Archive 4 checkpoints (1-4)

paper-releases/pbisc-diffusion/updates/20261003T103040Z-crosscohort-k1000-saturation/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/scores/giroux2022/xcalL_K1000/summary.json

@@ -0,0 +1,531 @@
+{
+ "schema": "paired_ft_scores.v1",
+ "dataset": "giroux2022",
+ "written_utc": "2026-10-02T15:38:49.572708+00:00",
+ "command": [
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/score_ft.py",
+  "--dataset",
+  "giroux2022",
+  "--threads",
+  "8",
+  "--donors",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/data/giroux2022/usable_samples.txt",
+  "--arm",
+  "xcalL_K1000=/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/generated/giroux2022/hoL_giroux2022/xcal_K1000.npz",
+  "--case-groups",
+  "severe",
+  "--control-groups",
+  "mild",
+  "--out",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/scores/giroux2022/xcalL_K1000"
+ ],
+ "arms": [
+  "xcalL_K1000"
+ ],
+ "donors": [
+  "450905_d0",
+  "450905_d14",
+  "0B943B_d0",
+  "0B943B_d14",
+  "7085CA_d0",
+  "7085CA_d7",
+  "7085CA_d14",
+  "2AD75E_d0",
+  "2AD75E_d7",
+  "2AD75E_d14",
+  "B85D75_d0",
+  "B85D75_d7",
+  "180E1A_d0",
+  "180E1A_d7",
+  "180E1A_d14",
+  "82CCF5_d0",
+  "82CCF5_d7",
+  "82CCF5_d14",
+  "BAAF62_d0",
+  "BAAF62_d7",
+  "BAAF62_d14",
+  "0BF51C_d0",
+  "0BF51C_d7",
+  "0BF51C_d14"
+ ],
+ "n_donors": 24,
+ "reference": {
+  "cache": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/giroux2022",
+  "reference_sha256": "dd0152f7bde41da80e1c71dabf54ee561982f36dc99a336b65fca419880a1593",
+  "features_sha256": "fb41b91617e3ff786d5153c4dc7869d66191ca5947ff28161d24215a24f819e5",
+  "n_r1_cells": 3600,
+  "pca": {
+   "n_components": 50,
+   "explained_variance_ratio_sum": 0.23158283531665802,
+   "n_fit_cells": 3600
+  },
+  "r2_control": {
+   "precision": 0.9980555555555556,
+   "coverage": 1.0,
+   "density": 3.7886666666666664,
+   "n_query": 3600,
+   "depth_matched": {
+    "precision": 0.9980555555555556,
+    "coverage": 1.0,
+    "density": 3.7033611111111115
+   }
+  },
+  "observed_l1_loo_null_mean": 0.30584824369245694
+ },
+ "contrast_definition": {
+  "enabled": true,
+  "rule": "explicit --case-groups / --control-groups",
+  "group_source": "features groups",
+  "case_groups": [
+   "severe"
+  ],
+  "control_groups": [
+   "mild"
+  ],
+  "n_case_donors": 11,
+  "n_control_donors": 13,
+  "n_donors_in_neither": 0,
+  "case": [
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   false,
+   false,
+   false,
+   true,
+   true,
+   true
+  ],
+  "ctrl": [
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   true,
+   true,
+   true,
+   false,
+   false,
+   false
+  ],
+  "groups": [
+   "mild",
+   "mild",
+   "mild",
+   "mild",
+   "mild",
+   "mild",
+   "mild",
+   "mild",
+   "mild",
+   "mild",
+   "severe",
+   "severe",
+   "severe",
+   "severe",
+   "severe",
+   "severe",
+   "severe",
+   "severe",
+   "mild",
+   "mild",
+   "mild",
+   "severe",
+   "severe",
+   "severe"
+  ]
+ },
+ "deg_definition": {
+  "enabled": true,
+  "score": "top-50 DEG overlap per cell type (benchmark total_top50_overlap, Welch backend)",
+  "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/bulk2sc-paper/scripts/evaluate_generated_vs_truth.py",
+  "functions": [
+   "welch_bh_deg",
+   "top_deg_genes",
+   "jaccard"
+  ],
+  "reproduces": "utility_studies_final6k_20260925/sle top-50 DEG 81 (1,006 genes) / 84 (2,279 genes) = sum over the six gated types of top50_overlap_count",
+  "top_n": 50,
+  "fdr": 0.05,
+  "n_genes": 5563,
+  "types": [
+   "T cell",
+   "B cell",
+   "Monocyte",
+   "NK cell",
+   "DC"
+  ],
+  "aggregates_source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/giroux2022/deg_reference.npz",
+  "aggregates_sha256": "412c44bba7481b0b851c3b12e695f1b7ff662163e1f7849d3a64359dfce6d50e",
+  "aggregates_note": null,
+  "types_not_scored_observed": [],
+  "observed": {
+   "T cell": {
+    "n_significant": 306,
+    "n_tested": 5354,
+    "n_case_donors": 11,
+    "n_control_donors": 13,
+    "n_cells": 145034,
+    "top_list_size": 50
+   },
+   "B cell": {
+    "n_significant": 9,
+    "n_tested": 5142,
+    "n_case_donors": 11,
+    "n_control_donors": 13,
+    "n_cells": 17383,
+    "top_list_size": 9
+   },
+   "Monocyte": {
+    "n_significant": 7,
+    "n_tested": 5295,
+    "n_case_donors": 11,
+    "n_control_donors": 13,
+    "n_cells": 47311,
+    "top_list_size": 7
+   },
+   "NK cell": {
+    "n_significant": 83,
+    "n_tested": 5159,
+    "n_case_donors": 11,
+    "n_control_donors": 13,
+    "n_cells": 29061,
+    "top_list_size": 50
