Archive 4 checkpoints (1-4)
Browse files- paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/cross_run.log +52 -0
- paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/verify_confound_cross_run.log +0 -0
- paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/verify_confound_within_run.log +100 -0
- paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/verify_recompute_cross_extra_report.txt +59 -0
paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/cross_run.log
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
informative: 15 of 15 | min obs thinned: 629
|
| 2 |
+
locoL_scratchsat_s162000_K1000 vs locoL_lora16sat_s26750_K1000 [own_ref]: scratch closer 1, comparator closer 5, n.s. 7
|
| 3 |
+
locoL_scratchsat_s162000_K1000 vs locoL_lora16sat_s26750_K1000 [pretrained_depth_ref]: scratch closer 3, comparator closer 3, n.s. 7
|
| 4 |
+
locoL_scratchsat_s162000_K1000 vs locoL_lora64sat_s59000_K1000 [own_ref]: scratch closer 1, comparator closer 4, n.s. 8
|
| 5 |
+
locoL_scratchsat_s162000_K1000 vs locoL_lora64sat_s59000_K1000 [pretrained_depth_ref]: scratch closer 4, comparator closer 2, n.s. 7
|
| 6 |
+
locoL_scratchsat_s162000_K1000 vs locoL_fullsat_s108500_K1000 [own_ref]: scratch closer 2, comparator closer 6, n.s. 5
|
| 7 |
+
locoL_scratchsat_s162000_K1000 vs locoL_fullsat_s108500_K1000 [pretrained_depth_ref]: scratch closer 3, comparator closer 3, n.s. 7
|
| 8 |
+
locoL_scratchsat_s162000_K1000 vs xcalL_K1000 [own_ref]: scratch closer 1, comparator closer 6, n.s. 6
|
| 9 |
+
locoL_scratchsat_s162000_K1000 vs xcalL_K1000 [pretrained_depth_ref]: scratch closer 2, comparator closer 5, n.s. 6
|
| 10 |
+
locoL_scratchsat_s162000_K1000 vs locoL_fullsat_s20000_K1000 [own_ref]: scratch closer 0, comparator closer 6, n.s. 7
|
| 11 |
+
locoL_scratchsat_s162000_K1000 vs locoL_fullsat_s20000_K1000 [pretrained_depth_ref]: scratch closer 0, comparator closer 3, n.s. 10
|
| 12 |
+
locoL_scratchsat_s50000_K1000 vs locoL_lora16sat_s26750_K1000 [own_ref]: scratch closer 2, comparator closer 3, n.s. 8
|
| 13 |
+
locoL_scratchsat_s50000_K1000 vs locoL_lora16sat_s26750_K1000 [pretrained_depth_ref]: scratch closer 5, comparator closer 3, n.s. 5
|
| 14 |
+
locoL_scratchsat_s162000_K1000 vs locoL_scratchsat_s50000_K1000 [own_ref]: scratch closer 3, comparator closer 5, n.s. 5
|
| 15 |
+
locoL_scratchsat_s162000_K1000 vs locoL_scratchsat_s50000_K1000 [pretrained_depth_ref]: scratch closer 3, comparator closer 5, n.s. 5
|
| 16 |
+
locoL_scratchsat_s162000_K1000 vs locoL_scratchsat_s100000_K1000 [own_ref]: scratch closer 2, comparator closer 3, n.s. 8
|
| 17 |
+
locoL_scratchsat_s162000_K1000 vs locoL_scratchsat_s100000_K1000 [pretrained_depth_ref]: scratch closer 2, comparator closer 3, n.s. 8
|
| 18 |
+
locoL_scratchsat_s162000_K1000 vs locoL_scratchsat_s150000_K1000 [own_ref]: scratch closer 0, comparator closer 2, n.s. 11
|
| 19 |
+
locoL_scratchsat_s162000_K1000 vs locoL_scratchsat_s150000_K1000 [pretrained_depth_ref]: scratch closer 0, comparator closer 2, n.s. 11
|
| 20 |
+
arm variant states eligibility n_pairs kept reversed flat median_gen_over_obs_diff_kept lost_pairs
|
| 21 |
+
zs__raw own_ref rare_states eligible_own 21 15 6 0 0.091808 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cycling monocyte (Ye 2026) | adult sepsis vs adult HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 22 |
+
zs__raw own_ref rare_states eligible_common 21 15 6 0 0.091808 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cycling monocyte (Ye 2026) | adult sepsis vs adult HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 23 |
