Jerrychen229 commited on
Commit
e8ce6fd
·
verified ·
1 Parent(s): 9ee7b5c

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +2 -0
  2. tasks/caicedo22-variant-impact/data/expression_profiles.parquet +3 -0
  3. tasks/caicedo22-variant-impact/data/morphology_profiles.parquet +3 -0
  4. tasks/glunk23-genotype/data/README.md +97 -0
  5. tasks/glunk23-genotype/data/genotype_profiles.parquet +3 -0
  6. tasks/glunk23-genotype/data/genotype_samples.csv +437 -0
  7. tasks/glunk23-genotype/data/knockdown_profiles.parquet +3 -0
  8. tasks/glunk23-genotype/evaluator/tolerance.json +109 -0
  9. tasks/j1mm-crossmodal/data/expression_metadata.csv +0 -0
  10. tasks/j1mm-crossmodal/data/expression_signatures.parquet +3 -0
  11. tasks/j1mm-crossmodal/data/morphology/plate_metadata.csv +17 -0
  12. tasks/j1mm-crossmodal/data/morphology/profiles/BR00116991.csv.gz +3 -0
  13. tasks/j1mm-crossmodal/data/morphology/profiles/BR00116992.csv.gz +3 -0
  14. tasks/j1mm-crossmodal/data/morphology/profiles/BR00116993.csv.gz +3 -0
  15. tasks/j1mm-crossmodal/data/morphology/profiles/BR00116994.csv.gz +3 -0
  16. tasks/j1mm-crossmodal/data/morphology/profiles/BR00116995.csv.gz +3 -0
  17. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117010.csv.gz +3 -0
  18. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117011.csv.gz +3 -0
  19. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117012.csv.gz +3 -0
  20. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117013.csv.gz +3 -0
  21. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117015.csv.gz +3 -0
  22. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117016.csv.gz +3 -0
  23. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117017.csv.gz +3 -0
  24. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117019.csv.gz +3 -0
  25. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117024.csv.gz +3 -0
  26. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117025.csv.gz +3 -0
  27. tasks/j1mm-crossmodal/data/morphology/profiles/BR00117026.csv.gz +3 -0
  28. tasks/kelley23-bortezomib/data/README.md +57 -0
  29. tasks/kelley23-bortezomib/data/clones.csv +91 -0
  30. tasks/kelley23-bortezomib/data/profiles.parquet +3 -0
  31. tasks/kelley23-bortezomib/data/site_counts.csv +0 -0
  32. tasks/kelley23-bortezomib/evaluator/score.py +98 -0
  33. tasks/kelley23-bortezomib/evaluator/threshold.json +13 -0
  34. tasks/lacoste24-mislocalisation/data/MANIFEST.json +30 -0
  35. tasks/lacoste24-mislocalisation/data/README.md +32 -0
  36. tasks/lacoste24-mislocalisation/data/constructs.csv +0 -0
  37. tasks/lacoste24-mislocalisation/data/morphology_profiles.parquet +3 -0
  38. tasks/lacoste24-mislocalisation/data/task.txt +1 -0
  39. tasks/lacoste24-mislocalisation/evaluator/score.py +84 -0
  40. tasks/lazar24-proxbias/data/crispr_profiles.parquet +3 -0
  41. tasks/rohban17-gene-pairs/data/expression_profiles.parquet +3 -0
  42. tasks/rohban17-gene-pairs/data/morphology_profiles.parquet +3 -0
  43. tasks/rohban17-gene-pairs/evaluator/accepted_endpoints/reference_05.csv +16 -0
  44. tasks/rohban22-mapk14/data/compound_profiles.parquet +3 -0
  45. tasks/rohban22-mapk14/data/compound_samples.csv +3 -0
  46. tasks/rohban22-mapk14/data/compounds.csv +0 -0
  47. tasks/rohban22-mapk14/data/gene_profiles.parquet +3 -0
  48. tasks/rohban22-mapk14/evaluator/GT/CP/tolerance.json +62 -0
  49. tasks/rohban22-mapk14/evaluator/GT/EP/ground_truth.csv +32 -0
  50. tasks/rohban22-mapk14/evaluator/evaluator/reference_endpoint.csv +31 -0
.gitattributes CHANGED
@@ -58,3 +58,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
 
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
+ tasks/tegtmeyer25-22q11/data/astrocyte_expression.h5ad filter=lfs diff=lfs merge=lfs -text
62
+ tasks/rohban22-mapk14/data/compound_samples.csv filter=lfs diff=lfs merge=lfs -text
tasks/caicedo22-variant-impact/data/expression_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b689413d1955e540bd59bc4bb4d62fd44e0f02af7501fe96c1b09589d8ad495
3
+ size 15296393
tasks/caicedo22-variant-impact/data/morphology_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a092602e8b72b6348de9e7dc92292396aa79f652b37ef26c033e9a4ce7869b7
3
+ size 97413585
tasks/glunk23-genotype/data/README.md ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Adipocyte differentiation morphology workspace
2
+
3
+ Two high-content imaging experiments on primary human adipose-derived
4
+ mesenchymal stem cells (AMSCs) undergoing adipogenic differentiation. The assay
5
+ is a four-channel CellProfiler readout -- DNA, AGP (actin/Golgi/plasma
6
+ membrane), Mito, and BODIPY (neutral lipid) -- reported as per-well feature
7
+ vectors. There are no images.
8
+
9
+ The two experiments use different donor cohorts and different perturbations.
10
+ They are not paired at the sample level and share only their feature space.
11
+
12
+ ## Files
13
+
14
+ - `genotype_profiles.parquet`: 436 wells x 1,183 morphology features, keyed by
15
+ `genotype_sample_id`. Primary AMSCs from donors genotyped at one biallelic
16
+ locus.
17
+ - `genotype_samples.csv`: metadata aligned one-to-one with the profiles.
18
+ Columns: `genotype_sample_id`, `donor_id` (40 donors),
19
+ `genotype_class` (`G1`/`G2`/`G3` -- three classes at one locus, allele
20
+ identities masked), `sex`, `age`, `bmi`, `depot` (`sc` subcutaneous, `vc`
21
+ visceral), `well_position`, `ffa` (0 = control medium, 1 = free-fatty-acid
22
+ challenge), `batch`, `day` (0/3/8/14 of differentiation), `plate_id`.
23
+ - `knockdown_profiles.parquet`: 48 wells x 1,154 morphology features, keyed by
24
+ `knockdown_sample_id`. AMSCs treated with one of two siRNAs.
25
+ - `knockdown_samples.csv`: metadata aligned one-to-one with the profiles.
26
+ Columns: `knockdown_sample_id`, `plate_id`, `well_id`,
27
+ `assay_plate_barcode`, `plate_map`, `well_position`, `donor_id` (two donors),
28
+ `sirna` (`siRNA_1` = non-targeting control, `siRNA_2` = targeting one gene;
29
+ the target is masked), `si_day`, `si_day_plusminus`, `day` (0/3/9/14).
30
+ - `features.csv`: the feature index, with `in_genotype` and `in_knockdown`
31
+ flags. 1,154 of 1,183 feature names occur in both experiments.
32
+
33
+ ## Row semantics and cardinalities
34
+
35
+ One row is one imaged well.
36
+
37
+ - Genotype experiment: one row per `donor_id` x `depot` x `day` x `ffa`.
38
+ Coverage is structurally incomplete -- donors contribute between 5 and 16
39
+ rows, and depot x day strata are not balanced. Class sizes across all rows
40
+ are G1 104, G2 222, G3 110; donors recur across strata, so a class's row
41
+ count is not its donor count.
42
+ - Knockdown experiment: a balanced two donors x 4 days x 2 siRNAs x 3 wells.
43
+ The three wells of a (donor, day, siRNA) cell are wells of the same donor.
44
+
45
+ ## Joining the two experiments
46
+
47
+ The only join is by feature name. `donor_id` values in the two files come from
48
+ different cohorts and are not comparable. There is no row-level correspondence
49
+ of any kind between the two experiments.
50
+
51
+ ## Missing values
52
+
53
+ 754 missing feature values in the genotype profiles are structural -- those
54
+ features were not measured on some plates. 14 rows have no `bmi`. Both are
55
+ encoded as empty fields / NaN and are left as they are in the source.
56
+
57
+ ## What has and has not been applied
58
+
59
+ Feature values were normalised per feature by median/IQR against reference
60
+ control wells by the data producers, upstream of this workspace, and that step
61
+ is not reversible from these files. Nothing else has been applied here: no
62
+ aggregation of wells or donors, no feature filtering, no scaling, no batch
63
+ correction, no testing. Feature values are the source values unchanged.
64
+
65
+ Feature names are the original CellProfiler names and carry real compartment
66
+ and channel semantics. Group labels (`genotype_class`, `sirna`) are masked
67
+ tokens; the mapping to real alleles and to the targeted gene is evaluator-side
68
+ and deliberately not recoverable from this workspace.
69
+
70
+ ## Reporting
71
+
72
+ `effect` is signed. The class tokens carry no meaning outside this workspace, so
73
+ state in `report.md` which group the sign is relative to.
74
+
75
+ ## Analysis contract
76
+
77
+ - **Work only from the data in `/app`.** Do not retrieve external information:
78
+ no literature search, no code repositories, no reference or annotation
79
+ databases, no published result tables. This holds even if you recognise the
80
+ study the data came from -- in that case, still do not look it up. The point
81
+ of the task is what the shipped data supports, and an external label used to
82
+ choose among pipeline settings has selected that pipeline on the answer.
83
+ - Preprocessing and thresholds must be query-independent -- decided from
84
+ controls, replicate structure, plate layout and QC fields, never from how they
85
+ affect a particular feature's result.