+   },
+   "DC": {
+    "n_significant": 7,
+    "n_tested": 4962,
+    "n_case_donors": 11,
+    "n_control_donors": 13,
+    "n_cells": 4328,
+    "top_list_size": 7
+   }
+  },
+  "broadcast": "welch_bh_deg + top_deg_genes on the arm's condition bulk_counts (scoring genes, case|control donors), one list compared with every type's observed list (sle_broadcast.py broadcast_input_bulk); not part of the published benchmark"
+ },
+ "headline": {
+  "xcalL_K1000": {
+   "comp_l1": 0.20217456364466954,
+   "comp_l1_obs_loo_null": 0.30584824369245694,
+   "precision": 0.9880555555555556,
+   "coverage": 0.8911111111111111,
+   "precision_dm": 0.9791666666666666,
+   "coverage_dm": 0.8886111111111111,
+   "pb_r_median": 0.885349841361257,
+   "pb_top1": 0.25,
+   "mono_pct_gen": 19.11111111111111,
+   "mono_pct_obs": 19.247866360995445,
+   "median_umi_gen": 693.0,
+   "median_umi_obs": 471.0,
+   "contrast_mean_cp10k_r_gen": 0.31966885822675173,
+   "contrast_pseudobulk_log1p_cpm_r_gen": 0.3986518439756842,
+   "contrast_mean_cp10k_r_broadcast": 0.6370077946331492,
+   "contrast_pseudobulk_log1p_cpm_r_broadcast": 0.7884339553946482,
+   "deg_top50_T": 7,
+   "deg_top50_B": 4,
+   "deg_top50_Mono": 4,
+   "deg_top50_NK": 5,
+   "deg_top50_DC": 1,
+   "deg_top50_sum": 21,
+   "deg_top50_max": 123,
+   "deg_top50_sum_4type": 20,
+   "deg_top50_sum_broadcast": 36,
+   "seconds": 8.5
+  }
+ },
+ "per_arm": {
+  "xcalL_K1000": {
+   "arm": "xcalL_K1000",
+   "npz": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/generated/giroux2022/hoL_giroux2022/xcal_K1000.npz",
+   "npz_sha256": "a87b772fdd1c50698b49e53eaef51f96544a37b1887b02645b88f41cee928d8c",
+   "receipt_verified": true,
+   "generation": {
+    "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/checkpoint_epoch_016.pt",
+    "checkpoint_sha256": "8c7cd0e83e2bc19fe03dfcbc084919a1e2f61be8cc8bc12f4caf543dfda91633",
+    "input_condition_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/splits/hoL_giroux2022/test_xcal_K1000.npz",
+    "input_condition_sha256": "b791302e74c54f0117048083448e10447112b0ca9ded8af001a107792532ba53",
+    "seed": 20260927,
+    "steps": 50,
+    "support_readout": "bernoulli",
+    "weights": "ema"
+   },
+   "condition_for_broadcast": {
+    "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/splits/hoL_giroux2022/test_xcal_K1000.npz",
+    "sha256": "b791302e74c54f0117048083448e10447112b0ca9ded8af001a107792532ba53"
+   },
+   "poisson": {
+    "seed": 20260916,
+    "rule": "whole NPZ, rows in NPZ order",
+    "nnz_whole_npz": 12287317,
+    "median_library_whole_npz_panel": 2000.5,
+    "n_rows_whole_npz": 36000
+   },
+   "annotation": {
+    "model_pt_sha256": "bf3655899e5abee435340fc008656c4a17342da0f30b70cfd43de8e9223e1930",
+    "zero_padded_model_genes": 0,
+    "label_counts": {
+     "T cell": 22683,
+     "Monocyte": 6880,
+     "NK cell": 3779,
+     "B cell": 1743,
+     "DC": 536,
+     "Other": 379
+    },
+    "mean_l2_max_prob": 0.9470252502337098,
+    "predict_seconds": 3.9
+   },
+   "n_donors": 24,
+   "n_generated_cells": 36000,
+   "composition": {
+    "l1_mean": 0.20217456364466954,
+    "l1_median": 0.19673270000947024,
+    "l1_obs_loo_null_mean": 0.30584824369245694,
+    "l1_generated_loo_mean_vs_observed_mean": 0.3099280594774676,
+    "composition_top1": 0.16666666666666666,
+    "mean_fraction_generated": {
+     "T cell": 0.630083,
+     "B cell": 0.048417,
+     "Monocyte": 0.191111,
+     "NK cell": 0.104972,
+     "DC": 0.014889,
+     "Other": 0.010528
+    },
+    "mean_fraction_observed": {
+     "T cell": 0.566621,
+     "B cell": 0.065423,
+     "Monocyte": 0.192479,
+     "NK cell": 0.114152,
+     "DC": 0.016843,
+     "Other": 0.044482
+    }
+   },
+   "pseudobulk": {
+    "n_genes": 5563,
+    "r_median": 0.885349841361257,
+    "r_mean": 0.8821736818553639,
+    "top1": 0.25,
+    "n_top1": 6,
+    "own_rank_median": 5.0,
+    "chance_top1": 0.041666666666666664
+   },
+   "precision_coverage": {
+    "precision": 0.9880555555555556,
+    "coverage": 0.8911111111111111,
+    "density": 1.5855000000000001,
+    "n_query": 3600,
+    "n_reference": 3600
+   },
+   "precision_coverage_depth_matched": {
+    "precision": 0.9791666666666666,
+    "coverage": 0.8886111111111111,
+    "density": 1.4932777777777777,
+    "detected_median_before": {
+     "R1": 254.0,
+     "Q": 258.0
+    },
+    "detected_median_after": {
+     "R1": 239.0,
+     "Q": 239.0
+    }
+   },
+   "monocyte_pct": {
+    "generated_mean_over_donors": 19.11111111111111,
+    "observed_mean_over_donors": 19.247866360995445,