+
xcalL_K1000 own_ref rare_states eligible_own 25 19 6 0 0.228907 Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; DN2 / ABC B | adult sepsis vs adult HC; DN2 / ABC B | paediatric sepsis vs paediatric HC
|
| 24 |
+
xcalL_K1000 own_ref rare_states eligible_common 21 19 2 0 0.228907 Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC
|
| 25 |
+
locoL_lora16sat_s26750_K1000 own_ref rare_states eligible_own 23 11 12 0 0.137021 C1Q-high monocyte (reference) | adult sepsis vs adult HC; C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; C1Q-high monocyte (reference) | all sepsis vs all controls; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | adult sepsis vs adult HC; Granulocyte-like (LDG) | all sepsis vs all controls
|
| 26 |
+
locoL_lora16sat_s26750_K1000 own_ref rare_states eligible_common 21 11 10 0 0.137021 C1Q-high monocyte (reference) | adult sepsis vs adult HC; C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; C1Q-high monocyte (reference) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | adult sepsis vs adult HC; Granulocyte-like (LDG) | all sepsis vs all controls
|
| 27 |
+
locoL_lora64sat_s59000_K1000 own_ref rare_states eligible_own 22 10 12 0 0.131103 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; C1Q-high monocyte (reference) | all sepsis vs all controls; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cycling monocyte (Ye 2026) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | adult sepsis vs adult HC; Granulocyte-like (LDG) | paediatric sepsis vs paediatric HC; Granulocyte-like (LDG) | all sepsis vs all controls
|
| 28 |
+
locoL_lora64sat_s59000_K1000 own_ref rare_states eligible_common 21 10 11 0 0.131103 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; C1Q-high monocyte (reference) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cycling monocyte (Ye 2026) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | adult sepsis vs adult HC; Granulocyte-like (LDG) | paediatric sepsis vs paediatric HC; Granulocyte-like (LDG) | all sepsis vs all controls
|
| 29 |
+
locoL_fullsat_s20000_K1000 own_ref rare_states eligible_own 23 15 8 0 0.289959 C1Q-high monocyte (reference) | adult sepsis vs adult HC; C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; C1Q-high monocyte (reference) | all sepsis vs all controls; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 30 |
+
locoL_fullsat_s20000_K1000 own_ref rare_states eligible_common 21 15 6 0 0.289959 C1Q-high monocyte (reference) | adult sepsis vs adult HC; C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; C1Q-high monocyte (reference) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 31 |
+
locoL_fullsat_s108500_K1000 own_ref rare_states eligible_own 23 16 7 0 0.149520 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 32 |
+
locoL_fullsat_s108500_K1000 own_ref rare_states eligible_common 21 16 5 0 0.149520 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 33 |
+
locoL_scratchsat_s50000_K1000 own_ref rare_states eligible_own 26 17 9 0 0.073919 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; DN2 / ABC B | adult sepsis vs adult HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 34 |
+
locoL_scratchsat_s50000_K1000 own_ref rare_states eligible_common 21 15 6 0 0.073919 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 35 |
+
locoL_scratchsat_s50000_K1000 pretrained_depth_ref rare_states eligible_own 23 15 8 0 0.116441 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 36 |
+