86
+
87
+ ## Environment
88
+
89
+ The scientific stack (`numpy`, `pandas`, `pyarrow`, `scipy`, `statsmodels`)
90
+ lives in one interpreter. Call it by absolute path:
91
+
92
+ ```
93
+ /fs/scratch/PCON0023/jiayuan/ccc_subgraph/envs/ccc-pilot/bin/python <script.py>
94
+ ```
95
+
96
+ The bare `python` first on PATH may be a minimal shim without those packages,
97
+ and a login shell (`bash -l`) resets PATH, so neither is reliable here.
tasks/glunk23-genotype/data/genotype_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:238fda1269708249ec52fa6418cc86a9144b4c0d36431d8153c0920047fe28a5
3
+ size 5559944
tasks/glunk23-genotype/data/genotype_samples.csv ADDED
@@ -0,0 +1,437 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ genotype_sample_id,donor_id,genotype_class,sex,age,bmi,depot,well_position,ffa,batch,day,plate_id
2
+ m113_14_sc_1,m113,G2,2,46,47.32,sc,H02,1,5,14,BR00101146
3
+ m113_0_sc_0,m113,G2,2,46,47.32,sc,H02,0,5,0,BR00101117
4
+ m113_3_sc_0,m113,G2,2,46,47.32,sc,B02,0,5,3,BR00101116
5
+ m113_3_sc_1,m113,G2,2,46,47.32,sc,F02,1,5,3,BR00101116
6
+ m113_14_sc_0,m113,G2,2,46,47.32,sc,A02,0,5,14,BR00101146
7
+ m115_14_sc_1,m115,G2,2,61,53.74,sc,G05,1,4,14,BR00101149
8
+ m115_14_sc_0,m115,G2,2,61,53.74,sc,A05,0,4,14,BR00101149
9
+ m115_3_sc_1,m115,G2,2,61,53.74,sc,F05,1,4,3,BR00101113
10
+ m115_8_sc_1,m115,G2,2,61,53.74,sc,E05,1,4,8,BR00101114
11
+ m115_0_sc_0,m115,G2,2,61,53.74,sc,E05,0,4,0,BR00101112
12
+ m115_8_sc_0,m115,G2,2,61,53.74,sc,A05,0,4,8,BR00101114
13
+ m115_3_sc_0,m115,G2,2,61,53.74,sc,B05,0,4,3,BR00101113
14
+ m116_3_sc_1,m116,G1,1,41,69.26,sc,H07,1,4,3,BR00101113
15
+ m116_14_sc_1,m116,G1,1,41,69.26,sc,F07,1,4,14,BR00101149
16
+ m116_0_sc_0,m116,G1,1,41,69.26,sc,D07,0,4,0,BR00101112
17
+ m116_8_sc_0,m116,G1,1,41,69.26,sc,C07,0,4,8,BR00101114
18
+ m116_3_sc_0,m116,G1,1,41,69.26,sc,C07,0,4,3,BR00101113
19
+ m116_8_sc_1,m116,G1,1,41,69.26,sc,F07,1,4,8,BR00101114
20
+ m116_14_sc_0,m116,G1,1,41,69.26,sc,A07,0,4,14,BR00101149
21
+ m130_3_sc_1,m130,G3,2,41,52.34,sc,F03,1,6,3,BR00101144
22
+ m130_3_sc_0,m130,G3,2,41,52.34,sc,D03,0,6,3,BR00101144
23
+ m130_14_sc_1,m130,G3,2,41,52.34,sc,G03,1,6,14,BR00101139
24
+ m130_8_sc_1,m130,G3,2,41,52.34,sc,F03,1,6,8,BR00101141
25
+ m130_0_sc_0,m130,G3,2,41,52.34,sc,E03,0,6,0,BR00101145
26
+ m130_14_sc_0,m130,G3,2,41,52.34,sc,A03,0,6,14,BR00101139
27
+ m130_8_sc_0,m130,G3,2,41,52.34,sc,D03,0,6,8,BR00101141
28
+ m161_3_sc_1,m161,G2,2,48,65.87,sc,H09,1,4,3,BR00101113
29
+ m161_3_sc_0,m161,G2,2,48,65.87,sc,B09,0,4,3,BR00101113
30
+ m161_14_sc_0,m161,G2,2,48,65.87,sc,A09,0,4,14,BR00101149
31
+ m161_0_sc_0,m161,G2,2,48,65.87,sc,D09,0,4,0,BR00101112
32
+ m161_8_sc_0,m161,G2,2,48,65.87,sc,C09,0,4,8,BR00101114
33
+ m161_14_sc_1,m161,G2,2,48,65.87,sc,G09,1,4,14,BR00101149
34
+ m161_8_sc_1,m161,G2,2,48,65.87,sc,F09,1,4,8,BR00101114
35
+ m162_14_sc_0,m162,G2,2,28,44.78,sc,H06,0,5,14,BR00101147
36
+ m162_0_sc_0,m162,G2,2,28,44.78,sc,D06,0,5,0,BR00101115
37
+ m162_14_sc_1,m162,G2,2,28,44.78,sc,H06,1,5,14,BR00101148
38
+ m162_3_sc_0,m162,G2,2,28,44.78,sc,A06,0,5,3,BR00101118
39
+ m162_3_sc_1,m162,G2,2,28,44.78,sc,G06,1,5,3,BR00101118
40
+ m164_8_sc_1,m164,G2,1,46,53.4,sc,F07,1,1,8,BR00101079
41
+ m164_14_sc_0,m164,G2,1,46,53.4,sc,D07,0,1,14,BR00101077
42
+ m164_0_sc_0,m164,G2,1,46,53.4,sc,F07,0,1,0,BR00101083
43
+ m164_3_sc_0,m164,G2,1,46,53.4,sc,C07,0,1,3,BR00101080
44
+ m164_8_sc_0,m164,G2,1,46,53.4,sc,D07,0,1,8,BR00101079
45
+ m164_3_sc_1,m164,G2,1,46,53.4,sc,E07,1,1,3,BR00101080
46
+ m164_14_sc_1,m164,G2,1,46,53.4,sc,F07,1,1,14,BR00101077
47
+ m192_0_sc_0,m192,G3,2,17,44.2,sc,D02,0,5,0,BR00101115
48
+ m192_14_sc_0,m192,G3,2,17,44.2,sc,E02,0,5,14,BR00101147
49
+ m192_3_sc_0,m192,G3,2,17,44.2,sc,B02,0,5,3,BR00101118
50
+ m192_14_sc_1,m192,G3,2,17,44.2,sc,D02,1,5,14,BR00101148
51
+ m192_3_sc_1,m192,G3,2,17,44.2,sc,F02,1,5,3,BR00101118
52
+ m194_14_sc_1,m194,G3,2,37,66.12,sc,G03,1,4,14,BR00101149
53
+ m194_3_sc_1,m194,G3,2,37,66.12,sc,H03,1,4,3,BR00101113
54
+ m194_0_sc_0,m194,G3,2,37,66.12,sc,B03,0,4,0,BR00101112
55
+ m194_8_sc_1,m194,G3,2,37,66.12,sc,F03,1,4,8,BR00101114
56
+ m194_3_sc_0,m194,G3,2,37,66.12,sc,B03,0,4,3,BR00101113
57
+ m194_8_sc_0,m194,G3,2,37,66.12,sc,B03,0,4,8,BR00101114
58
+ m194_14_sc_0,m194,G3,2,37,66.12,sc,B03,0,4,14,BR00101149
59
+ m214_8_sc_1,m214,G2,2,57,22.27,sc,E04,1,3,8,BR00000004
60
+ m214_14_sc_0,m214,G2,2,57,22.27,sc,A04,0,3,14,BR00000001
61
+ m214_8_sc_0,m214,G2,2,57,22.27,sc,B04,0,3,8,BR00000004
62
+ m214_3_sc_0,m214,G2,2,57,22.27,sc,C04,0,3,3,BR00000003
63
+ m214_0_sc_0,m214,G2,2,57,22.27,sc,D04,0,3,0,BR00000002
64
+ m214_3_sc_1,m214,G2,2,57,22.27,sc,H04,1,3,3,BR00000003
65
+ m214_14_sc_1,m214,G2,2,57,22.27,sc,G04,1,3,14,BR00000001
66
+ m227_3_sc_0,m227,G1,2,36,49.38,sc,C11,0,5,3,BR00101118
67
+ m227_14_sc_0,m227,G1,2,36,49.38,sc,C11,0,5,14,BR00101147
68
+ m227_14_sc_1,m227,G1,2,36,49.38,sc,F11,1,5,14,BR00101148
69
+ m227_0_sc_0,m227,G1,2,36,49.38,sc,F11,0,5,0,BR00101115
70
+ m227_3_sc_1,m227,G1,2,36,49.38,sc,G11,1,5,3,BR00101118
71
+ m235_3_sc_1,m235,G2,2,30,60.85,sc,F08,1,5,3,BR00101118
72
+ m235_14_sc_1,m235,G2,2,30,60.85,sc,E08,1,5,14,BR00101148
73
+ m235_0_sc_0,m235,G2,2,30,60.85,sc,G08,0,5,0,BR00101115
74
+ m235_14_sc_0,m235,G2,2,30,60.85,sc,G08,0,5,14,BR00101147
75
+ m235_3_sc_0,m235,G2,2,30,60.85,sc,A08,0,5,3,BR00101118
76
+ m239_8_sc_1,m239,G3,2,46,66.22,sc,E01,1,4,8,BR00101114
77
+ m239_0_sc_0,m239,G3,2,46,66.22,sc,A01,0,4,0,BR00101112
78
+ m239_3_sc_1,m239,G3,2,46,66.22,sc,F01,1,4,3,BR00101113
79
+ m239_14_sc_0,m239,G3,2,46,66.22,sc,B01,0,4,14,BR00101149
80
+ m239_3_sc_0,m239,G3,2,46,66.22,sc,B01,0,4,3,BR00101113
81
+ m239_14_sc_1,m239,G3,2,46,66.22,sc,H01,1,4,14,BR00101149
82
+ m239_8_sc_0,m239,G3,2,46,66.22,sc,D01,0,4,8,BR00101114
83
+ m246_8_sc_1,m246,G3,2,34,59.14,sc,F03,1,1,8,BR00101079
84
+ m246_14_sc_1,m246,G3,2,34,59.14,sc,E03,1,1,14,BR00101077
85
+ m246_3_sc_0,m246,G3,2,34,59.14,sc,B03,0,1,3,BR00101080
86
+ m246_8_sc_0,m246,G3,2,34,59.14,sc,C03,0,1,8,BR00101079
87
+ m246_3_sc_1,m246,G3,2,34,59.14,sc,E03,1,1,3,BR00101080
88
+ m246_0_sc_0,m246,G3,2,34,59.14,sc,E03,0,1,0,BR00101083
89
+ m246_14_sc_0,m246,G3,2,34,59.14,sc,D03,0,1,14,BR00101077
90
+ m252_8_sc_0,m252,G2,1,72,51.36,sc,A01,0,3,8,BR00000004
91
+ m252_3_sc_1,m252,G2,1,72,51.36,sc,H01,1,3,3,BR00000003
92
+ m252_8_sc_1,m252,G2,1,72,51.36,sc,H01,1,3,8,BR00000004