+    "generated_pooled_cells": 19.11111111111111
+   },
+   "median_umi_per_cell_scoring_genes": {
+    "generated_pooled": 693.0,
+    "observed_pooled": 471.0,
+    "generated_mean_of_donor_medians": 701.5625,
+    "observed_mean_of_donor_medians": 483.6458333333333
+   },
+   "contrast": {
+    "mean_cp10k": {
+     "per_type": {
+      "T cell": {
+       "r_generated": 0.46807968377476183,
+       "r_broadcast": 0.663884796401021,
+       "partial_r_generated_given_broadcast": 0.1378394024331679,
+       "r_generated_vs_broadcast": 0.578399935965255,
+       "n_genes": 2908
+      },
+      "B cell": {
+       "r_generated": 0.18489080648085457,
+       "r_broadcast": 0.5645155961562724,
+       "partial_r_generated_given_broadcast": 0.0642267115185676,
+       "r_generated_vs_broadcast": 0.23626910571239182,
+       "n_genes": 3053
+      },
+      "Monocyte": {
+       "r_generated": 0.29373332218950676,
+       "r_broadcast": 0.6980894090417796,
+       "partial_r_generated_given_broadcast": 0.1138857039889066,
+       "r_generated_vs_broadcast": 0.3097011856278731,
+       "n_genes": 3056
+      },
+      "NK cell": {
+       "r_generated": 0.3319716204618838,
+       "r_broadcast": 0.6215413769335235,
+       "partial_r_generated_given_broadcast": 0.16384350896165084,
+       "r_generated_vs_broadcast": 0.3398993588786841,
+       "n_genes": 2955
+      }
+     },
+     "mean_r_generated": 0.31966885822675173,
+     "mean_r_broadcast": 0.6370077946331492,
+     "mean_partial_r": 0.11994883172557323
+    },
+    "pseudobulk_log1p_cpm": {
+     "per_type": {
+      "T cell": {
+       "r_generated": 0.5781521360920436,
+       "r_broadcast": 0.8229501008768898,
+       "partial_r_generated_given_broadcast": 0.174416270575966,
+       "r_generated_vs_broadcast": 0.6068337956615151,
+       "n_genes": 2908
+      },
+      "B cell": {
+       "r_generated": 0.2293292420424647,
+       "r_broadcast": 0.7209755672221485,
+       "partial_r_generated_given_broadcast": 0.05554884466742807,
+       "r_generated_vs_broadcast": 0.2666241778418804,
+       "n_genes": 3053
+      },
+      "Monocyte": {
+       "r_generated": 0.3838923902659595,
+       "r_broadcast": 0.8425716586211682,
+       "partial_r_generated_given_broadcast": 0.08994132248074382,
+       "r_generated_vs_broadcast": 0.4030034433605325,
+       "n_genes": 3056
+      },
+      "NK cell": {
+       "r_generated": 0.4032336075022692,
+       "r_broadcast": 0.767238494858386,
+       "partial_r_generated_given_broadcast": 0.20685286722172372,
+       "r_generated_vs_broadcast": 0.36454851808197813,
+       "n_genes": 2955
+      }
+     },
+     "mean_r_generated": 0.3986518439756842,
+     "mean_r_broadcast": 0.7884339553946482,
+     "mean_partial_r": 0.1316898262364654
+    }
+   },
+   "deg_top50": {
+    "types_scored": [
+     "T cell",
+     "B cell",
+     "Monocyte",
+     "NK cell",
+     "DC"
+    ],
+    "types_not_scored": {},
+    "per_type": {
+     "T cell": {
+      "deg_scored": true,
+      "top50_overlap": 7,
+      "top50_jaccard": 0.1346153846153846,
+      "obs_top_list_size": 50,
+      "gen_top_list_size": 9,
+      "top50_overlap_broadcast": 9,
+      "top50_jaccard_broadcast": 0.16071428571428573,
+      "n_cells_obs": 145034,
+      "n_cells_gen": 22683
+     },
+     "B cell": {
+      "deg_scored": true,
+      "top50_overlap": 4,
+      "top50_jaccard": 0.4444444444444444,
+      "obs_top_list_size": 9,
+      "gen_top_list_size": 4,
+      "top50_overlap_broadcast": 7,
+      "top50_jaccard_broadcast": 0.4117647058823529,
+      "n_cells_obs": 17383,
+      "n_cells_gen": 1743
+     },
+     "Monocyte": {
+      "deg_scored": true,
+      "top50_overlap": 4,
+      "top50_jaccard": 0.5714285714285714,
+      "obs_top_list_size": 7,
+      "gen_top_list_size": 4,
+      "top50_overlap_broadcast": 7,
+      "top50_jaccard_broadcast": 0.4666666666666667,
+      "n_cells_obs": 47311,
+      "n_cells_gen": 6880
+     },
+     "NK cell": {
+      "deg_scored": true,
+      "top50_overlap": 5,
+      "top50_jaccard": 0.09090909090909091,
+      "obs_top_list_size": 50,
+      "gen_top_list_size": 10,
+      "top50_overlap_broadcast": 8,
+      "top50_jaccard_broadcast": 0.14035087719298245,
+      "n_cells_obs": 29061,
+      "n_cells_gen": 3779
+     },
+     "DC": {
+      "deg_scored": true,
+      "top50_overlap": 1,
+      "top50_jaccard": 0.14285714285714285,
+      "obs_top_list_size": 7,
+      "gen_top_list_size": 1,
+      "top50_overlap_broadcast": 5,
+      "top50_jaccard_broadcast": 0.29411764705882354,