locoL_scratchsat_s50000_K1000 pretrained_depth_ref rare_states eligible_common 21 15 6 0 0.116441 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 37 |
+
locoL_scratchsat_s100000_K1000 own_ref rare_states eligible_own 26 17 9 0 0.036039 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; DN2 / ABC B | paediatric sepsis vs paediatric HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 38 |
+
locoL_scratchsat_s100000_K1000 own_ref rare_states eligible_common 21 15 6 0 0.035548 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 39 |
+
locoL_scratchsat_s100000_K1000 pretrained_depth_ref rare_states eligible_own 23 15 8 0 0.054710 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 40 |
+
locoL_scratchsat_s100000_K1000 pretrained_depth_ref rare_states eligible_common 21 15 6 0 0.054710 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 41 |
+
locoL_scratchsat_s150000_K1000 own_ref rare_states eligible_own 26 18 8 0 0.035622 Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; DN2 / ABC B | paediatric sepsis vs paediatric HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 42 |
+
locoL_scratchsat_s150000_K1000 own_ref rare_states eligible_common 21 16 5 0 0.033845 Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 43 |
+
locoL_scratchsat_s150000_K1000 pretrained_depth_ref rare_states eligible_own 23 16 7 0 0.047973 Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 44 |
+
locoL_scratchsat_s150000_K1000 pretrained_depth_ref rare_states eligible_common 21 16 5 0 0.047973 Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls
|
| 45 |
+
locoL_scratchsat_s162000_K1000 own_ref rare_states eligible_own 26 16 10 0 0.049172 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; DN2 / ABC B | paediatric sepsis vs paediatric HC; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | paediatric sepsis vs paediatric HC
|
| 46 |
+
locoL_scratchsat_s162000_K1000 own_ref rare_states eligible_common 21 14 7 0 0.041622 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | paediatric sepsis vs paediatric HC
|
| 47 |
+
locoL_scratchsat_s162000_K1000 pretrained_depth_ref rare_states eligible_own 23 14 9 0 0.067088 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | paediatric sepsis vs paediatric HC
|
| 48 |
+
locoL_scratchsat_s162000_K1000 pretrained_depth_ref rare_states eligible_common 21 14 7 0 0.067088 C1Q-high monocyte (reference) | paediatric sepsis vs paediatric HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; Cytotoxic CD4 T | adult sepsis vs adult HC; Cytotoxic CD4 T | all sepsis vs all controls; Granulocyte-like (LDG) | paediatric sepsis vs paediatric HC
|
| 49 |
+
lora_n5__raw own_ref rare_states eligible_own 25 18 7 0 0.258039 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls; DN2 / ABC B | paediatric sepsis vs paediatric HC
|
| 50 |
+
lora_n5__raw own_ref rare_states eligible_common 21 17 4 0 0.245401 C1Q-high monocyte (reference) | adult sepsis vs adult HC; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC; ISG-high T | all sepsis vs all controls
|
| 51 |
+
real_lora__raw own_ref rare_states eligible_own 26 22 4 0 0.420212 Cycling activated T (CD38/HLA-DR) | paediatric sepsis vs paediatric HC; Cycling activated T (CD38/HLA-DR) | all sepsis vs all controls; Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC
|
| 52 |
+
real_lora__raw own_ref rare_states eligible_common 21 19 2 0 0.344018 Proliferating T (label) | all sepsis vs all controls; Plasmablast (markers) | paediatric sepsis vs paediatric HC
|
paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/verify_confound_cross_run.log