93
+ m252_0_sc_0,m252,G2,1,72,51.36,sc,D01,0,3,0,BR00000002
94
+ m252_14_sc_0,m252,G2,1,72,51.36,sc,C01,0,3,14,BR00000001
95
+ m252_14_sc_1,m252,G2,1,72,51.36,sc,G01,1,3,14,BR00000001
96
+ m252_3_sc_0,m252,G2,1,72,51.36,sc,D01,0,3,3,BR00000003
97
+ m266_14_sc_0,m266,G1,1,64,43.94,sc,D05,0,1,14,BR00101077
98
+ m266_8_sc_1,m266,G1,1,64,43.94,sc,F05,1,1,8,BR00101079
99
+ m266_0_sc_0,m266,G1,1,64,43.94,sc,E05,0,1,0,BR00101083
100
+ m266_14_sc_1,m266,G1,1,64,43.94,sc,F05,1,1,14,BR00101077
101
+ m266_3_sc_1,m266,G1,1,64,43.94,sc,F05,1,1,3,BR00101080
102
+ m266_8_sc_0,m266,G1,1,64,43.94,sc,C05,0,1,8,BR00101079
103
+ m266_3_sc_0,m266,G1,1,64,43.94,sc,B05,0,1,3,BR00101080
104
+ m269_8_sc_0,m269,G1,2,52,,sc,B05,0,6,8,BR00101141
105
+ m269_0_sc_0,m269,G1,2,52,,sc,G05,0,6,0,BR00101145
106
+ m269_3_sc_1,m269,G1,2,52,,sc,G05,1,6,3,BR00101144
107
+ m269_8_sc_1,m269,G1,2,52,,sc,F05,1,6,8,BR00101141
108
+ m269_14_sc_1,m269,G1,2,52,,sc,H05,1,6,14,BR00101139
109
+ m269_14_sc_0,m269,G1,2,52,,sc,A05,0,6,14,BR00101139
110
+ m269_3_sc_0,m269,G1,2,52,,sc,B05,0,6,3,BR00101144
111
+ m270_8_sc_0,m270,G2,2,33,47.97,sc,A06,0,3,8,BR00000004
112
+ m270_14_sc_0,m270,G2,2,33,47.97,sc,A06,0,3,14,BR00000001
113
+ m270_8_sc_1,m270,G2,2,33,47.97,sc,G06,1,3,8,BR00000004
114
+ m270_0_sc_0,m270,G2,2,33,47.97,sc,A06,0,3,0,BR00000002
115
+ m270_3_sc_1,m270,G2,2,33,47.97,sc,H06,1,3,3,BR00000003
116
+ m270_14_sc_1,m270,G2,2,33,47.97,sc,G06,1,3,14,BR00000001
117
+ m270_3_sc_0,m270,G2,2,33,47.97,sc,A06,0,3,3,BR00000003
118
+ m272_3_sc_1,m272,G1,2,25,44.08,sc,H08,1,3,3,BR00000003
119
+ m272_3_sc_0,m272,G1,2,25,44.08,sc,B08,0,3,3,BR00000003
120
+ m272_8_sc_1,m272,G1,2,25,44.08,sc,F08,1,3,8,BR00000004
121
+ m272_14_sc_1,m272,G1,2,25,44.08,sc,E08,1,3,14,BR00000001
122
+ m272_14_sc_0,m272,G1,2,25,44.08,sc,D08,0,3,14,BR00000001
123
+ m272_8_sc_0,m272,G1,2,25,44.08,sc,D08,0,3,8,BR00000004
124
+ m272_0_sc_0,m272,G1,2,25,44.08,sc,C08,0,3,0,BR00000002
125
+ m276_14_sc_0,m276,G3,1,46,49.86,sc,A10,0,3,14,BR00000001
126
+ m276_0_sc_0,m276,G3,1,46,49.86,sc,H10,0,3,0,BR00000002
127
+ m276_8_sc_0,m276,G3,1,46,49.86,sc,A10,0,3,8,BR00000004
128
+ m276_3_sc_1,m276,G3,1,46,49.86,sc,H10,1,3,3,BR00000003
129
+ m276_3_sc_0,m276,G3,1,46,49.86,sc,A10,0,3,3,BR00000003
130
+ m276_8_sc_1,m276,G3,1,46,49.86,sc,F10,1,3,8,BR00000004
131
+ m276_14_sc_1,m276,G3,1,46,49.86,sc,E10,1,3,14,BR00000001
132
+ m282_14_sc_1,m282,G1,2,44,41.27,sc,F07,1,7,14,BR00112243
133
+ m282_8_sc_1,m282,G1,2,44,41.27,sc,H07,1,7,8,BR00112244
134
+ m282_14_sc_0,m282,G1,2,44,41.27,sc,B07,0,7,14,BR00112243
135
+ m282_0_sc_1,m282,G1,2,44,41.27,sc,G07,1,7,0,BR00112226
136
+ m282_3_sc_0,m282,G1,2,44,41.27,sc,D07,0,7,3,BR00112227
137
+ m282_0_sc_0,m282,G1,2,44,41.27,sc,B07,0,7,0,BR00112226
138
+ m282_3_sc_1,m282,G1,2,44,41.27,sc,G07,1,7,3,BR00112227
139
+ m282_8_sc_0,m282,G1,2,44,41.27,sc,D07,0,7,8,BR00112244
140
+ m129_3_sc_1,m129,G1,2,40,41.15,sc,E09,1,7,3,BR00112227
141
+ m129_14_sc_1,m129,G1,2,40,41.15,sc,F09,1,7,14,BR00112243
142
+ m129_14_sc_0,m129,G1,2,40,41.15,sc,D09,0,7,14,BR00112243
143
+ m129_8_sc_1,m129,G1,2,40,41.15,sc,F09,1,7,8,BR00112244
144
+ m129_0_sc_0,m129,G1,2,40,41.15,sc,A09,0,7,0,BR00112226
145
+ m129_3_sc_0,m129,G1,2,40,41.15,sc,C09,0,7,3,BR00112227
146
+ m129_0_sc_1,m129,G1,2,40,41.15,sc,G09,1,7,0,BR00112226
147
+ m129_8_sc_0,m129,G1,2,40,41.15,sc,B09,0,7,8,BR00112244
148
+ m157_14_sc_1,m157,G3,2,35,41.66,sc,F07,1,8,14,BR00112237
149
+ m157_8_sc_0,m157,G3,2,35,41.66,sc,D07,0,8,8,BR00112235
150
+ m157_0_sc_0,m157,G3,2,35,41.66,sc,C07,0,8,0,BR00112228
151
+ m157_14_sc_0,m157,G3,2,35,41.66,sc,D07,0,8,14,BR00112237
152
+ m157_0_sc_1,m157,G3,2,35,41.66,sc,F07,1,8,0,BR00112228
153
+ m157_8_sc_1,m157,G3,2,35,41.66,sc,H07,1,8,8,BR00112235
154
+ m157_3_sc_1,m157,G3,2,35,41.66,sc,E07,1,8,3,BR00112234
155
+ m157_3_sc_0,m157,G3,2,35,41.66,sc,B07,0,8,3,BR00112234
156
+ m216_14_sc_0,m216,G2,2,53,51.2,sc,B08,0,8,14,BR00112237
157
+ m216_0_sc_0,m216,G2,2,53,51.2,sc,C08,0,8,0,BR00112228
158
+ m216_14_sc_1,m216,G2,2,53,51.2,sc,H08,1,8,14,BR00112237
159
+ m216_8_sc_1,m216,G2,2,53,51.2,sc,E08,1,8,8,BR00112235
160
+ m216_3_sc_0,m216,G2,2,53,51.2,sc,C08,0,8,3,BR00112234
161
+ m216_0_sc_1,m216,G2,2,53,51.2,sc,H08,1,8,0,BR00112228
162
+ m216_8_sc_0,m216,G2,2,53,51.2,sc,B08,0,8,8,BR00112235
163
+ m216_3_sc_1,m216,G2,2,53,51.2,sc,E08,1,8,3,BR00112234
164
+ m256_14_sc_0,m256,G2,1,58,42.05,sc,B09,0,8,14,BR00112237
165
+ m256_8_sc_0,m256,G2,1,58,42.05,sc,C09,0,8,8,BR00112235
166
+ m256_14_sc_1,m256,G2,1,58,42.05,sc,E09,1,8,14,BR00112237
167
+ m256_0_sc_1,m256,G2,1,58,42.05,sc,H09,1,8,0,BR00112228
168
+ m256_3_sc_0,m256,G2,1,58,42.05,sc,B09,0,8,3,BR00112234
169
+ m256_3_sc_1,m256,G2,1,58,42.05,sc,G09,1,8,3,BR00112234
170
+ m256_0_sc_0,m256,G2,1,58,42.05,sc,C09,0,8,0,BR00112228
171
+ m256_8_sc_1,m256,G2,1,58,42.05,sc,F09,1,8,8,BR00112235
172
+ m280_14_sc_1,m280,G1,2,45,55.06,sc,E11,1,8,14,BR00112237
173
+ m280_14_sc_0,m280,G1,2,45,55.06,sc,C11,0,8,14,BR00112237
174
+ m280_0_sc_0,m280,G1,2,45,55.06,sc,A11,0,8,0,BR00112228
175
+ m280_3_sc_1,m280,G1,2,45,55.06,sc,G11,1,8,3,BR00112234
176
+ m280_8_sc_0,m280,G1,2,45,55.06,sc,D11,0,8,8,BR00112235
177
+ m280_3_sc_0,m280,G1,2,45,55.06,sc,B11,0,8,3,BR00112234
178
+ m280_8_sc_1,m280,G1,2,45,55.06,sc,E11,1,8,8,BR00112235
179
+ m280_0_sc_1,m280,G1,2,45,55.06,sc,E11,1,8,0,BR00112228
180
+ m120_14_sc_1,m120,G2,2,27,56.83,sc,E09,1,9,14,BR00112229
181
+ m120_0_sc_1,m120,G2,2,27,56.83,sc,E09,1,9,0,BR00112231
182
+ m120_0_sc_0,m120,G2,2,27,56.83,sc,A09,0,9,0,BR00112231
183
+ m120_8_sc_1,m120,G2,2,27,56.83,sc,G09,1,9,8,BR00114012
184
+ m120_14_sc_0,m120,G2,2,27,56.83,sc,D09,0,9,14,BR00112229
185
+ m120_3_sc_1,m120,G2,2,27,56.83,sc,E09,1,9,3,BR00114011
186
+ m120_8_sc_0,m120,G2,2,27,56.83,sc,A09,0,9,8,BR00114012
187
+ m120_3_sc_0,m120,G2,2,27,56.83,sc,D09,0,9,3,BR00114011
188
+ m153_3_sc_1,m153,G2,2,23,41.98,sc,G10,1,9,3,BR00114011
189
+ m153_0_sc_1,m153,G2,2,23,41.98,sc,F10,1,9,0,BR00112231
190
+ m153_14_sc_0,m153,G2,2,23,41.98,sc,A10,0,9,14,BR00112229
191
+ m153_14_sc_1,m153,G2,2,23,41.98,sc,G10,1,9,14,BR00112229
192
+ m153_8_sc_1,m153,G2,2,23,41.98,sc,G10,1,9,8,BR00114012
193
+ m153_0_sc_0,m153,G2,2,23,41.98,sc,C10,0,9,0,BR00112231
194
+ m153_8_sc_0,m153,G2,2,23,41.98,sc,B10,0,9,8,BR00114012
195
+ m153_3_sc_0,m153,G2,2,23,41.98,sc,B10,0,9,3,BR00114011
196