+      "n_cells_obs": 4328,
+      "n_cells_gen": 536
+     }
+    },
+    "sum": 21,
+    "sum_4type": 20,
+    "max": 123,
+    "mean": 4.2,
+    "sum_broadcast": 36,
+    "sum_broadcast_4type": 31,
+    "broadcast_n_significant": 15,
+    "seconds": 0.07
+   },
+   "seconds": 8.5
+  }
+ },
+ "code_sha256": {
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/score_ft.py": "ec5f40b4cea6d5062d0138f3b934abb6465e34cbfeac869c8a0906924c033aec",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/ft_common.py": "4a6366b0f22208cd58714fc0261b8840c643b9d4569704d560cfd93534cbc747",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tools/stage_b_embed.py": "d056cfdf38fb0e5f478756f370174e993b75d0c973c48e02e21749175a4ca2da",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/bulk2sc-paper/scripts/evaluate_generated_vs_truth.py": "48b12d794782b840838853e93e0fc55fab00532c4b3698f4588e926b985d2881",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/tools/immicro_common.py": "c5c30e917148cc5e664e91ea30c11bbdff9b6d1f53a3e0a13e16b1917eff37e9",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/aifi_scanvi_rebuild_20260916/predict_scanvi.py": "482ee5dcf4d72b54bc36cbb6dcff62e5030bc7afc46cb5fb01eca73a82ee45a5",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/aifi_scanvi_rebuild_20260916/poisson_readout.py": "8bfa080942fbcdc89ae23b276524a1a83266fb8cbc56b9af279e50f82e6cde38"
+ },
+ "seconds": 9.9
+}

paper-releases/pbisc-diffusion/updates/20261003T103040Z-crosscohort-k1000-saturation/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/scores/giroux2022/zs_K2000/README.md

@@ -0,0 +1,73 @@
+# score_ft.py output: giroux2022, 24 scored donors
+
+Arms: zs_K2000.  Every arm is scored on the same donors against the same observed reference.
+
+## Inputs and fixed contracts
+- Observed reference: `/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/giroux2022/reference.npz` (sha dd0152f7bde4), built from datadir:/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/giroux2022;
+  observed cells sum exactly to the features pseudobulk of every donor. Observed labels: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/giroux2022/observed_labels/labels/labels.parquet.
+- Generated cells: `generated_counts` read out once with Poisson(seed 20260916) over the **whole NPZ in row order**
+  (`poisson_readout.poisson_integerise`), then the scored donors' rows. These integer counts are used for labels, embedding,
+  depth and the contrast. Pseudobulk uses the summed `generated_counts` (rates), as `score_generated.py`.
+- Labels: frozen f01common scANVI (`predict_scanvi.py`, CPU, zero query epochs), six-type gate ['T cell', 'B cell', 'Monocyte', 'NK cell', 'DC', 'Other']; query genes =
+  the 1,006-gene annotation list minus genes the dataset's sc assay did not measure in every sample (1006 genes).
+- Scoring genes: 6k panel genes measured by both assays in every sample of the dataset (5563 genes).
+
+## Definitions
+1. **Composition**: fraction of each donor's cells per six-type label (all generated cells; all observed QC cells).
+   `comp_l1` = sum over the six labels of |generated - observed| (immicro_common.l1). `comp_l1_obs_loo_null` = L1 of the
+   donor's observed composition to the mean of the *other scored donors'* observed compositions. `composition_top1` =
+   fraction of donors whose generated composition is closest (L1) to their own observed composition.
+2. **Pseudobulk**: `pb_r` = Pearson r of CPM-log1p(summed generated_counts) vs CPM-log1p(observed pseudobulk) on the scoring
+   genes (evaluate_realbulk.cpm_log1p / paired_r). `pb_top1` = the donor's generated pseudobulk is closest (Pearson) to its own
+   observed pseudobulk among the 24 scored donors (immicro_common.retrieval = evaluate_realbulk.top1 arithmetic).
+3. **Precision / coverage / density** (stage_b_embed.py functions): per-cell CP10K over the scoring genes + log1p; PCA-50
+   (randomized, random_state 20260925) and k = 10 real-real k-NN balls fitted on **R1 of the scored donors only**.
+   Reference rule (independent of fold and arm): per donor, cells in canonical order (library_id+barcode), one
+   permutation from `default_rng([20260925, uint64(sha256(donor)[:8])])`; R1 = first n, R2 = next n,
+   n = min(150, n_cells // 2) (3600 R1 cells here). Generated query: per donor n cells (the donor's R1 size),
+   without replacement, `default_rng([20260926, same key])`. Precision = fraction of generated cells inside any R1 ball
+   (Kynkaanniemi 2019); coverage = fraction of R1 cells whose ball holds a generated cell; density = Naeem 2020.
