
+28,18,23,19,Figure 3,Fig. 3b detail,Consequence: cells off-manifold,
+29,19,5,,Figure 3,Fig. 3b,Healthy paired data: full FT and LoRA,guide:availability
+30,20,51,,Figure 3,Fig. 3c,Full FT and LoRA ranks: quality and parameters,review:8
+31,21,49,26,Figure 3,Fig. 3d,Healthy paired-data budget: n training + 7 validation,
+32,22,54,30,Figure 3,Fig. 3e,"Five pairs, no extra validation: LoRA vs scSemi",
+33,23,55,,Figure 3,Fig. 3f,Composition: error and donor-level wins,review:13
+34,24,60,35,Figure 4,Figure 4,Figure 4. Disease signal preservation,
+35,25,61,36,Figure 4,Figure 4 plan,Figure 4: disease signal preservation,
+36,26,63,38,Figure 4,Fig. 4a,Disease signals with five training donors,
+37,27,67,39,Figure 4,Fig. 4b,5-fold CV: disease signal,
+38,28,68,40,Figure 4,Fig. 4c,Disease genes at every list length (CAT),
+39,29,71,41,Figure 4,Fig. 4d,"Cell-type deconvolution, same pairs: disease signal",
+40,30,86,47,Figure 5,Figure 5,Figure 5. Rare-population identity,
+41,31,87,48,Figure 5,Figure 5 plan,Figure 5: paired rare-population identity,
+42,32,102,53,Figure 6,Figure 6,Figure 6. Published-finding reproduction,
+43,33,103,54,Figure 6,Figure 6 plan,Figure 6: published-finding reproduction,
+44,34,107,,Figure 6,Figure 6,Figure 6: what the reproduction must establish,guide:figure6
+45,35,108,56,Supplementary,Supplementary,Supplementary evidence,
+46,36,109,57,Supplementary,Supplementary,Supplementary evidence map,
+47,37,8,6,Supp S01,Supp. S01,Backbone: SDD → PBISC-Diffusion,
+48,38,10,8,Supp S01,Supp. S01,Bulk conditioning,
+49,39,17,14,Supp S01,Supp. S01,Denoiser: joint value + support diffusion,
+50,40,38,20,Supp S01,Supp. S01,Fine-tuning with paired bulk + sc,
+51,41,39,21,Supp S01,Supp. S01,Evaluation design,
+52,42,41,22,Supp S01,Supp. S01,How we score generated cells,
+53,44,18,,Supp S02,Supp. S02,Does the model follow its bulk conditioning?,healthy_sle_conditioning_interventions
+54,45,72,42,Supp S02,Supp. S02,SLE benchmark: PBISC vs scSemiProfiler,
+55,46,73,43,Supp S02,Supp. S02,SLE benchmark: per cell type,
+56,47,74,,Supp S02,Supp. S02,Disjoint-cell controls test the bulk-broadcast baseline,healthy_sle_disjoint_truth_control
+57,48,75,,Supp S02,Supp. S02,SLE: what is added beyond the bulk broadcast?,disease_sle_broadcast
+58,49,22,18,Supp S03,Supp. S03,Same shift in three paired cohorts,
+59,50,24,,Supp S03,Supp. S03,Input corrections on one real-reference UMAP,healthy_stored_real_reference_umap
+60,51,25,,Supp S03,Supp. S03,Where generated-cell density differs from real cells,healthy_saved_density_difference
+61,52,26,,Supp S03,Supp. S03,Input calibration changes cell-type composition,healthy_input_control_composition
+62,53,27,,Supp S03,Supp. S03,"Same noise, different input: matched-cell shifts",healthy_stored_matched_trajectories
+63,54,28,,Supp S03,Supp. S03,Matched noise separates cell state from cell identity,healthy_matched_similarity
+64,55,29,,Supp S03,Supp. S03,Input changes can switch the generated cell label,healthy_label_transitions
+65,56,30,,Supp S03,Supp. S03,Nearest real cells: distance and label agreement,healthy_nearest_real_by_type
+66,57,31,,Supp S03,Supp. S03,Composition agreement does not ensure state agreement,healthy_within_type_centroids
+67,58,32,,Supp S03,Supp. S03,Gene-length effects in the bulk input,healthy_gene_length_input