+ m262_14_sc_0,m262,G2,2,28,38.77,sc,B12,0,9,14,BR00112229
197
+ m262_0_sc_1,m262,G2,2,28,38.77,sc,H12,1,9,0,BR00112231
198
+ m262_14_sc_1,m262,G2,2,28,38.77,sc,E12,1,9,14,BR00112229
199
+ m262_8_sc_1,m262,G2,2,28,38.77,sc,G12,1,9,8,BR00114012
200
+ m262_3_sc_1,m262,G2,2,28,38.77,sc,G12,1,9,3,BR00114011
201
+ m262_0_sc_0,m262,G2,2,28,38.77,sc,B12,0,9,0,BR00112231
202
+ m262_3_sc_0,m262,G2,2,28,38.77,sc,B12,0,9,3,BR00114011
203
+ m262_8_sc_0,m262,G2,2,28,38.77,sc,B12,0,9,8,BR00114012
204
+ m113_3_vc_1,m113,G2,2,46,47.32,vc,F03,1,5,3,BR00101116
205
+ m113_14_vc_1,m113,G2,2,46,47.32,vc,F03,1,5,14,BR00101146
206
+ m113_3_vc_0,m113,G2,2,46,47.32,vc,A03,0,5,3,BR00101116
207
+ m113_0_vc_0,m113,G2,2,46,47.32,vc,A03,0,5,0,BR00101117
208
+ m113_14_vc_0,m113,G2,2,46,47.32,vc,B03,0,5,14,BR00101146
209
+ m114_8_vc_1,m114,G2,1,29,83.26,vc,F10,1,6,8,BR00101141
210
+ m114_0_vc_0,m114,G2,1,29,83.26,vc,A10,0,6,0,BR00101145
211
+ m114_3_vc_1,m114,G2,1,29,83.26,vc,E10,1,6,3,BR00101144
212
+ m114_14_vc_1,m114,G2,1,29,83.26,vc,G10,1,6,14,BR00101139
213
+ m114_8_vc_0,m114,G2,1,29,83.26,vc,C10,0,6,8,BR00101141
214
+ m114_3_vc_0,m114,G2,1,29,83.26,vc,C10,0,6,3,BR00101144
215
+ m114_14_vc_0,m114,G2,1,29,83.26,vc,D10,0,6,14,BR00101139
216
+ m115_14_vc_1,m115,G2,2,61,53.74,vc,E06,1,4,14,BR00101149
217
+ m115_0_vc_0,m115,G2,2,61,53.74,vc,G06,0,4,0,BR00101112
218
+ m115_8_vc_1,m115,G2,2,61,53.74,vc,G06,1,4,8,BR00101114
219
+ m115_3_vc_1,m115,G2,2,61,53.74,vc,G06,1,4,3,BR00101113
220
+ m115_3_vc_0,m115,G2,2,61,53.74,vc,C06,0,4,3,BR00101113
221
+ m115_8_vc_0,m115,G2,2,61,53.74,vc,A06,0,4,8,BR00101114
222
+ m115_14_vc_0,m115,G2,2,61,53.74,vc,A06,0,4,14,BR00101149
223
+ m116_14_vc_0,m116,G1,1,41,69.26,vc,A08,0,4,14,BR00101149
224
+ m116_14_vc_1,m116,G1,1,41,69.26,vc,G08,1,4,14,BR00101149
225
+ m116_8_vc_1,m116,G1,1,41,69.26,vc,G08,1,4,8,BR00101114
226
+ m116_8_vc_0,m116,G1,1,41,69.26,vc,A08,0,4,8,BR00101114
227
+ m116_0_vc_0,m116,G1,1,41,69.26,vc,G08,0,4,0,BR00101112
228
+ m116_3_vc_1,m116,G1,1,41,69.26,vc,H08,1,4,3,BR00101113
229
+ m116_3_vc_0,m116,G1,1,41,69.26,vc,C08,0,4,3,BR00101113
230
+ m130_8_vc_0,m130,G3,2,41,52.34,vc,A04,0,6,8,BR00101141
231
+ m130_8_vc_1,m130,G3,2,41,52.34,vc,G04,1,6,8,BR00101141
232
+ m130_0_vc_0,m130,G3,2,41,52.34,vc,B04,0,6,0,BR00101145
233
+ m130_3_vc_0,m130,G3,2,41,52.34,vc,D04,0,6,3,BR00101144
234
+ m130_14_vc_1,m130,G3,2,41,52.34,vc,H04,1,6,14,BR00101139
235
+ m130_3_vc_1,m130,G3,2,41,52.34,vc,F04,1,6,3,BR00101144
236
+ m130_14_vc_0,m130,G3,2,41,52.34,vc,A04,0,6,14,BR00101139
237
+ m134_0_vc_0,m134,G2,2,44,51.12,vc,B01,0,5,0,BR00101117
238
+ m134_3_vc_1,m134,G2,2,44,51.12,vc,G01,1,5,3,BR00101116
239
+ m134_14_vc_1,m134,G2,2,44,51.12,vc,G01,1,5,14,BR00101146
240
+ m134_14_vc_0,m134,G2,2,44,51.12,vc,C01,0,5,14,BR00101146
241
+ m134_3_vc_0,m134,G2,2,44,51.12,vc,A01,0,5,3,BR00101116
242
+ m141_14_vc_1,m141,G2,1,47,24.62,vc,E01,1,5,14,BR00101148
243
+ m141_3_vc_1,m141,G2,1,47,24.62,vc,G01,1,5,3,BR00101118
244
+ m141_3_vc_0,m141,G2,1,47,24.62,vc,C01,0,5,3,BR00101118
245
+ m141_0_vc_0,m141,G2,1,47,24.62,vc,F01,0,5,0,BR00101115
246
+ m141_14_vc_0,m141,G2,1,47,24.62,vc,B01,0,5,14,BR00101147
247
+ m143_0_vc_0,m143,G3,2,45,40.49,vc,A07,0,6,0,BR00101145
248
+ m143_8_vc_0,m143,G3,2,45,40.49,vc,A07,0,6,8,BR00101141
249
+ m143_14_vc_0,m143,G3,2,45,40.49,vc,D07,0,6,14,BR00101139
250
+ m143_8_vc_1,m143,G3,2,45,40.49,vc,H07,1,6,8,BR00101141
251
+ m143_3_vc_1,m143,G3,2,45,40.49,vc,E07,1,6,3,BR00101144
252
+ m143_14_vc_1,m143,G3,2,45,40.49,vc,F07,1,6,14,BR00101139
253
+ m143_3_vc_0,m143,G3,2,45,40.49,vc,D07,0,6,3,BR00101144
254
+ m161_8_vc_0,m161,G2,2,48,65.87,vc,D10,0,4,8,BR00101114
255
+ m161_14_vc_0,m161,G2,2,48,65.87,vc,C10,0,4,14,BR00101149
256
+ m161_0_vc_0,m161,G2,2,48,65.87,vc,G10,0,4,0,BR00101112
257
+ m161_3_vc_1,m161,G2,2,48,65.87,vc,E10,1,4,3,BR00101113
258
+ m161_8_vc_1,m161,G2,2,48,65.87,vc,G10,1,4,8,BR00101114
259
+ m161_3_vc_0,m161,G2,2,48,65.87,vc,C10,0,4,3,BR00101113
260
+ m162_14_vc_1,m162,G2,2,28,44.78,vc,B07,1,5,14,BR00101148
261
+ m162_14_vc_0,m162,G2,2,28,44.78,vc,F07,0,5,14,BR00101147
262
+ m162_3_vc_0,m162,G2,2,28,44.78,vc,A07,0,5,3,BR00101118
263
+ m162_3_vc_1,m162,G2,2,28,44.78,vc,H07,1,5,3,BR00101118
264
+ m162_0_vc_0,m162,G2,2,28,44.78,vc,E07,0,5,0,BR00101115
265
+ m164_3_vc_0,m164,G2,1,46,53.4,vc,D08,0,1,3,BR00101080
266
+ m164_8_vc_1,m164,G2,1,46,53.4,vc,G08,1,1,8,BR00101079
267
+ m164_8_vc_0,m164,G2,1,46,53.4,vc,D08,0,1,8,BR00101079
268
+ m164_14_vc_0,m164,G2,1,46,53.4,vc,D08,0,1,14,BR00101077
269
+ m164_14_vc_1,m164,G2,1,46,53.4,vc,F08,1,1,14,BR00101077
270
+ m164_3_vc_1,m164,G2,1,46,53.4,vc,G08,1,1,3,BR00101080
271
+ m164_0_vc_0,m164,G2,1,46,53.4,vc,G08,0,1,0,BR00101083
272
+ m192_0_vc_0,m192,G3,2,17,44.2,vc,G03,0,5,0,BR00101115
273
+ m192_14_vc_0,m192,G3,2,17,44.2,vc,D03,0,5,14,BR00101147
274
+ m192_3_vc_1,m192,G3,2,17,44.2,vc,G03,1,5,3,BR00101118
275
+ m192_3_vc_0,m192,G3,2,17,44.2,vc,C03,0,5,3,BR00101118
276
+ m192_14_vc_1,m192,G3,2,17,44.2,vc,E03,1,5,14,BR00101148
277
+ m194_8_vc_0,m194,G3,2,37,66.12,vc,A04,0,4,8,BR00101114
278
+ m194_3_vc_0,m194,G3,2,37,66.12,vc,A04,0,4,3,BR00101113
279
+ m194_14_vc_0,m194,G3,2,37,66.12,vc,A04,0,4,14,BR00101149
280
+ m194_0_vc_0,m194,G3,2,37,66.12,vc,G04,0,4,0,BR00101112
281
+ m194_8_vc_1,m194,G3,2,37,66.12,vc,F04,1,4,8,BR00101114
282
+ m194_14_vc_1,m194,G3,2,37,66.12,vc,H04,1,4,14,BR00101149
283
+ m194_3_vc_1,m194,G3,2,37,66.12,vc,F04,1,4,3,BR00101113
284
+ m206_3_vc_1,m206,G2,2,48,37.64,vc,F05,1,5,3,BR00101118
285
+ m206_14_vc_1,m206,G2,2,48,37.64,vc,G05,1,5,14,BR00101148
286
+ m206_0_vc_0,m206,G2,2,48,37.64,vc,B05,0,5,0,BR00101115
287
+ m206_14_vc_0,m206,G2,2,48,37.64,vc,C05,0,5,14,BR00101147
288
+ m206_3_vc_0,m206,G2,2,48,37.64,vc,B05,0,5,3,BR00101118
289
+ m214_3_vc_1,m214,G2,2,57,22.27,vc,H05,1,3,3,BR00000003
290
+ m214_8_vc_0,m214,G2,2,57,22.27,vc,B05,0,3,8,BR00000004
291
+ m214_14_vc_0,m214,G2,2,57,22.27,vc,C05,0,3,14,BR00000001
292
+ m214_14_vc_1,m214,G2,2,57,22.27,vc,H05,1,3,14,BR00000001
293
+ m214_0_vc_0,m214,G2,2,57,22.27,vc,A05,0,3,0,BR00000002
294
+ m214_8_vc_1,m214,G2,2,57,22.27,vc,G05,1,3,8,BR00000004
295
+ m214_3_vc_0,m214,G2,2,57,22.27,vc,A05,0,3,3,BR00000003
296
+ m222_14_vc_1,m222,G3,2,37,47.4,vc,G01,1,6,14,BR00101139
297
+ m222_0_vc_0,m222,G3,2,37,47.4,vc,A01,0,6,0,BR00101145
298
+ m222_8_vc_0,m222,G3,2,37,47.4,vc,D01,0,6,8,BR00101141