+   Per-donor precision = that donor's generated cells; per-donor coverage = that donor's R1 cells (the pooled values are
+   their n-weighted means; plain means when every donor has 150 reference cells). R2 (held-out real cells of the same donors) is the
+   perfect-generator control: precision 0.9981, coverage 1.0000.
+   **Depth matched**: stage_b_embed depth matching run pairwise (R1 vs this arm): per donor and within-donor rank of detected
+   genes, both cells are thinned by UMI sampling without replacement to the smaller detected-gene count; PCA kept from
+   un-thinned R1, radii recomputed on thinned R1; fresh rng(20260929) per arm. R2 control: precision
+   0.9981, coverage 1.0000.
+4. **Monocyte %** = 100 x Monocyte fraction (all cells) per donor; **median UMI per cell** = median over cells of the
+   Poisson-read (generated) / observed counts summed over the scoring genes (per donor, and pooled).
+5. **Group contrast** (T, B, Monocyte, NK): case groups ['severe'] (n = 11) vs control groups ['mild'] (n = 13); rule: explicit --case-groups / --control-groups; groups from features groups.
+   Genes per type = scoring genes detected (count > 0) in >= 1 % of the scored donors' observed
+   cells of that type. Primary statistic `mean_cp10k`: per-cell CP10K over the scoring genes; per donor the mean over its
+   cells of the type (donors with >= 1 such cell); group mean = mean over the group's donors (donor-balanced);
+   logFC = log2(case mean + 1) - log2(control mean + 1). Secondary `pseudobulk_log1p_cpm`: the SLE benchmark evaluator's
+   statistic (donor x type summed counts, log1p CPM over the scoring genes, mean case - mean control;
+   evaluate_generated_vs_truth.group_log_fold_change). Both are computed on the generated cells (their frozen labels) and on
+   the observed cells; `r_generated` = Pearson r of the two logFC vectors. Broadcast baseline `r_broadcast`: the same
+   observed vector against the logFC of the input bulk (condition NPZ `bulk_counts` of the arm, scoring genes, per-donor
+   CP10K resp. log1p CPM, group means). `partial_r_generated_given_broadcast`: r of the OLS residuals of observed and
+   generated logFC on the bulk logFC (residual_specificity.py).
+6. **Top-50 DEG overlap** (contrast.csv rows `statistic = welch_top50_deg`; summary `deg_top50`): the SLE benchmark's score (`total_top50_overlap` of
+   `bulk2sc-paper/scripts/evaluate_generated_vs_truth.py`, Welch backend: the "top-50 DEG 81 / 84" of the final model), run
+   with the evaluator's own functions on the same donor x type pseudobulks as the contrast. Per cell type, observed and
+   generated separately: donor x type summed counts on all 5563 scoring genes (donors of the two groups with
+   >= 1 cell of the type) -> `welch_bh_deg` (log1p CPM, two-sided Welch t-test case vs control, Benjamini-Hochberg over the
+   genes with a defined p-value; needs >= 2 donors per group) -> `top_deg_genes` (FDR < 0.05, descending |logFC|,
+   first 50; shorter when fewer genes pass). `top50_overlap` = |observed list & generated list|; `top50_jaccard`
+   the evaluator's Jaccard. Summary `sum` = total over the scored types (the quantity reported as 81), `max` = the sum of
+   the observed list lengths (the "of 262"), `sum_4type` = T + B + Monocyte + NK. A type is scored when its observed test
+   is defined; an arm whose generated test is undefined for a scored type gets overlap 0. Types: T cell, B cell, Monocyte, NK cell, DC
+   (aggregates: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/giroux2022/deg_reference.npz).
+   **Broadcast analogue** (`top50_overlap_broadcast`, as `sle_broadcast.py` `broadcast_input_bulk`; not part of the
+   published benchmark): the same test on the arm's condition bulk_counts (scoring genes, the 11 + 13
+   donors), one whole-PBMC top-50 list, compared with every type's observed list.
+
+## Files
+- `per_donor.csv`: one row per arm x donor (all per-donor quantities above).
+- `summary.json`: per arm the donor means / pooled values, provenance (NPZ sha, receipt check, generation metadata,
+  Poisson nnz of the whole NPZ, annotation contract), and the reference block (PCA, R2 control).
+- `contrast.csv`: per arm x cell type x statistic (`mean_cp10k`, `pseudobulk_log1p_cpm`: logFC r; `welch_top50_deg`:
+  top-50 DEG overlap, T / B / Monocyte / NK / DC; columns of the other kind are empty).
+- `arms/<arm>/labels/` (predict_scanvi.py labels + manifest), `arms/<arm>/generated_cells.parquet` (NPZ row, donor, label,
+  UMI on the scoring genes, P/C subsample flag), `arms/<arm>/deg_top50_lists.csv` (observed, generated and broadcast
+  top-50 gene lists, Ensembl ids, rank order).