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/verify_confound_within_run.log
ADDED
|
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[ 0.0s] donors 17 {'adult_sepsis': np.int64(9), 'ped_sepsis': np.int64(3), 'adult_HC': np.int64(2), 'adult_ICU_ctrl': np.int64(1), 'ped_HC': np.int64(1), 'ped_ICU_ctrl': np.int64(1)} rare states 13
|
| 2 |
+
[ 0.1s] wrote verify_confound_within_repro.csv (240, 36)
|
| 3 |
+
[ 0.1s] max |diff| vs published: {'D': 4.883762290130989e-07, 'D_ci_lo': 4.962792746754374e-07, 'D_ci_hi': 4.992577684603816e-07, 'D_bc_ci_lo': 4.913921749505334e-07, 'D_bc_ci_hi': 4.897366950862292e-07}
|
| 4 |
+
[ 0.1s] state-level verdict counts (main, 13 rare): {'not significant': 42, 'scratch closer': 7, 'pretrained closer': 3} BC: {'not significant': 32, 'scratch closer': 11, 'pretrained closer': 9}
|
| 5 |
+
[ 0.1s] wrote verify_confound_within_stateset.csv (49, 11)
|
| 6 |
+
[ 1.0s] wrote verify_confound_within_lodo.csv (112, 16)
|
| 7 |
+
[ 1.0s] wrote verify_confound_within_donormean.csv (112, 13)
|
| 8 |
+
[ 1.0s] wrote verify_confound_within_concentration.csv (15, 28)
|
| 9 |
+
[ 1.0s] loading observed cells via tools/sepsis_states.observed() ...
|
| 10 |
+
[ 4.6s] observed cells (131629, 73)
|
| 11 |
+
[ 6.3s] V0 replay: per-donor count rows not reproduced (arm x state): 0 of 120
|
| 12 |
+
[ 10.9s] wrote verify_confound_within_depth_arms.csv (520, 14)
|
| 13 |
+
[ 11.1s] wrote verify_confound_within_depth.csv (576, 16)
|
| 14 |
+
[ 11.1s] V0_published main state verdicts: {'not significant': 42, 'scratch closer': 7, 'pretrained closer': 3} | aggregate: [('5 pairs: scratch vs Lo', 0.224, -0.078, 0.406), ('58 pairs: scratch vs L', 0.019, -0.35, 0.255), ('58 pairs: scratch vs f', 0.137, -0.138, 0.406), ('58 pairs + recipes: sc', -0.089, -0.379, 0.178), ('ref: scratch 5 pairs v', -0.239, -0.515, 0.005), ('ref: scratch 58 pairs ', -0.35, -0.66, -0.128), ('ref: scratch 58 + reci', -0.457, -0.729, -0.238)]
|
| 15 |
+
[ 11.1s] V0_reseed main state verdicts: {'not significant': 41, 'scratch closer': 7, 'pretrained closer': 4} | aggregate: [('5 pairs: scratch vs Lo', 0.224, -0.079, 0.405), ('58 pairs: scratch vs L', 0.023, -0.349, 0.255), ('58 pairs: scratch vs f', 0.132, -0.14, 0.401), ('58 pairs + recipes: sc', -0.092, -0.37, 0.173), ('ref: scratch 5 pairs v', -0.238, -0.515, 0.011), ('ref: scratch 58 pairs ', -0.351, -0.667, -0.112), ('ref: scratch 58 + reci', -0.452, -0.729, -0.232)]
|
| 16 |
+
[ 11.1s] V1_lib_trained main state verdicts: {'not significant': 41, 'scratch closer': 8, 'pretrained closer': 3} | aggregate: [('5 pairs: scratch vs Lo', 0.096, -0.226, 0.345), ('58 pairs: scratch vs L', 0.02, -0.363, 0.257), ('58 pairs: scratch vs f', 0.16, -0.127, 0.427), ('58 pairs + recipes: sc', -0.092, -0.379, 0.177)]
|
| 17 |
+
[ 11.1s] V1_lib_all main state verdicts: {'not significant': 42, 'scratch closer': 8, 'pretrained closer': 2} | aggregate: [('5 pairs: scratch vs Lo', -0.035, -0.465, 0.236), ('58 pairs: scratch vs L', -0.105, -0.471, 0.234), ('58 pairs: scratch vs f', 0.033, -0.282, 0.362), ('58 pairs + recipes: sc', -0.127, -0.467, 0.177), ('ref: scratch 5 pairs v', -0.11, -0.435, 0.169), ('ref: scratch 58 pairs ', -0.293, -0.645, -0.003), ('ref: scratch 58 + reci', -0.428, -0.729, -0.17)]
|
| 18 |
+
[ 11.1s] V2_msum_trained main state verdicts: {'not significant': 39, 'scratch closer': 7, 'pretrained closer': 6} | aggregate: [('5 pairs: scratch vs Lo', 0.262, -0.076, 0.49), ('58 pairs: scratch vs L', -0.011, -0.381, 0.332), ('58 pairs: scratch vs f', 0.057, -0.204, 0.373), ('58 pairs + recipes: sc', -0.048, -0.359, 0.253)]