299
+ m222_3_vc_0,m222,G3,2,37,47.4,vc,D01,0,6,3,BR00101144
300
+ m222_8_vc_1,m222,G3,2,37,47.4,vc,E01,1,6,8,BR00101141
301
+ m222_3_vc_1,m222,G3,2,37,47.4,vc,G01,1,6,3,BR00101144
302
+ m222_14_vc_0,m222,G3,2,37,47.4,vc,C01,0,6,14,BR00101139
303
+ m227_3_vc_1,m227,G1,2,36,49.38,vc,E12,1,5,3,BR00101118
304
+ m227_3_vc_0,m227,G1,2,36,49.38,vc,D12,0,5,3,BR00101118
305
+ m227_14_vc_1,m227,G1,2,36,49.38,vc,D12,1,5,14,BR00101148
306
+ m227_14_vc_0,m227,G1,2,36,49.38,vc,E12,0,5,14,BR00101147
307
+ m227_0_vc_0,m227,G1,2,36,49.38,vc,G12,0,5,0,BR00101115
308
+ m235_14_vc_1,m235,G2,2,30,60.85,vc,C09,1,5,14,BR00101148
309
+ m235_3_vc_0,m235,G2,2,30,60.85,vc,B09,0,5,3,BR00101118
310
+ m235_14_vc_0,m235,G2,2,30,60.85,vc,D09,0,5,14,BR00101147
311
+ m235_3_vc_1,m235,G2,2,30,60.85,vc,F09,1,5,3,BR00101118
312
+ m239_8_vc_0,m239,G3,2,46,66.22,vc,D02,0,4,8,BR00101114
313
+ m239_3_vc_1,m239,G3,2,46,66.22,vc,E02,1,4,3,BR00101113
314
+ m239_0_vc_0,m239,G3,2,46,66.22,vc,H02,0,4,0,BR00101112
315
+ m239_8_vc_1,m239,G3,2,46,66.22,vc,E02,1,4,8,BR00101114
316
+ m239_14_vc_0,m239,G3,2,46,66.22,vc,A02,0,4,14,BR00101149
317
+ m239_3_vc_0,m239,G3,2,46,66.22,vc,B02,0,4,3,BR00101113
318
+ m239_14_vc_1,m239,G3,2,46,66.22,vc,E02,1,4,14,BR00101149
319
+ m246_8_vc_0,m246,G3,2,34,59.14,vc,D04,0,1,8,BR00101079
320
+ m246_14_vc_1,m246,G3,2,34,59.14,vc,F04,1,1,14,BR00101077
321
+ m246_3_vc_0,m246,G3,2,34,59.14,vc,C04,0,1,3,BR00101080
322
+ m246_0_vc_0,m246,G3,2,34,59.14,vc,B04,0,1,0,BR00101083
323
+ m246_3_vc_1,m246,G3,2,34,59.14,vc,G04,1,1,3,BR00101080
324
+ m246_8_vc_1,m246,G3,2,34,59.14,vc,F04,1,1,8,BR00101079
325
+ m246_14_vc_0,m246,G3,2,34,59.14,vc,D04,0,1,14,BR00101077
326
+ m252_14_vc_1,m252,G2,1,72,51.36,vc,E02,1,3,14,BR00000001
327
+ m252_3_vc_0,m252,G2,1,72,51.36,vc,C02,0,3,3,BR00000003
328
+ m252_8_vc_0,m252,G2,1,72,51.36,vc,B02,0,3,8,BR00000004
329
+ m252_0_vc_0,m252,G2,1,72,51.36,vc,D02,0,3,0,BR00000002
330
+ m252_14_vc_0,m252,G2,1,72,51.36,vc,B02,0,3,14,BR00000001
331
+ m252_8_vc_1,m252,G2,1,72,51.36,vc,H02,1,3,8,BR00000004
332
+ m252_3_vc_1,m252,G2,1,72,51.36,vc,H02,1,3,3,BR00000003
333
+ m261_14_vc_0,m261,G2,2,38,38.42,vc,C02,0,1,14,BR00101077
334
+ m261_0_vc_0,m261,G2,2,38,38.42,vc,E02,0,1,0,BR00101083
335
+ m261_14_vc_1,m261,G2,2,38,38.42,vc,F02,1,1,14,BR00101077
336
+ m261_8_vc_1,m261,G2,2,38,38.42,vc,F02,1,1,8,BR00101079
337
+ m261_8_vc_0,m261,G2,2,38,38.42,vc,C02,0,1,8,BR00101079
338
+ m261_3_vc_1,m261,G2,2,38,38.42,vc,F02,1,1,3,BR00101080
339
+ m261_3_vc_0,m261,G2,2,38,38.42,vc,C02,0,1,3,BR00101080
340
+ m266_8_vc_1,m266,G1,1,64,43.94,vc,G06,1,1,8,BR00101079
341
+ m266_3_vc_1,m266,G1,1,64,43.94,vc,G06,1,1,3,BR00101080
342
+ m266_14_vc_1,m266,G1,1,64,43.94,vc,G06,1,1,14,BR00101077
343
+ m266_0_vc_0,m266,G1,1,64,43.94,vc,B06,0,1,0,BR00101083
344
+ m266_14_vc_0,m266,G1,1,64,43.94,vc,C06,0,1,14,BR00101077
345
+ m266_8_vc_0,m266,G1,1,64,43.94,vc,B06,0,1,8,BR00101079
346
+ m266_3_vc_0,m266,G1,1,64,43.94,vc,D06,0,1,3,BR00101080
347
+ m269_0_vc_0,m269,G1,2,52,,vc,B06,0,6,0,BR00101145
348
+ m269_3_vc_1,m269,G1,2,52,,vc,G06,1,6,3,BR00101144
349
+ m269_8_vc_0,m269,G1,2,52,,vc,C06,0,6,8,BR00101141
350
+ m269_8_vc_1,m269,G1,2,52,,vc,G06,1,6,8,BR00101141
351
+ m269_14_vc_1,m269,G1,2,52,,vc,H06,1,6,14,BR00101139
352
+ m269_14_vc_0,m269,G1,2,52,,vc,C06,0,6,14,BR00101139
353
+ m269_3_vc_0,m269,G1,2,52,,vc,C06,0,6,3,BR00101144
354
+ m270_8_vc_0,m270,G2,2,33,47.97,vc,A07,0,3,8,BR00000004
355
+ m270_14_vc_0,m270,G2,2,33,47.97,vc,D07,0,3,14,BR00000001
356
+ m270_3_vc_0,m270,G2,2,33,47.97,vc,D07,0,3,3,BR00000003
357
+ m270_0_vc_0,m270,G2,2,33,47.97,vc,A07,0,3,0,BR00000002
358
+ m270_14_vc_1,m270,G2,2,33,47.97,vc,G07,1,3,14,BR00000001
359
+ m270_3_vc_1,m270,G2,2,33,47.97,vc,H07,1,3,3,BR00000003
360
+ m270_8_vc_1,m270,G2,2,33,47.97,vc,E07,1,3,8,BR00000004
361
+ m272_3_vc_0,m272,G1,2,25,44.08,vc,A09,0,3,3,BR00000003
362
+ m272_14_vc_0,m272,G1,2,25,44.08,vc,D09,0,3,14,BR00000001
363
+ m272_0_vc_0,m272,G1,2,25,44.08,vc,F09,0,3,0,BR00000002
364
+ m272_3_vc_1,m272,G1,2,25,44.08,vc,E09,1,3,3,BR00000003
365
+ m272_8_vc_0,m272,G1,2,25,44.08,vc,B09,0,3,8,BR00000004
366
+ m272_8_vc_1,m272,G1,2,25,44.08,vc,F09,1,3,8,BR00000004
367
+ m272_14_vc_1,m272,G1,2,25,44.08,vc,G09,1,3,14,BR00000001
368
+ m274_0_vc_0,m274,G2,2,53,55.18,vc,F10,0,5,0,BR00101115
369
+ m274_14_vc_1,m274,G2,2,53,55.18,vc,E10,1,5,14,BR00101148
370
+ m274_14_vc_0,m274,G2,2,53,55.18,vc,D10,0,5,14,BR00101147
371
+ m274_3_vc_1,m274,G2,2,53,55.18,vc,H10,1,5,3,BR00101118
372
+ m274_3_vc_0,m274,G2,2,53,55.18,vc,D10,0,5,3,BR00101118
373
+ m276_0_vc_0,m276,G3,1,46,49.86,vc,B11,0,3,0,BR00000002
374
+ m276_14_vc_0,m276,G3,1,46,49.86,vc,B11,0,3,14,BR00000001
375
+ m276_8_vc_0,m276,G3,1,46,49.86,vc,B11,0,3,8,BR00000004
376
+ m276_8_vc_1,m276,G3,1,46,49.86,vc,E11,1,3,8,BR00000004
377
+ m276_3_vc_1,m276,G3,1,46,49.86,vc,F11,1,3,3,BR00000003
378
+ m276_3_vc_0,m276,G3,1,46,49.86,vc,A11,0,3,3,BR00000003
379
+ m276_14_vc_1,m276,G3,1,46,49.86,vc,E11,1,3,14,BR00000001
380
+ m277_8_vc_0,m277,G2,2,62,41.04,vc,D02,0,6,8,BR00101141
381
+ m277_0_vc_0,m277,G2,2,62,41.04,vc,C02,0,6,0,BR00101145
382
+ m277_14_vc_1,m277,G2,2,62,41.04,vc,E02,1,6,14,BR00101139
383
+ m277_8_vc_1,m277,G2,2,62,41.04,vc,G02,1,6,8,BR00101141
384
+ m277_3_vc_0,m277,G2,2,62,41.04,vc,D02,0,6,3,BR00101144
385
+ m277_3_vc_1,m277,G2,2,62,41.04,vc,G02,1,6,3,BR00101144
386
+ m277_14_vc_0,m277,G2,2,62,41.04,vc,C02,0,6,14,BR00101139
387
+ m278_0_vc_0,m278,G2,1,33,40.63,vc,A04,0,5,0,BR00101115
388
+ m278_14_vc_1,m278,G2,1,33,40.63,vc,H04,1,5,14,BR00101148
389
+ m278_3_vc_0,m278,G2,1,33,40.63,vc,B04,0,5,3,BR00101118
390
+ m278_14_vc_0,m278,G2,1,33,40.63,vc,F04,0,5,14,BR00101147
391
+ m278_3_vc_1,m278,G2,1,33,40.63,vc,F04,1,5,3,BR00101118
392
+ m282_3_vc_1,m282,G1,2,44,41.27,vc,H08,1,7,3,BR00112227
393
+ m282_14_vc_0,m282,G1,2,44,41.27,vc,B08,0,7,14,BR00112243
394
+ m282_8_vc_1,m282,G1,2,44,41.27,vc,H08,1,7,8,BR00112244
395
+ m282_3_vc_0,m282,G1,2,44,41.27,vc,D08,0,7,3,BR00112227
396
+ m282_8_vc_0,m282,G1,2,44,41.27,vc,A08,0,7,8,BR00112244
397
+ m282_14_vc_1,m282,G1,2,44,41.27,vc,E08,1,7,14,BR00112243
398
+ m282_0_vc_1,m282,G1,2,44,41.27,vc,E08,1,7,0,BR00112226
399
+ m282_0_vc_0,m282,G1,2,44,41.27,vc,A08,0,7,0,BR00112226
400
+ m129_14_vc_0,m129,G1,2,40,41.15,vc,B10,0,7,14,BR00112243
401
+ m129_8_vc_1,m129,G1,2,40,41.15,vc,F10,1,7,8,BR00112244
402