paper-releases/pbisc-diffusion/updates/20261003T103040Z-crosscohort-k1000-saturation/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/scores/giroux2022/zs_K2000/contrast.csv

@@ -0,0 +1,14 @@
+arm,celltype,statistic,n_genes,r_generated,r_broadcast,partial_r_generated_given_broadcast,r_generated_vs_broadcast,n_case_donors_obs,n_control_donors_obs,n_case_donors_gen,n_control_donors_gen,n_cells_obs,n_cells_gen,deg_scored,top50_overlap,top50_jaccard,obs_top_list_size,gen_top_list_size,n_sig_obs,n_sig_gen,top50_overlap_broadcast,top50_jaccard_broadcast,broadcast_top_list_size,n_sig_broadcast
+zs_K2000,T cell,mean_cp10k,2908,0.5703418449570322,0.6665223768938867,0.24397802332866586,0.647817719104607,11,13,11,13,145034,18065,,,,,,,,,,,
+zs_K2000,T cell,pseudobulk_log1p_cpm,2908,0.6603100348831036,0.829771536027708,0.20581230579622808,0.6964336632564426,11,13,11,13,145034,18065,,,,,,,,,,,
+zs_K2000,B cell,mean_cp10k,3053,0.27622352722365384,0.5741068019580355,0.03138623816908417,0.44096063532049995,11,13,11,13,17383,4245,,,,,,,,,,,
+zs_K2000,B cell,pseudobulk_log1p_cpm,3053,0.3397277050225295,0.7295906969005199,0.041760137910813996,0.430306894745198,11,13,11,13,17383,4245,,,,,,,,,,,
+zs_K2000,Monocyte,mean_cp10k,3056,0.4325303404151604,0.6998157471451955,0.2506330494230102,0.3815927576104611,11,13,11,13,47311,9725,,,,,,,,,,,
+zs_K2000,Monocyte,pseudobulk_log1p_cpm,3056,0.5394322930557204,0.8460762046394,0.22967230623175044,0.5133921120199054,11,13,11,13,47311,9725,,,,,,,,,,,
+zs_K2000,NK cell,mean_cp10k,2955,0.4050195322059327,0.6119759414285395,0.23728452630354793,0.3779130008150191,11,13,11,13,29061,2975,,,,,,,,,,,
+zs_K2000,NK cell,pseudobulk_log1p_cpm,2955,0.39215266217692546,0.7735167267842559,0.18916457239118353,0.3625264555905113,11,13,11,13,29061,2975,,,,,,,,,,,
+zs_K2000,T cell,welch_top50_deg,5563,,,,,11,13,11,13,145034,18065,True,5,0.09090909090909091,50,10,306,10,10,0.1724137931034483,18,18
+zs_K2000,B cell,welch_top50_deg,5563,,,,,11,13,11,13,17383,4245,True,4,0.4444444444444444,9,4,9,4,7,0.35,18,18
+zs_K2000,Monocyte,welch_top50_deg,5563,,,,,11,13,11,13,47311,9725,True,5,0.5555555555555556,7,7,7,7,7,0.3888888888888889,18,18
+zs_K2000,NK cell,welch_top50_deg,5563,,,,,11,13,11,13,29061,2975,True,4,0.08,50,4,83,4,10,0.1724137931034483,18,18
+zs_K2000,DC,welch_top50_deg,5563,,,,,11,13,11,13,4328,657,True,3,0.42857142857142855,7,3,7,3,5,0.25,18,18

paper-releases/pbisc-diffusion/updates/20261003T103040Z-crosscohort-k1000-saturation/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/scores/giroux2022/zs_K2000/per_donor.csv

@@ -0,0 +1,25 @@
+arm,donor,group,n_generated_cells,n_observed_cells,n_reference_cells,gen_T cell,obs_T cell,gen_B cell,obs_B cell,gen_Monocyte,obs_Monocyte,gen_NK cell,obs_NK cell,gen_DC,obs_DC,gen_Other,obs_Other,comp_l1,comp_l1_obs_loo_null,pb_r,pb_own_rank,pb_top1,precision,coverage,precision_depth_matched,coverage_depth_matched,mono_pct_gen,mono_pct_obs,median_umi_gen,median_umi_obs
+zs_K2000,450905_d0,mild,1500,5914,150,0.216,0.30943523841731485,0.03266666666666666,0.036185322962461956,0.642,0.4293202570172472,0.051333333333333335,0.12089956036523504,0.04733333333333333,0.009976327358809604,0.010666666666666666,0.09418329387893135,0.5000734979145531,0.6266504923198378,0.8817465064034697,2,False,0.8933333333333333,0.24666666666666667,0.9866666666666667,0.32666666666666666,64.2,42.93202570172472,1211.0,574.5
+zs_K2000,450905_d14,mild,1500,7407,150,0.46266666666666667,0.4422843256379101,0.07933333333333334,0.02929661131362225,0.324,0.27420008100445525,0.11133333333333334,0.15782368030241664,0.014,0.02983664101525584,0.008666666666666666,0.06655866072633995,0.2404379640880248,0.36155260046011806,0.8902371144458188,7,False,0.9133333333333333,0.32666666666666666,0.96,0.36666666666666664,32.4,27.420008100445525,1048.5,558.0
+zs_K2000,0B943B_d0,mild,1500,11344,150,0.52,0.6220028208744711,0.09266666666666666,0.08312764456981664,0.298,0.12588152327221439,0.064,0.11865303244005641,0.019333333333333334,0.023977433004231313,0.006,0.026357545839210156,0.3633149976492713,0.1791278384873501,0.886844649737469,11,False,0.8666666666666667,0.28,0.94,0.32,29.799999999999997,12.588152327221438,949.5,539.0