+68,59,33,,Supp S03,Supp. S03,Gene-length effects within label-stable cells,healthy_gene_length_stable
+69,60,34,,Supp S03,Supp. S03,Raw and TPM shifts persist within cell labels,healthy_raw_tpm_top_genes
+70,61,35,,Supp S03,Supp. S03,Reference calibration leaves specific gene-state shifts,healthy_refcal_top_genes
+71,62,36,,Supp S03,Supp. S03,Donor trait axes and donor identification differ,healthy_donor_axes
+72,63,37,,Supp S03,Supp. S03,Reference choice changes different quality endpoints,healthy_reference_choice
+73,64,40,,Supp S04,Supp. S04,Why the historical and strict protocols differ,guide:protocol
+74,65,42,,Supp S04,Supp. S04,Technical repeats and follow-up are different controls,healthy_immicro_replicate_controls
+75,66,43,,Supp S04,Supp. S04,Annotation agreement and difficult categories,healthy_annotation_agreement
+76,67,46,,Supp S04,Supp. S04,Depth matching changes precision and coverage,healthy_depth_matched_quality
+77,68,47,,Supp S04,Supp. S04,Detection and library size are separate diagnostics,healthy_detection_and_library
+78,69,44,23,Supp S05,Supp. S05,Healthy cohort: LoRA cells cover the real ones,
+79,70,45,24,Supp S05,Supp. S05,Healthy cohort: realism beyond precision/coverage,
+80,71,48,25,Supp S05,Supp. S05,Healthy cohort: full vs LoRA vs condition encoder,
+81,72,57,32,Supp S05,Supp. S05,Representative-dependent composition,
+82,73,116,64,Supp S05,Supp. S05,Full vs LoRA vs cond.-encoder-only,
+83,74,50,27,Supp S06,Supp. S06,Healthy cohort: LoRA rank and K,
+84,75,56,31,Supp S06,Supp. S06,Healthy paired-data learning curves,
+85,76,58,33,Supp S06,Supp. S06,"Time: days for scSemiProfiler, minutes for LoRA",
+86,77,59,34,Supp S06,Supp. S06,5 donors: time to apply,
+87,78,117,65,Supp S06,Supp. S06,"Pairs needed: composition, coverage, training steps",
+88,79,118,66,Supp S06,Supp. S06,"LoRA rank: r = 4, 16, 64",
+89,80,119,67,Supp S06,Supp. S06,K diversification: no gain,
+90,81,62,37,Supp S07,Supp. S07,Five paired cohorts,
+91,82,64,,Supp S07,Supp. S07,Sepsis: exact-five reconstruction across five draws,disease_strict_sepsis_quality
+92,83,65,,Supp S07,Supp. S07,Sepsis: realism beyond precision and coverage,disease_strict_sepsis_realism
+93,84,66,,Supp S07,Supp. S07,Imperial: first completed exact-five comparison,disease_strict_imperial_quality
+94,85,110,58,Supp S07,Supp. S07,5 donors: LoRA vs scSemiProfiler — cells,
+95,86,111,59,Supp S07,Supp. S07,5 donors: realism beyond precision/coverage,
+96,87,112,60,Supp S07,Supp. S07,"5 donors: precision/coverage at k = 3, 5, 10",
+97,88,113,61,Supp S07,Supp. S07,5-fold: realism beyond precision/coverage,
+98,89,114,62,Supp S07,Supp. S07,Same training pairs: LoRA vs scSemiProfiler,
+99,90,115,63,Supp S07,Supp. S07,scSemiProfiler and the representative-copy baseline,
+100,91,69,,Supp S08,Supp. S08,Sepsis: patient-aware disease-gene agreement,disease_strict_ye2026_sepsis_deg
+101,92,70,,Supp S08,Supp. S08,Imperial: patient-aware disease-gene agreement,disease_strict_imperial_eskd_deg
+102,93,82,44,Supp S08,Supp. S08,Sepsis: fine-tuning variants,
+103,94,83,45,Supp S08,Supp. S08,Pairs needed: donor and disease signal,
+104,95,84,46,Supp S08,Supp. S08,"Cell-type deconvolution, same pairs: composition",
+105,96,85,,Supp S08,Supp. S08,Deconvolution: expression coverage limits the comparison,disease_deconv_expression_coverage