|
| 19 |
+
[ 11.1s] V2_msum_all main state verdicts: {'not significant': 39, 'scratch closer': 7, 'pretrained closer': 6} | aggregate: [('5 pairs: scratch vs Lo', 0.259, -0.088, 0.494), ('58 pairs: scratch vs L', -0.02, -0.395, 0.319), ('58 pairs: scratch vs f', 0.022, -0.251, 0.357), ('58 pairs + recipes: sc', -0.037, -0.34, 0.259), ('ref: scratch 5 pairs v', -0.383, -0.792, -0.093), ('ref: scratch 58 pairs ', -0.578, -0.948, -0.308), ('ref: scratch 58 + reci', -0.695, -1.057, -0.424)]
|
| 20 |
+
[ 15.1s] wrote verify_confound_within_composition_arms.csv (128, 10)
|
| 21 |
+
[ 15.1s] wrote verify_confound_within_composition.csv (119, 15)
|
| 22 |
+
[ 15.1s] composition aggregate:
|
| 23 |
+
comparison std_ratio_scratch std_ratio_pretrained D_std D_std_ci_lo D_std_ci_hi verdict_std verdict_std_bc n_states_pretrained_closer n_states_scratch_closer
|
| 24 |
+
0 5 pairs: scratch vs LoRA 1.522 1.526 -0.004 -0.231 0.170 not significant not significant 1 4
|
| 25 |
+
1 58 pairs: scratch vs LoRA 1.403 1.426 -0.024 -0.310 0.170 not significant not significant 2 1
|
| 26 |
+
2 58 pairs: scratch vs full FT 1.403 1.270 0.133 -0.074 0.321 not significant not significant 2 2
|
| 27 |
+
3 58 pairs + recipes: scratch vs LoRA 1.267 1.473 -0.206 -0.438 0.058 not significant not significant 1 3
|
| 28 |
+
4 ref: scratch 5 pairs vs zero-shot 1.522 1.842 -0.321 -0.628 -0.131 scratch closer scratch closer 1 2
|
| 29 |
+
5 ref: scratch 58 pairs vs zero-shot 1.403 1.842 -0.440 -0.754 -0.216 scratch closer scratch closer 1 3
|
| 30 |
+
6 ref: scratch 58 + recipes vs zero-shot 1.267 1.842 -0.575 -0.881 -0.337 scratch closer scratch closer 0 3
|
| 31 |
+
[ 15.6s] wrote verify_confound_within_stateset_by_scheme.csv (147, 10)
|
| 32 |
+
[ 15.6s] state-set x scheme (main comparisons):
|
| 33 |
+
scheme V0_published V0_reseed V1_lib_all V1_lib_trained V2_msum_all V2_msum_trained composition_std
|
| 34 |
+
state_set comparison
|
| 35 |
+
11 marker states (no label states) 5 pairs: scratch vs LoRA +0.410[-0.00,+0.65]b +0.410[-0.00,+0.65]b +0.104[-0.45,+0.44] +0.258[-0.17,+0.58] +0.451[-0.00,+0.75]b +0.454[+0.01,+0.75]*b +0.099[-0.21,+0.34]
|
| 36 |
+
58 pairs + recipes: scratch vs LoRA +0.055[-0.30,+0.31] +0.051[-0.29,+0.31] +0.010[-0.39,+0.30] +0.051[-0.31,+0.31] +0.116[-0.25,+0.39] +0.103[-0.27,+0.39] -0.100[-0.41,+0.14]
|
| 37 |
+
58 pairs: scratch vs LoRA +0.194[-0.23,+0.42] +0.198[-0.23,+0.43] +0.047[-0.36,+0.37] +0.195[-0.23,+0.41] +0.148[-0.27,+0.45] +0.158[-0.26,+0.47] +0.147[-0.19,+0.31]
|
| 38 |
+
58 pairs: scratch vs full FT +0.354[+0.02,+0.62]*b +0.348[+0.02,+0.61]*b +0.231[-0.14,+0.57] +0.381[+0.04,+0.66]*b +0.218[-0.12,+0.55] +0.259[-0.06,+0.57] +0.344[+0.06,+0.53]*b
|
| 39 |
+
13 rare 5 pairs: scratch vs LoRA +0.224[-0.08,+0.41]b +0.224[-0.08,+0.40]b -0.035[-0.47,+0.24] +0.096[-0.23,+0.34] +0.259[-0.09,+0.49] +0.262[-0.08,+0.49] -0.004[-0.23,+0.17]
|
| 40 |
+
58 pairs + recipes: scratch vs LoRA -0.089[-0.38,+0.18] -0.092[-0.37,+0.17] -0.127[-0.47,+0.18] -0.092[-0.38,+0.18] -0.037[-0.34,+0.26] -0.048[-0.36,+0.25] -0.206[-0.44,+0.06]
|
| 41 |
+
58 pairs: scratch vs LoRA +0.019[-0.35,+0.25] +0.023[-0.35,+0.25] -0.105[-0.47,+0.23] +0.020[-0.36,+0.26] -0.020[-0.39,+0.32] -0.011[-0.38,+0.33] -0.024[-0.31,+0.17]
|
| 42 |
+
58 pairs: scratch vs full FT +0.137[-0.14,+0.41] +0.132[-0.14,+0.40] +0.033[-0.28,+0.36] +0.160[-0.13,+0.43] +0.022[-0.25,+0.36] +0.057[-0.20,+0.37] +0.133[-0.07,+0.32]
|
| 43 |
+
T/NK states (5) 5 pairs: scratch vs LoRA +0.281[-0.22,+0.73] +0.281[-0.22,+0.73] +0.252[-0.74,+0.89] +0.218[-0.27,+0.63] +0.295[-0.23,+0.76] +0.307[-0.22,+0.78] +0.166[-0.26,+0.50]