+zs_K2000,0B943B_d14,mild,1500,15650,150,0.376,0.5652396166134185,0.154,0.11175718849840256,0.342,0.17086261980830672,0.10333333333333333,0.09789137380191694,0.016666666666666666,0.009968051118210863,0.008,0.04428115015974441,0.4510415335463258,0.10063602424453227,0.8868164993237605,1,True,0.92,0.2,0.9866666666666667,0.3,34.2,17.08626198083067,921.0,461.0
+zs_K2000,7085CA_d0,mild,1500,12585,150,0.18933333333333333,0.21660707191100517,0.036,0.024791418355184745,0.7053333333333334,0.5874453714739769,0.04466666666666667,0.13198251887167262,0.018666666666666668,0.017719507350019866,0.006,0.021454112038140644,0.26008740564163696,0.8829287440863147,0.8988383151665826,1,True,0.86,0.20666666666666667,0.9533333333333334,0.24,70.53333333333333,58.74453714739769,1315.5,640.0
+zs_K2000,7085CA_d7,mild,1500,9515,150,0.5633333333333334,0.521492380451918,0.11066666666666666,0.0722017866526537,0.16133333333333333,0.10110352075669994,0.13866666666666666,0.2725170782974251,0.018666666666666668,0.019863373620599055,0.007333333333333333,0.012821860220704152,0.2810712909441233,0.3632282731474772,0.8932516033210693,7,False,0.9066666666666666,0.26666666666666666,0.9933333333333333,0.30666666666666664,16.133333333333333,10.110352075669994,931.0,502.0
+zs_K2000,7085CA_d14,mild,1500,11887,150,0.5826666666666667,0.472533019264743,0.1,0.07133843694792631,0.17133333333333334,0.16210986792294102,0.12733333333333333,0.24270211155043325,0.012666666666666666,0.018591738874400604,0.006,0.032724825439555814,0.2960373517287793,0.2993283707516418,0.8950101255856584,2,False,0.9266666666666666,0.34,0.96,0.42,17.133333333333333,16.2109867922941,890.5,536.0
+zs_K2000,2AD75E_d0,mild,1500,7693,150,0.4633333333333333,0.6034056934875861,0.134,0.06408423241908229,0.268,0.14064734173924348,0.09266666666666666,0.15949564539191471,0.024666666666666667,0.025347718705316522,0.017333333333333333,0.007019368256856883,0.41516478183630146,0.20681142084856263,0.8671495507952579,23,False,0.94,0.17333333333333334,1.0,0.22,26.8,14.064734173924348,775.5,532.0
+zs_K2000,2AD75E_d7,mild,1500,10499,150,0.4246666666666667,0.5808172206876845,0.22066666666666668,0.14372797409277074,0.224,0.10277169254214687,0.10466666666666667,0.14715687208305553,0.008,0.013715591961139155,0.018,0.011810648633203162,0.4087127027970917,0.27874882058987843,0.8235212959384495,23,False,0.9733333333333334,0.12,0.9933333333333333,0.18666666666666668,22.400000000000002,10.277169254214687,619.0,443.0
+zs_K2000,2AD75E_d14,mild,1500,13692,150,0.5426666666666666,0.5811422728600643,0.15,0.12503651767455448,0.19333333333333333,0.14198071866783524,0.078,0.12496348232544552,0.025333333333333333,0.016579024247735902,0.010666666666666666,0.010297984224364592,0.17087817703768624,0.19762923258298543,0.8859144645092987,19,False,0.9333333333333333,0.06666666666666667,0.9866666666666667,0.12,19.333333333333332,14.198071866783524,941.0,428.0
+zs_K2000,B85D75_d0,severe,1500,11441,150,0.6913333333333334,0.7173324010138974,0.09266666666666666,0.031203566121842496,0.13466666666666666,0.11441307578008915,0.06733333333333333,0.06039681845992483,0.008666666666666666,0.016519535005681322,0.005333333333333333,0.06013460361856481,0.17730641260962032,0.35882358586377083,0.8892217884406611,10,False,0.94,0.36,0.9866666666666667,0.5666666666666667,13.466666666666665,11.441307578008916,955.0,347.0
+zs_K2000,B85D75_d7,severe,1500,12328,150,0.5846666666666667,0.6435756002595717,0.14,0.06448734587929916,0.18466666666666667,0.13741077222582737,0.07,0.04201817001946788,0.011333333333333334,0.027741726151849447,0.009333333333333334,0.08476638546398442,0.30150075708414453,0.28301793033829503,0.8921712312986567,4,False,0.9333333333333333,0.2,0.9933333333333333,0.2866666666666667,18.46666666666667,13.741077222582737,936.0,363.0
+zs_K2000,180E1A_d0,severe,1500,7157,150,0.532,0.7019700992035769,0.13533333333333333,0.04135811094033813,0.21333333333333335,0.13175911694844208,0.08133333333333333,0.0655302501047925,0.024,0.013273718038284197,0.014,0.04610870476456616,0.40415760793628613,0.31372800769589615,0.8875183693952524,7,False,0.9533333333333334,0.44666666666666666,0.98,0.4666666666666667,21.333333333333336,13.175911694844208,867.0,433.0
+zs_K2000,180E1A_d7,severe,1500,10591,150,0.6266666666666667,0.7348692285903126,0.12,0.04909829100179398,0.15933333333333333,0.09064299877254273,0.07733333333333334,0.08280615617033331,0.01,0.010480596733075253,0.006666666666666667,0.03210272873194221,0.27918408711799325,0.37628768415060193,0.8891025008776866,9,False,0.94,0.3933333333333333,1.0,0.5133333333333333,15.933333333333334,9.064299877254273,901.5,408.0