+106,97,120,,Supp S08,Supp. S08,"K=4,000 recipe: paired reconstruction quality",disease_k4000_quality
+107,98,121,,Supp S08,Supp. S08,"K=4,000 recipe: disease signal does not improve uniformly",disease_k4000_disease
+108,99,122,,Supp S08,Supp. S08,"K=4,000 recipe: calibration gains are selective",disease_k4000_calibration
+109,100,76,,Supp S09,Supp. S09,Influenza: response order versus response genes,disease_influenza_response
+110,101,77,,Supp S09,Supp. S09,COVID: calibration on a shared real-cell map,disease_covid_map
+111,102,78,,Supp S09,Supp. S09,COVID monocytes: direction after depth matching,disease_covid_direction
+112,103,79,,Supp S09,Supp. S09,"COVID monocytes: direction, magnitude and null",disease_covid_monocyte_direction_detail
+113,104,80,,Supp S09,Supp. S09,"COVID CD14 monocytes: direction, magnitude and null",disease_covid_cd14_direction_detail
+114,105,81,,Supp S09,Supp. S09,COVID: biological endpoints give a mixed verdict,disease_covid_endpoints
+115,106,89,,Supp S10,Supp. S10,Rare-state frequencies in observed cells,review:22
+116,107,91,,Supp S10,Supp. S10,Rare-state misses: rules 1–8,disease_rare_misses_1
+117,108,92,,Supp S10,Supp. S10,Rare-state misses: rules 9–15,disease_rare_misses_2
+118,109,93,,Supp S10,Supp. S10,Rare-state fidelity: zero-aware reanalysis,review:23
+119,110,95,,Supp S10,Supp. S10,Sepsis rare states across paired-data budgets,disease_sepsis_rare_support
+120,111,96,,Supp S10,Supp. S10,Sepsis rare-state direction across support budgets,disease_sepsis_rare_direction
+121,112,97,,Supp S11,Supp. S11,"Rare-state screen in DRESS, Kawasaki and SLE",disease_external_rare_screen
+122,113,98,52,Supp S11,Supp. S11,Consequence: disease direction flips,
+123,114,99,,Supp S11,Supp. S11,C1Q: single-gene detection versus co-detection,disease_c1q_cross_disease
+124,115,100,,Supp S11,Supp. S11,"Behçet: K=4,000 changes direction more than identity",disease_behcet_k4000
+125,116,101,,Supp S11,Supp. S11,Behcet: one-gene and multi-gene C1Q rules,review:24
+126,117,123,68,Supp S12,Supp. S12,Pairs from other cohorts do not transfer,
+127,118,124,69,Supp S12,Supp. S12,Longer training on other cohorts: gap stays,
+128,119,125,70,Supp S12,Supp. S12,"No pretraining: same cohort, from scratch",
+129,120,126,71,Supp S12,Supp. S12,"No pretraining: other cohorts, from scratch",
+130,121,127,72,Supp S12,Supp. S12,Rare states: from scratch vs pretrained,
+131,122,129,74,Supp S13,Supp. S13,Appendix: scSemiProfiler without output threshold,
+132,123,106,,Supp S14,Supp. S14,Bulk-only exploration: calibration populations differ,disease_bulkonly_imperial_wave_shift
+133,124,128,73,Historical archive,Historical archive,Historical scoring and plans,
+134,125,88,49,Historical archive,Historical only,Rare-state fractions: original analysis,
+135,126,90,50,Historical archive,Historical only,Rare-state ranking: original analysis,
+136,127,94,51,Historical archive,Historical only,Rare states: original five-fold analysis,
+137,128,104,55,Historical archive,Historical archive,Bulk-only cohort plan: original snapshot,
+138,129,105,,Historical archive,Historical archive,Historical bulk-only search: no confirmed new finding,disease_bulkonly_verification_history
+139,130,130,75,Discussion,Discussion,Discussion and outlook,
+140,131,131,76,Discussion,Discussion,Next: other tissues from a few dozen pairs,