|
| 44 |
+
58 pairs + recipes: scratch vs LoRA +0.483[+0.01,+0.91]*b +0.476[+0.02,+0.90]*b +0.417[-0.21,+0.89] +0.476[+0.02,+0.90]*b +0.493[+0.08,+0.93]*b +0.487[+0.08,+0.92]*b +0.140[-0.20,+0.50]
|
| 45 |
+
58 pairs: scratch vs LoRA +0.481[+0.07,+0.88]*b +0.501[+0.08,+0.90]*b +0.601[+0.03,+1.08]*b +0.514[+0.12,+0.90]*b +0.640[+0.16,+1.11]*b +0.643[+0.18,+1.11]*b +0.420[+0.08,+0.68]*b
|
| 46 |
+
58 pairs: scratch vs full FT +0.455[+0.01,+1.02]*b +0.452[+0.01,+1.02]*b +0.545[-0.17,+1.19] +0.464[+0.02,+1.01]*b +0.435[-0.07,+1.05] +0.471[-0.04,+1.07] +0.446[+0.01,+0.77]*b
|
| 47 |
+
[ 15.6s] wrote verify_confound_within_arm_mean_abslog2.csv (56, 5)
|
| 48 |
+
[ 15.6s] arm mean |log2| by scheme:
|
| 49 |
+
scheme V0_published V0_reseed V1_lib_all V1_lib_trained V2_msum_all V2_msum_trained composition_std
|
| 50 |
+
arm
|
| 51 |
+
lora_n5__raw 1.300 1.300 1.686 1.473 1.305 1.299 1.526
|
| 52 |
+
lora_recipe__raw 1.395 1.402 1.460 1.409 1.289 1.283 1.473
|
| 53 |
+
real_full__raw 1.275 1.278 1.435 1.278 1.347 1.322 1.270
|
| 54 |
+
real_lora__raw 1.393 1.388 1.573 1.418 1.389 1.390 1.426
|
| 55 |
+
scratch_full__raw 1.412 1.410 1.468 1.438 1.369 1.379 1.403
|
| 56 |
+
scratch_n5__raw 1.524 1.524 1.651 1.569 1.564 1.560 1.522
|
| 57 |
+
scratch_recipe__raw 1.306 1.310 1.333 1.317 1.252 1.235 1.267
|
| 58 |
+
zs__raw 1.763 1.762 1.761 1.878 1.947 1.966 1.842
|
| 59 |
+
[ 15.7s] wrote verify_confound_within_labelcheck.csv (12, 40)
|
| 60 |
+
[ 15.7s] eligible (state, contrast) pairs: 25
|
| 61 |
+
[ 16.2s] wrote verify_confound_within_contrasts.csv (351, 31)
|
| 62 |
+
[ 16.2s] contrast summary:
|
| 63 |
+
arm n_eligible kept reversed_strong_abs reversed_weak_abs flat reversed_strong_rel reversed_weak_rel kept_lodo_min kept_lodo_max kept_by_contrast kept_diff_pretrained_minus_scratch discordant_pretrained_only_kept discordant_scratch_only_kept exact_sign_test_p kept_diff_lodo_min kept_diff_lodo_max
|
| 64 |
+
zs__raw 25 21.0 1.0 3.0 0.0 1.0 3.0 12.0 22.0 adult sepsis vs adult HC:6/9; paediatric sepsis vs paediatric HC:8/8; all sepsis vs all controls:7/8 NaN NaN NaN NaN NaN NaN
|
| 65 |
+
lora_n5__raw 25 23.0 2.0 0.0 0.0 0.0 2.0 14.0 23.0 adult sepsis vs adult HC:7/9; paediatric sepsis vs paediatric HC:8/8; all sepsis vs all controls:8/8 NaN NaN NaN NaN NaN NaN
|
| 66 |
+
scratch_n5__raw 25 17.0 3.0 4.0 1.0 3.0 4.0 10.0 19.0 adult sepsis vs adult HC:6/9; paediatric sepsis vs paediatric HC:7/8; all sepsis vs all controls:4/8 NaN NaN NaN NaN NaN NaN
|
| 67 |
+
real_lora__raw 25 20.0 0.0 5.0 0.0 1.0 4.0 13.0 20.0 adult sepsis vs adult HC:7/9; paediatric sepsis vs paediatric HC:7/8; all sepsis vs all controls:6/8 NaN NaN NaN NaN NaN NaN
|
| 68 |
+
real_full__raw 25 20.0 0.0 5.0 0.0 2.0 3.0 14.0 21.0 adult sepsis vs adult HC:7/9; paediatric sepsis vs paediatric HC:6/8; all sepsis vs all controls:7/8 NaN NaN NaN NaN NaN NaN
|
| 69 |
+
scratch_full__raw 25 19.0 0.0 6.0 0.0 3.0 3.0 14.0 20.0 adult sepsis vs adult HC:8/9; paediatric sepsis vs paediatric HC:6/8; all sepsis vs all controls:5/8 NaN NaN NaN NaN NaN NaN
|
| 70 |
+
lora_recipe__raw 25 23.0 0.0 2.0 0.0 0.0 2.0 15.0 23.0 adult sepsis vs adult HC:8/9; paediatric sepsis vs paediatric HC:8/8; all sepsis vs all controls:7/8 NaN NaN NaN NaN NaN NaN
|
| 71 |
+
scratch_recipe__raw 25 16.0 0.0 9.0 0.0 2.0 7.0 10.0 19.0 adult sepsis vs adult HC:5/9; paediatric sepsis vs paediatric HC:7/8; all sepsis vs all controls:4/8 NaN NaN NaN NaN NaN NaN
|
| 72 |
+
PAIR 5 pairs: scratch vs LoRA 25 NaN NaN NaN NaN NaN NaN NaN NaN NaN 6.0 8.0 2.0 0.109375 3.0 7.0