+zs_K2000,180E1A_d14,severe,1500,12511,150,0.542,0.680840860043162,0.12533333333333332,0.05147470226200943,0.22,0.12277196067460634,0.08733333333333333,0.0800095915594277,0.016666666666666666,0.009831348413396212,0.008666666666666666,0.05507153704739829,0.37049146084778717,0.28939519394073066,0.8892148570866728,8,False,0.9466666666666667,0.47333333333333333,1.0,0.6066666666666667,22.0,12.277196067460634,951.0,342.0
+zs_K2000,82CCF5_d0,severe,1500,10125,150,0.398,0.5422222222222223,0.22466666666666665,0.11832098765432099,0.26466666666666666,0.1539753086419753,0.08866666666666667,0.09303703703703704,0.015333333333333332,0.016691358024691356,0.008666666666666666,0.07575308641975309,0.4340740740740741,0.18419072822032903,0.8745873760166466,6,False,0.9666666666666667,0.36666666666666664,0.9866666666666667,0.42,26.466666666666665,15.397530864197531,793.5,390.0
+zs_K2000,82CCF5_d7,severe,1500,10860,150,0.472,0.619060773480663,0.17,0.09972375690607735,0.22333333333333333,0.14152854511970533,0.10333333333333333,0.09410681399631676,0.023333333333333334,0.013627992633517496,0.008,0.03195211786372008,0.3420257826887662,0.19419086678828923,0.8896584568807178,2,False,0.9333333333333333,0.18,0.9866666666666667,0.24,22.333333333333332,14.152854511970533,911.0,497.0
+zs_K2000,82CCF5_d14,severe,1500,11056,150,0.4866666666666667,0.6330499276410999,0.17266666666666666,0.10121201157742403,0.22066666666666668,0.11369392185238784,0.09,0.06828871201157742,0.021333333333333333,0.014562228654124458,0.008666666666666666,0.0691931982633864,0.4138195851423059,0.2787071044352101,0.8934433090857309,2,False,0.9066666666666666,0.44,0.9933333333333333,0.5266666666666666,22.066666666666666,11.369392185238784,946.0,394.0
+zs_K2000,BAAF62_d0,mild,1500,14741,150,0.5713333333333334,0.5279153381724442,0.114,0.0384641476154942,0.18666666666666668,0.1854690997897022,0.10733333333333334,0.1717658232141646,0.014666666666666666,0.023132758971575875,0.006,0.053252832236618954,0.24030282884471887,0.14602913946264867,0.8925786995255315,8,False,0.9066666666666666,0.25333333333333335,0.9866666666666667,0.37333333333333335,18.666666666666668,18.54690997897022,956.0,449.0
+zs_K2000,BAAF62_d7,mild,1500,7297,150,0.598,0.6932986158695355,0.126,0.04070165821570509,0.16133333333333333,0.13389063998903658,0.09133333333333334,0.0941482801151158,0.017333333333333333,0.016033986569823216,0.006,0.021926819240783883,0.2280807637842036,0.25991971231178185,0.8833145319268779,12,False,0.92,0.24666666666666667,0.9533333333333334,0.26666666666666666,16.133333333333333,13.389063998903659,888.5,580.0
+zs_K2000,BAAF62_d14,mild,1500,11923,150,0.6533333333333333,0.5431518913025245,0.07266666666666667,0.019206575526293717,0.16666666666666666,0.2843244149962258,0.08,0.10416841398976767,0.020666666666666667,0.032793759959741674,0.006666666666666667,0.016354944225446616,0.3272830663423635,0.2201331132920326,0.893220945041193,2,False,0.92,0.12666666666666668,0.9666666666666667,0.20666666666666667,16.666666666666664,28.43244149962258,894.5,513.0
+zs_K2000,0BF51C_d0,severe,1500,6510,150,0.23266666666666666,0.4012288786482335,0.019333333333333334,0.01674347158218126,0.6786666666666666,0.4609831029185868,0.024666666666666667,0.07726574500768049,0.018,0.006144393241167435,0.02666666666666667,0.03763440860215054,0.46425806451612905,0.5556611725994968,0.8783655798270676,1,True,0.8,0.2866666666666667,0.98,0.26,67.86666666666666,46.09831029185868,925.5,628.0
+zs_K2000,0BF51C_d7,severe,1500,10105,150,0.6473333333333333,0.612370113805047,0.096,0.05700148441365661,0.19266666666666668,0.1607125185551707,0.048666666666666664,0.05294408708560119,0.012666666666666666,0.0073231073725878275,0.0026666666666666666,0.10964868876793667,0.22251888504040904,0.22247826940326798,0.8829493888202711,16,False,0.8733333333333333,0.35333333333333333,0.98,0.38,19.26666666666667,16.071251855517072,990.0,509.0
+zs_K2000,0BF51C_d14,severe,1500,11419,150,0.6666666666666666,0.6330676941938874,0.11133333333333334,0.07960416849111131,0.14733333333333334,0.151589456169542,0.05,0.07907872843506436,0.018666666666666668,0.010508801120938787,0.006,0.04615115158945617,0.1469720057214583,0.1611535225979176,0.8920295589920224,9,False,0.9333333333333333,0.2866666666666667,0.96,0.32666666666666666,14.733333333333334,15.1589456169542,942.5,541.0