+ m129_3_vc_0,m129,G1,2,40,41.15,vc,A10,0,7,3,BR00112227
403
+ m129_8_vc_0,m129,G1,2,40,41.15,vc,B10,0,7,8,BR00112244
404
+ m129_14_vc_1,m129,G1,2,40,41.15,vc,E10,1,7,14,BR00112243
405
+ m129_3_vc_1,m129,G1,2,40,41.15,vc,G10,1,7,3,BR00112227
406
+ m216_14_vc_1,m216,G2,2,53,51.2,vc,E11,1,7,14,BR00112243
407
+ m216_8_vc_0,m216,G2,2,53,51.2,vc,D11,0,7,8,BR00112244
408
+ m216_0_vc_1,m216,G2,2,53,51.2,vc,G11,1,7,0,BR00112226
409
+ m216_14_vc_0,m216,G2,2,53,51.2,vc,C11,0,7,14,BR00112243
410
+ m216_8_vc_1,m216,G2,2,53,51.2,vc,G11,1,7,8,BR00112244
411
+ m216_3_vc_0,m216,G2,2,53,51.2,vc,A11,0,7,3,BR00112227
412
+ m216_0_vc_0,m216,G2,2,53,51.2,vc,A11,0,7,0,BR00112226
413
+ m216_3_vc_1,m216,G2,2,53,51.2,vc,H11,1,7,3,BR00112227
414
+ m256_8_vc_0,m256,G2,1,58,42.05,vc,B12,0,7,8,BR00112244
415
+ m256_14_vc_1,m256,G2,1,58,42.05,vc,H12,1,7,14,BR00112243
416
+ m256_0_vc_1,m256,G2,1,58,42.05,vc,H12,1,7,0,BR00112226
417
+ m256_3_vc_1,m256,G2,1,58,42.05,vc,E12,1,7,3,BR00112227
418
+ m256_14_vc_0,m256,G2,1,58,42.05,vc,A12,0,7,14,BR00112243
419
+ m256_0_vc_0,m256,G2,1,58,42.05,vc,B12,0,7,0,BR00112226
420
+ m256_8_vc_1,m256,G2,1,58,42.05,vc,E12,1,7,8,BR00112244
421
+ m256_3_vc_0,m256,G2,1,58,42.05,vc,C12,0,7,3,BR00112227
422
+ m262_3_vc_0,m262,G2,2,28,38.77,vc,D10,0,8,3,BR00112234
423
+ m262_14_vc_0,m262,G2,2,28,38.77,vc,C10,0,8,14,BR00112237
424
+ m262_8_vc_0,m262,G2,2,28,38.77,vc,D10,0,8,8,BR00112235
425
+ m262_0_vc_1,m262,G2,2,28,38.77,vc,G10,1,8,0,BR00112228
426
+ m262_8_vc_1,m262,G2,2,28,38.77,vc,F10,1,8,8,BR00112235
427
+ m262_14_vc_1,m262,G2,2,28,38.77,vc,G10,1,8,14,BR00112237
428
+ m262_3_vc_1,m262,G2,2,28,38.77,vc,G10,1,8,3,BR00112234
429
+ m262_0_vc_0,m262,G2,2,28,38.77,vc,D10,0,8,0,BR00112228
430
+ m260_14_vc_0,m260,G3,2,39,40.04,vc,B11,0,9,14,BR00112229
431
+ m260_14_vc_1,m260,G3,2,39,40.04,vc,H11,1,9,14,BR00112229
432
+ m260_8_vc_0,m260,G3,2,39,40.04,vc,A11,0,9,8,BR00114012
433
+ m260_3_vc_1,m260,G3,2,39,40.04,vc,H11,1,9,3,BR00114011
434
+ m260_0_vc_1,m260,G3,2,39,40.04,vc,F11,1,9,0,BR00112231
435
+ m260_8_vc_1,m260,G3,2,39,40.04,vc,G11,1,9,8,BR00114012
436
+ m260_3_vc_0,m260,G3,2,39,40.04,vc,A11,0,9,3,BR00114011
437
+ m260_0_vc_0,m260,G3,2,39,40.04,vc,C11,0,9,0,BR00112231
tasks/glunk23-genotype/data/knockdown_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd2541a0321789b97619aca2e426036a0316bc57f7cbb985a63b3bb68e27289d
3
+ size 1555772
tasks/glunk23-genotype/evaluator/tolerance.json ADDED
@@ -0,0 +1,109 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "1.0",
3
+ "endpoint": "EP",
4
+ "status": "final",
5
+ "submission_file": "genotype_effects.csv",
6
+ "ground_truth_file": "ground_truth.csv",
7
+ "comparison": "auroc_of_effect_magnitude_for_correctly_signed_planted_features",
8
+ "key_columns": [
9
+ "feature"
10
+ ],
11
+ "value_column": "effect",
12
+ "minimum_features_scored": 1000,
13
+ "order_matters": false,
14
+ "duplicates_allowed": false,
15
+ "minimum_score": 0.666,
16
+ "chance_baseline": 0.5,
17
+ "key": "a planted programme, put in by reordering, as a GRADIENT. 400 of the shared features -- 190 whole correlation blocks -- were reordered inside each (depot, day, medium) cell so the risk class holds the upper values, at full strength in mature subcutaneous wells and tapering to 0.6 at mid differentiation, 0.5 in the other depot and 0.3 in both, with days 0 and 3 carrying nothing. Each feature keeps its own values inside every cell, so the multiset is byte-identical there and in any union of cells; no label-free statistic on any well set moves. The knockdown experiment ships unmodified.",
18
+ "calibration": {
19
+ "shipped_instance": {
20
+ "alpha": 4.0,
21
+ "seed": 20260826,
22
+ "programme_features": 400,
23
+ "correlation_blocks": 190,
24
+ "block_threshold": 0.995,
25
+ "median_cohens_d": 2.0,
26
+ "planted_stratum": "gradient over depot x day: sc/day14 1.0, sc/day8 0.6, vc/day14 0.5, vc/day8 0.3, days 0 and 3 zero, both media equal",
27
+ "wells_touched": 229
28
+ },
29
+ "lowest_acceptable_arm": {
30
+ "arm": "acc-day8-only",
31
+ "point": 0.7359,
32
+ "lower_edge": 0.7102
33
+ },
34
+ "highest_broken_arm": {
35
+ "arm": "brk-day3-only",
36
+ "point": 0.5817,
37
+ "upper_edge": 0.6215
38
+ },
39
+ "margin_at_edges": 0.0888,
40
+ "arms": {
41
+ "acc-all-donors": 0.9589,
42
+ "acc-day8-only": 0.7359,
43
+ "acc-donor-agg": 0.974,
44
+ "acc-drop-sex": 0.8985,
45
+ "acc-mean-effect": 0.8378,
46
+ "acc-plate-center": 0.9457,
47
+ "acc-pool-all": 0.7998,
48
+ "acc-pool-days": 0.7611,
49
+ "acc-pool-depots": 0.9934,
50
+ "acc-pool-ffa": 0.9559,
51
+ "acc-vc-only": 0.9608,
52
+ "acc-welch": 0.9996,
53
+ "acc-winsorise": 0.8306,
54
+ "brk-day0-only": 0.5816,
55
+ "brk-day3-only": 0.5817,
56
+ "brk-heterozygotes": 0.4719,
57
+ "brk-shuffle": 0.4334,
58
+ "brk-variance-only": 0.5688,
59
+ "reference": 0.825
60
+ },
61
+ "note": "Rebuilt 2026-08-28 after three independent trials failed. The programme previously sat in ONE (depot x day) cell and was exactly zero elsewhere -- a step function. gpt56 (0.6243), gemini36 (0.5815) and opus48 (0.6403) all pooled the strata and modelled day, depot and medium additively, which is what competent analysts do, and pooling averages a step function to nothing. The task was testing whether the agent guessed one cell. A gradient is only DILUTED by pooling: `acc-pool-all`, the shape all three used, now scores 0.7998 and passes. Alpha was searched only over the declared ladder {2.0, 3.0, 4.0} on six seeds, three read and three held out: 2.0 separates on 0 of 6, 3.0 on 3 of 6, 4.0 on all six. The flatness guard was also fixed: it enforced its bound over the pooled contrast wells only, and passed a draw at 0.0781 whose 15-well reference cell sat at 0.1230 -- the variance shortcut then scored 0.6408 against its ~0.53 baseline. It now enforces the bound over every well set a graded arm ranks features on, and that shortcut is back to 0.5688.",
62
+ "checked_end_to_end": {
63
+ "workspace_only_reference": 0.7856,
64
+ "workspace_only_welch": 0.994,
65
+ "best_label_free_attack": 0.588,
66
+ "how": "code/69_workspace_only_reference.py reads only workspace/ files and GT/EP/score.py scores its output; code/73_shortcut_audit.py tries to reach the key without a genotype label; code/75_release_gate.py runs both plus the structural, fidelity, semantic and leakage checks."