|
| 73 |
+
PAIR 58 pairs: scratch vs LoRA 25 NaN NaN NaN NaN NaN NaN NaN NaN NaN 1.0 2.0 1.0 1.000000 -1.0 2.0
|
| 74 |
+
PAIR 58 pairs: scratch vs full FT 25 NaN NaN NaN NaN NaN NaN NaN NaN NaN 1.0 3.0 2.0 1.000000 0.0 3.0
|
| 75 |
+
PAIR 58 pairs + recipes: scratch vs LoRA 25 NaN NaN NaN NaN NaN NaN NaN NaN NaN 7.0 7.0 0.0 0.015625 4.0 8.0
|
| 76 |
+
PAIR ref: scratch 5 pairs vs zero-shot 25 NaN NaN NaN NaN NaN NaN NaN NaN NaN 4.0 7.0 3.0 0.343750 -1.0 6.0
|
| 77 |
+
PAIR ref: scratch 58 pairs vs zero-shot 25 NaN NaN NaN NaN NaN NaN NaN NaN NaN 2.0 5.0 3.0 0.726562 -2.0 3.0
|
| 78 |
+
PAIR ref: scratch 58 + recipes vs zero-shot 25 NaN NaN NaN NaN NaN NaN NaN NaN NaN 5.0 7.0 2.0 0.179688 2.0 6.0
|
| 79 |
+
[ 16.2s] observed-direction flips per left-out donor (arm lora_n5):
|
| 80 |
+
left_out group obs_direction_flips undefined
|
| 81 |
+
Abd-PS19 ped_sepsis 0 0
|
| 82 |
+
Abd-PS23 ped_sepsis 0 0
|
| 83 |
+
Abd-S123 adult_sepsis 0 0
|
| 84 |
+
Abd-S39 adult_sepsis 0 0
|
| 85 |
+
Abd-S47 adult_sepsis 0 0
|
| 86 |
+
Abd-S66 adult_sepsis 0 0
|
| 87 |
+
HC01 adult_HC 0 0
|
| 88 |
+
HC07 adult_HC 1 0
|
| 89 |
+
IC23 adult_ICU_ctrl 0 0
|
| 90 |
+
PHC03 ped_HC 0 8
|
| 91 |
+
PIC04 ped_ICU_ctrl 0 0
|
| 92 |
+
Res-PS25 ped_sepsis 0 0
|
| 93 |
+
Res-S139 adult_sepsis 1 0
|
| 94 |
+
Res-S26 adult_sepsis 0 0
|
| 95 |
+
Res-S77 adult_sepsis 0 0
|
| 96 |
+
Skin-S140 adult_sepsis 0 0
|
| 97 |
+
Uri-S147 adult_sepsis 0 0
|
| 98 |
+
[ 21.9s] wrote verify_confound_within_spearman.csv (124, 22)
|
| 99 |
+
[ 22.1s] wrote verify_confound_within_folds.csv (1003, 19)
|
| 100 |
+
[ 22.1s] done
|
paper-releases/pbisc-diffusion/updates/20261005T142225Z-rare-states-scratch-vs-pretrained/diffusion/dataset/server_1/model8/plans/rare_states_screen_20261004/out/scratch_vs_lora_20261005/verify_recompute_cross_extra_report.txt
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
=== cross-7: medians over each arm's own kept set vs over the pairs kept by both ===
|
| 2 |
+
scratch_variant comparator n_kept_scratch n_kept_comp n_kept_both median_kept_scratch median_kept_comp ratio_comp_over_scratch ratio_ci_lo ratio_ci_hi frac_boot_ratio_gt1 median_bothkept_scratch median_bothkept_comp ratio_bothkept_comp_over_scratch ratio_bothkept_ci_lo ratio_bothkept_ci_hi frac_boot_bothkept_ratio_gt1 n_boot_finite
|
| 3 |
+
own r16 14 11 10 0.042 0.137 3.292 1.224 11.763 0.988 0.059 0.138 2.317 1.077 5.391 0.980 2000
|
| 4 |
+
own r64 14 10 10 0.042 0.131 3.150 0.999 9.664 0.974 0.059 0.131 2.204 0.866 4.899 0.956 2000
|
| 5 |
+
own full20k 14 15 12 0.042 0.290 6.967 2.515 13.280 0.999 0.049 0.288 5.866 2.327 9.492 0.998 2000
|
| 6 |
+
own full108k 14 16 13 0.042 0.150 3.592 1.878 9.433 0.997 0.042 0.255 6.101 2.152 8.904 0.999 2000
|
| 7 |
+
own xcal 14 19 14 0.042 0.229 5.500 3.262 13.398 1.000 0.042 0.401 9.645 3.328 12.086 1.000 2000
|
| 8 |
+
own zs 14 15 13 0.042 0.092 2.206 0.892 5.832 0.958 0.041 0.092 2.216 0.976 5.675 0.966 2000
|
| 9 |
+
own own58 14 19 14 0.042 0.344 8.265 3.153 13.745 1.000 0.042 0.307 7.379 3.021 12.716 1.000 2000
|
| 10 |
+
depth_r16 r16 14 11 10 0.067 0.137 2.042 0.760 7.114 0.926 0.107 0.138 1.286 0.661 3.225 0.834 2000
|
| 11 |
+
depth_r16 r64 14 10 10 0.067 0.131 1.954 0.584 6.220 0.814 0.107 0.131 1.224 0.519 3.039 0.692 2000
|
| 12 |
+
depth_r16 full20k 14 15 12 0.067 0.290 4.322 1.521 8.355 0.993 0.083 0.288 3.466 1.308 5.572 0.994 2000
|
| 13 |
+
depth_r16 full108k 14 16 13 0.067 0.150 2.229 1.145 5.882 0.984 0.072 0.255 3.553 1.182 5.190 0.990 2000
|
| 14 |
+