67
+ },
68
+ "independent_estimators": {
69
+ "why": "Every arm in the grid runs through one OLS/Type-II-ANOVA function and differs only in wells and covariates, so the acceptable band had never been tested against a different code path. These are written from scratch against the shipped workspace (code/80_independent_estimators.py) and scored by GT/EP/score.py.",
70
+ "results": {
71
+ "Mann-Whitney U, reference slice": 1.0,
72
+ "Welch t, reference slice": 0.9997,
73
+ "donor-level Welch t, reference slice": 0.9997,
74
+ "Spearman vs dose, reference slice": 0.9981,
75
+ "mixed model + donor random intercept, ALL wells": 0.9845,
76
+ "Spearman vs dose, ALL wells": 0.9797,
77
+ "cluster-robust OLS, day 14 only": 0.9688,
78
+ "cluster-robust OLS, ALL wells": 0.9016
79
+ },
80
+ "weakest": 0.9016,
81
+ "note": "All eight pass, the weakest by 0.236 over the threshold. The last three are deliberately the shapes the three failed trials used. Note they score ABOVE my own grid arms (0.735-0.825): the reference arm spends 11 parameters on 13 wells and is one of the weakest reasonable analyses, so the floor and hence the threshold are conservative."
82
+ },
83
+ "edge_stability": {
84
+ "why": "The threshold is the midpoint of two bootstrapped interval edges, so it is only as trustworthy as those edges. way_21's moved by about 0.02 under a different bootstrap seed because it resamples 70 readouts; this one resamples 1,183 features.",
85
+ "bootstrap_seeds": [
86
+ 11,
87
+ 101,
88
+ 202,
89
+ 303,
90
+ 404,
91
+ 505
92
+ ],
93
+ "floor_edge_range": [
94
+ 0.7098,
95
+ 0.7114
96
+ ],
97
+ "ceiling_edge_range": [
98
+ 0.6215,
99
+ 0.6251
100
+ ],
101
+ "midpoint_range": [
102
+ 0.6658,
103
+ 0.6675
104
+ ],
105
+ "threshold_inside_under_all_seeds": true
106
+ }
107
+ },
108
+ "checkpoint": "none. There is one modality here and the spec places a checkpoint on the image side of a multimodal task. The endpoint already tests the image side end to end: the variance-ranking shortcut, which never runs the contrast, scores 0.536 against an acceptable band of 0.779 to 0.995."
109
+ }
tasks/j1mm-crossmodal/data/expression_metadata.csv ADDED
The diff for this file is too large to render. See raw diff
 
tasks/j1mm-crossmodal/data/expression_signatures.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58a52a4b5d175530a43aff1f76665b027844d68036fcda47f0e24385fafa82a5
3
+ size 139303781
tasks/j1mm-crossmodal/data/morphology/plate_metadata.csv ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ plate,cell_line,exposure_h,cell_background,seeding_density,antibiotics,staining_anomaly,n_images
2
+ BR00116991,A549,24,Parental,100,absent,WGA,27648
3
+ BR00116992,A549,24,Parental,100,absent,WGA,27640
4
+ BR00116993,A549,24,Parental,100,absent,WGA,27352
5
+ BR00116994,A549,24,Parental,100,absent,WGA,27576
6
+ BR00116995,U2OS,24,Parental,100,absent,WGA,27648
7
+ BR00117010,U2OS,48,Parental,100,absent,Mitotracker,27648
8
+ BR00117011,U2OS,48,Parental,100,absent,none,27648
9
+ BR00117012,U2OS,48,Parental,100,absent,none,27648
10
+ BR00117013,U2OS,48,Parental,100,absent,Mitotracker,27648
11
+ BR00117015,A549,48,Parental,100,absent,none,49144
12
+ BR00117016,A549,48,Parental,100,absent,none,49152
13
+ BR00117017,A549,48,Parental,100,absent,none,49144
14
+ BR00117019,A549,48,Parental,100,absent,none,49152
15
+ BR00117024,U2OS,24,Parental,100,absent,WGA,27648
16
+ BR00117025,U2OS,24,Parental,100,absent,WGA,27648
17
+ BR00117026,U2OS,24,Parental,100,absent,WGA,27648
tasks/j1mm-crossmodal/data/morphology/profiles/BR00116991.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdccfe2718513382e5306c624c21088d9e1cdee1dea46d384bcaa7d053025da9
3
+ size 7088408
tasks/j1mm-crossmodal/data/morphology/profiles/BR00116992.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26b05b8508293d4f4ec735c8aa8002035509aa1c246d32b786e27549fcef5047
3
+ size 7085894
tasks/j1mm-crossmodal/data/morphology/profiles/BR00116993.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9170596d95e83c68b37db2e11e06149aa783663852da7ca2f278cb3f5881e58e
3
+ size 7068074
tasks/j1mm-crossmodal/data/morphology/profiles/BR00116994.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aeae526e0448dd8d7d2deeb29c92c7e3deecbac87b3090400e774aea7c2b544a
3
+ size 7097115
tasks/j1mm-crossmodal/data/morphology/profiles/BR00116995.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:616ae621f7b201be1b9c0573377362d08e05d502d87a236f1ffec86a497bfea4
3
+ size 5488308
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117010.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:041b3b76db8a85d8f8eeae1d0903dde03ba6267a60ee95d1745d02db1556762a
3
+ size 7033632
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117011.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3bb111fe0d907bc18802086db30628241db6942a7e9d5740e4d475ae389c20f3
3
+ size 7034982
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117012.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:430d81aba1a1e56b9004f3fc49a039ec1fb384b911d398ac62dba173487fa71f
3
+ size 7036236
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117013.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:84252ef97cfb233301019bd2b058506e2a854d1a6788b37425efed634d94b88d
3
+ size 7034159
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117015.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:65588a3f07d451d45f2d53f6f461fc6656b9fea88f4ebed181943c694a2f309c
3
+ size 7086125
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117016.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db0caf7d355097ef480bea7d6b1f6dabc11dc4e059d227f6e30b212024318e60
3
+ size 7085469
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117017.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2492f32e3616eb1cbf1c93a2b24bc1c970fc989f4f629ed908f075d321cf8b28
3
+ size 7093525
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117019.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:330f4a3370d8296ea6bff3834f99c7887cda7cd8f80c7aef9d32110ae435b809
3
+ size 7082779
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117024.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19b44ecd85c83beef8cd79f18b77da14fabac60e88aee3f5f536c6e9ed2ef0da
3
+ size 7019983
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117025.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21d0ca2532320027e7ebd9393ab9cde90f369298f8bfd6f21c5ac7734284bf5c
3
+ size 7015711
tasks/j1mm-crossmodal/data/morphology/profiles/BR00117026.csv.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:df377a73002f21b0d9637998baa61f9955a492702b9a00f581397a8e775c06f3
3
+ size 7025011
tasks/kelley23-bortezomib/data/README.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Cell Painting profiles of clonal cell lines
2
+
3
+ Image-derived morphology profiles from a Cell Painting experiment on clonal cell
4
+ lines, every well treated with 0.1% DMSO. No images are included.
5
+
6
+ ## Files
7
+
8
+ | File | Rows | Columns | Row granularity |
9
+ |---|---:|---:|---|
10
+ | `profiles.parquet` | 480 | 517 | one well |
11
+ | `samples.csv` | 480 | 5 | one well |
12
+ | `clones.csv` | 90 | 2 | one clone |
13
+ | `site_counts.csv` | 5,759 | 5 | one imaging site |
14
+
15
+ ## Join keys
16
+
17
+ ```
18
+ profiles.well_id <-> samples.well_id one-to-one
19
+ samples.clone_id <-> clones.clone_id many-to-one
20
+ site_counts.well_id <-> samples.well_id many-to-one
21
+ ```
22
+
23
+ ## profiles.parquet
24
+
25
+ `well_id` plus 516 CellProfiler measurements. Column names follow
26
+ `<compartment>_<family>_<detail>`, where compartment is `Cells`, `Cytoplasm` or
27
+ `Nuclei`, and the stains are named `DNA`, `ER`, `RNA`, `AGP` and `Mito`. Values
28
+ are per-well summaries of the single cells in that well. No values are missing.
29
+
30
+ ## samples.csv
31
+
32
+ | Column | Meaning |
33
+ |---|---|
34
+ | `well_id` | join key |
35
+ | `clone_id` | which clone was seeded in this well |
36
+ | `plate_id` | the plate the well belongs to |
37
+ | `batch` | the imaging batch the plate belongs to |
38
+ | `treatment_time` | how long the treatment was applied before imaging |
39
+
40
+ ## clones.csv
41
+
42
+ | Column | Meaning |
43
+ |---|---|
44
+ | `clone_id` | join key |
45
+ | `resistance` | `resistant`, `sensitive`, or empty where no label is provided |
46
+
47
+ ## site_counts.csv
48
+
49
+ Each well was imaged at several sites. One row per site.