depth_r16 xcal 14 19 14 0.067 0.229 3.412 2.000 8.928 0.997 0.067 0.401 5.984 2.102 8.202 1.000 2000
|
| 15 |
+
depth_r16 zs 14 15 13 0.067 0.092 1.368 0.552 3.678 0.688 0.062 0.092 1.472 0.627 3.778 0.716 2000
|
| 16 |
+
depth_r16 own58 14 19 14 0.067 0.344 5.128 1.837 9.303 0.995 0.067 0.307 4.578 1.780 8.581 0.997 2000
|
| 17 |
+
pair s162k_frac r16_frac r64_frac full108k_frac full20k_frac xcal_frac own58_frac s162k_at_r16depth_frac
|
| 18 |
+
C1Q-high monocyte (reference)|adult 0.041 -0.042 0.001 -0.043 -0.024 0.035 0.312 0.046
|
| 19 |
+
C1Q-high monocyte (reference)|all 0.012 -0.013 -0.011 0.021 -0.004 0.019 0.302 0.013
|
| 20 |
+
C1Q-high monocyte (reference)|paediatric -0.014 -0.006 -0.035 0.106 -0.016 0.015 0.344 -0.016
|
| 21 |
+
Cycling monocyte (Ye 2026)|adult 0.057 0.139 0.179 0.279 0.341 0.417 0.298 0.120
|
| 22 |
+
Cycling monocyte (Ye 2026)|all 0.034 0.105 0.069 0.160 0.193 0.229 0.160 0.062
|
| 23 |
+
Cycling monocyte (Ye 2026)|paediatric 0.017 0.081 -0.003 0.054 0.075 0.116 0.142 0.029
|
| 24 |
+
Cytotoxic CD4 T|adult -1.756 -0.723 -0.494 -1.935 0.383 1.255 2.481 -2.095
|
| 25 |
+
Cytotoxic CD4 T|all -1.190 -0.815 -1.069 -2.021 -0.448 0.048 0.982 -1.350
|
| 26 |
+
Granulocyte-like (LDG)|adult 0.021 -0.026 -0.000 0.005 0.014 0.103 0.038 0.027
|
| 27 |
+
Granulocyte-like (LDG)|all 0.024 -0.019 -0.022 0.001 0.003 0.147 0.181 0.028
|
| 28 |
+
Granulocyte-like (LDG)|paediatric -0.006 0.005 -0.036 0.010 0.005 0.083 0.136 -0.012
|
| 29 |
+
ISG-high T|adult 0.670 2.065 2.405 2.342 1.651 2.080 0.265 1.018
|
| 30 |
+
ISG-high T|all -0.243 -0.153 -0.034 0.139 0.290 0.101 0.559 -0.224
|
| 31 |
+
ISG-high T|paediatric 0.198 0.568 0.596 0.624 0.737 0.820 0.496 0.220
|
| 32 |
+
MS1 monocyte (IL1R2+)|adult 0.062 0.099 0.084 0.255 0.243 0.505 0.621 0.095
|
| 33 |
+
MS1 monocyte (IL1R2+)|all 0.034 0.137 0.083 0.330 0.334 0.591 0.826 0.051
|
| 34 |
+
MS1 monocyte (IL1R2+)|paediatric 0.042 0.070 0.050 0.120 0.170 0.386 0.666 0.072
|
| 35 |
+
Plasmablast (label)|adult 0.728 1.762 1.122 2.940 2.670 1.219 0.842 0.728
|
| 36 |
+
Plasmablast (markers)|adult 0.412 1.290 1.014 2.294 2.162 0.849 0.544 0.559
|
| 37 |
+
Plasmablast (markers)|paediatric -1.667 -2.162 -2.353 -4.263 -3.206 -1.284 -1.798 -1.173
|
| 38 |
+
Proliferating T (label)|all -0.286 -1.081 -0.767 -1.890 -1.466 -1.184 -0.656 -0.286
|
| 39 |
+
|
| 40 |
+
=== cross-4: depth-correction factor ranges ===
|
| 41 |
+
reference depth_r16 depth_r64 depth_full20k depth_full108k is_label
|
| 42 |
+
state
|
| 43 |
+
C1Q-high monocyte (reference) 1.158 1.171 1.201 1.179 False
|
| 44 |
+
Cycling T/NK 1.317 1.423 1.311 1.363 False
|
| 45 |
+
Cycling activated T (CD38/HLA-DR) 1.444 1.571 1.421 1.488 False
|
| 46 |
+
Cycling monocyte (Ye 2026) 1.818 1.874 2.042 1.919 False
|
| 47 |
+
Cytotoxic CD4 T 1.141 1.215 1.085 1.144 False
|
| 48 |
+
DN2 / ABC B 1.496 1.743 1.586 1.675 False
|
| 49 |
+
Granulocyte-like (LDG) 1.803 2.089 1.927 1.982 False
|
| 50 |
+
HSPC 1.121 1.163 1.142 1.144 False
|
| 51 |
+
ISG-high T 1.671 1.904 1.632 1.772 False
|
| 52 |
+
ISG-high monocyte (positive control) 2.014 2.117 2.319 2.153 False
|
| 53 |
+
MS1 monocyte (IL1R2+) 1.564 1.617 1.736 1.640 False
|
| 54 |
+
Plasmablast (label) 1.000 1.000 1.000 1.000 True
|
| 55 |
+
Plasmablast (markers) 1.566 1.707 1.639 1.657 False
|
| 56 |
+
Platelet signal in monocytes (negative control) 2.114 2.226 2.460 2.240 False
|
| 57 |
+
Proliferating T (label) 1.000 1.000 1.000 1.000 True
|
| 58 |
+
rare marker states: factor range 1.085 - 2.089 | r16 only: 1.121 - 1.818
|
| 59 |
+
controls: factor range 2.014 - 2.46
|