50
+
51
+ | Column | Meaning |
52
+ |---|---|
53
+ | `well_id` | join key |
54
+ | `site` | site index within the well |
55
+ | `count_cells` | objects segmented as cells in that site |
56
+ | `count_nuclei` | objects segmented as nuclei in that site |
57
+ | `count_cytoplasm` | objects segmented as cytoplasm in that site |
tasks/kelley23-bortezomib/data/clones.csv ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ clone_id,resistance
2
+ clone_01,
3
+ clone_02,
4
+ clone_03,
5
+ clone_04,
6
+ clone_05,
7
+ clone_06,
8
+ clone_07,
9
+ clone_08,
10
+ clone_09,
11
+ clone_10,
12
+ clone_11,resistant
13
+ clone_12,
14
+ clone_13,
15
+ clone_14,
16
+ clone_15,
17
+ clone_16,
18
+ clone_17,sensitive
19
+ clone_18,
20
+ clone_19,
21
+ clone_20,
22
+ clone_21,
23
+ clone_22,
24
+ clone_23,
25
+ clone_24,
26
+ clone_25,
27
+ clone_26,
28
+ clone_27,resistant
29
+ clone_28,
30
+ clone_29,
31
+ clone_30,
32
+ clone_31,
33
+ clone_32,
34
+ clone_33,
35
+ clone_34,
36
+ clone_35,
37
+ clone_36,
38
+ clone_37,
39
+ clone_38,
40
+ clone_39,
41
+ clone_40,
42
+ clone_41,
43
+ clone_42,
44
+ clone_43,
45
+ clone_44,
46
+ clone_45,
47
+ clone_46,
48
+ clone_47,
49
+ clone_48,
50
+ clone_49,
51
+ clone_50,
52
+ clone_51,
53
+ clone_52,resistant
54
+ clone_53,
55
+ clone_54,
56
+ clone_55,
57
+ clone_56,
58
+ clone_57,
59
+ clone_58,
60
+ clone_59,
61
+ clone_60,
62
+ clone_61,
63
+ clone_62,
64
+ clone_63,
65
+ clone_64,
66
+ clone_65,
67
+ clone_66,
68
+ clone_67,
69
+ clone_68,
70
+ clone_69,
71
+ clone_70,
72
+ clone_71,
73
+ clone_72,
74
+ clone_73,resistant
75
+ clone_74,
76
+ clone_75,resistant
77
+ clone_76,sensitive
78
+ clone_77,
79
+ clone_78,
80
+ clone_79,
81
+ clone_80,
82
+ clone_81,
83
+ clone_82,
84
+ clone_83,sensitive
85
+ clone_84,
86
+ clone_85,sensitive
87
+ clone_86,
88
+ clone_87,
89
+ clone_88,
90
+ clone_89,sensitive
91
+ clone_90,
tasks/kelley23-bortezomib/data/profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4021b26e56ca25f14667d48f40d42c3d3e730247250c07030a271f219e84930
3
+ size 2534159
tasks/kelley23-bortezomib/data/site_counts.csv ADDED
The diff for this file is too large to render. See raw diff
 
tasks/kelley23-bortezomib/evaluator/score.py ADDED
@@ -0,0 +1,98 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ """Score a kelley_23 submission against the endpoint threshold.
3
+
4
+ EP is the mean quantile of the dosed clones' scores within the vacuum clones'
5
+ distribution -- equivalently P(dosed score > vacuum score). Ties count a half,
6
+ which matches the calibration for any submission whose scores are continuous
7
+ (every calibrated arm was), and is the fair convention for one that is not.
8
+ """
9
+
10
+ from __future__ import annotations
11
+
12
+ import argparse
13
+ import csv
14
+ import json
15
+ import math
16
+ from pathlib import Path
17
+
18
+ HERE = Path(__file__).resolve().parent
19
+
20
+
21
+ class Invalid(Exception):
22
+ pass
23
+
24
+
25
+ def read_submission(path: Path, key: str, value: str) -> dict[str, float]:
26
+ if not path.exists():
27
+ raise Invalid(f"submission not found: {path}")
28
+ with path.open(newline="") as handle:
29
+ reader = csv.DictReader(handle)
30
+ missing = [c for c in (key, value) if c not in (reader.fieldnames or [])]
31
+ if missing:
32
+ raise Invalid(f"missing column(s): {', '.join(missing)}")
33
+ rows = list(reader)
34
+ out: dict[str, float] = {}
35
+ for n, row in enumerate(rows, start=2):
36
+ name = (row[key] or "").strip()
37
+ if not name:
38
+ raise Invalid(f"line {n}: empty {key}")
39
+ if name in out:
40
+ raise Invalid(f"line {n}: duplicate {key} {name!r}")
41
+ raw = (row[value] or "").strip()
42
+ try:
43
+ v = float(raw)
44
+ except ValueError:
45
+ raise Invalid(f"line {n}: {value} {raw!r} is not a number")
46
+ if not math.isfinite(v):
47
+ raise Invalid(f"line {n}: {value} {raw!r} is not finite")
48
+ out[name] = v
49
+ return out
50
+
51
+
52
+ def ep(vacuum: list[float], dosed: list[float]) -> float:
53
+ total = 0.0
54
+ for x in dosed:
55
+ wins = sum(1.0 if v < x else 0.5 if v == x else 0.0 for v in vacuum)
56
+ total += wins/len(vacuum)
57
+ return total/len(dosed)
58
+
59
+
60
+ def main() -> None:
61
+ parser = argparse.ArgumentParser(description=__doc__)
62
+ parser.add_argument("submission", type=Path)
63
+ args = parser.parse_args()
64
+
65
+ config = json.loads((HERE/"threshold.json").read_text())
66
+ key, value = config["key_column"], config["value_column"]
67
+ threshold = float(config["threshold"])
68
+
69
+ truth: dict[str, int] = {}
70
+ with (HERE/"ground_truth.csv").open(newline="") as handle:
71
+ for row in csv.DictReader(handle):
72
+ truth[row["clone_id"]] = int(row["truth"])
73
+
74
+ result: dict[str, object] = {"threshold": threshold, "valid_submission": False,
75
+ "passes": False, "ep": None}
76
+ try:
77
+ sub = read_submission(args.submission, key, value)
78
+ expected = set(truth)
79
+ extra = sorted(set(sub) - expected)
80
+ absent = sorted(expected - set(sub))
81
+ if absent:
82
+ raise Invalid(f"{len(absent)} unlabelled clone(s) not scored, e.g. {absent[:5]}")
83
+ if extra:
84
+ raise Invalid(f"{len(extra)} scored clone(s) are not unlabelled, e.g. {extra[:5]}")
85
+ vacuum = [sub[c] for c, t in truth.items() if t == 0]
86
+ dosed = [sub[c] for c, t in truth.items() if t == 1]
87
+ result.update(valid_submission=True, n_vacuum=len(vacuum), n_dosed=len(dosed),
88
+ distinct_scores=len(set(sub.values())))
89
+ value_ep = ep(vacuum, dosed)
90
+ result.update(ep=value_ep, passes=bool(value_ep >= threshold))
91
+ except Invalid as exc:
92
+ result["error"] = str(exc)
93
+
94
+ print(json.dumps(result, indent=2, sort_keys=True))
95
+
96
+
97
+ if __name__ == "__main__":
98
+ main()
tasks/kelley23-bortezomib/evaluator/threshold.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "endpoint": "mean quantile of dosed clone scores within the vacuum clone distribution, equivalently P(dosed score > vacuum score)",
3
+ "submission_file": "clone_predictions.csv",
4
+ "key_column": "clone_id",
5
+ "value_column": "score",
6
+ "score_direction": "higher means more likely resistant",
7
+ "threshold": 0.8447,
8
+ "chance_baseline": 0.5037,
9
+ "reference_arm": 0.9238,
10
+ "reference_bootstrap": [0.8506, 0.9806],
11
+ "highest_broken_arm": 0.7344,
12
+ "highest_broken_bootstrap": [0.6206, 0.8388]
13
+ }
tasks/lacoste24-mislocalisation/data/MANIFEST.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha": 3.0,
3
+ "endpoint": "auroc",
4
+ "files": {
5
+ "README.md": {
6
+ "bytes": 1509,
7
+ "sha256": "1da9afd71d761316c690f4f7c1b390181c00fdce8a296414943abc9e578ba95c"
8
+ },
9
+ "constructs.csv": {
10
+ "bytes": 120854,
11
+ "sha256": "9edee3c12b9ec0274c69d379d0311963000f7123fb2c09a40d90af1f6ad05f8d"
12
+ },
13
+ "morphology_profiles.parquet": {
14
+ "bytes": 139318557,
15
+ "sha256": "d7f38d82b4a0476ba610366a96584cd63092251aa0e782732c3d8e58e5815c07"
16
+ },
17
+ "morphology_samples.csv": {
18
+ "bytes": 3453329,
19
+ "sha256": "53c134b77f9a018c5f94e3b96e79d34b4ee3c75cd00d698dcc751c768dac6d67"
20
+ },
21
+ "task.txt": {
22
+ "bytes": 511,
23
+ "sha256": "70668a390751e49da5b208e3f8efd7101ac601dfbf2787ecb45de5f4b10c98da"
24
+ }
25
+ },
26
+ "gamma": 1.0,
27
+ "mad_threshold": 0.05,
28
+ "planted": true,
29
+ "threshold": 0.8169
30
+ }
tasks/lacoste24-mislocalisation/data/README.md ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Workspace
2
+
3
+ Each construct is a FLAG-tagged open reading frame — either a reference sequence or a single amino-acid variant of it — transiently transfected into cells that were then fixed, stained, imaged, and segmented.
4
+
5
+ ## Morphology
6
+
7
+ `morphology_profiles.parquet` — CellProfiler measurements keyed by `sample_id`. `morphology_samples.csv` — the same rows with their metadata. Join on `sample_id`.
8
+
9
+ | column | meaning |
10
+ |---|---|
11
+ | `Metadata_Plate`, `Metadata_Well` | plate barcode and 96-well position |
12
+ | `transfection_status` | three-level row label |
13
+ | `Gene` | the transfected open reading frame |
14
+ | `Mutation` | amino-acid change; empty for the reference construct |
15
+ | `Class` | `Wild-type` or `Mutant` |
16
+ | `n_transf`, `n_untransf`, `transf_Ratio` | cell counts and their ratio |
17
+ | `Metadata_transfection` | whether the well holds a transfection control rather than a library construct |
18
+ | other `Metadata_*` | acquisition settings |
19
+
20
+ `constructs.csv` — one row per construct, giving its `construct_id` and the number of wells it was measured in.
21
+
22
+ `MANIFEST.json` lists every file with size and SHA-256.
23
+
24
+ ## Environment
25
+
26
+ The scientific stack (`numpy`, `pandas`, `pyarrow`, `scipy`, `scikit-learn`) lives in one interpreter. Call it by absolute path:
27
+
28
+ ```
29
+ /fs/scratch/PCON0023/jiayuan/ccc_subgraph/envs/ccc-pilot/bin/python <script.py>
30
+ ```
31
+
32
+ The bare `python` first on PATH may be a minimal shim without those packages, and a login shell (`bash -l`) resets PATH, so neither is reliable here.
tasks/lacoste24-mislocalisation/data/constructs.csv ADDED
The diff for this file is too large to render. See raw diff
 
tasks/lacoste24-mislocalisation/data/morphology_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7f38d82b4a0476ba610366a96584cd63092251aa0e782732c3d8e58e5815c07
3
+ size 139318557
tasks/lacoste24-mislocalisation/data/task.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ Read README.md, then analyse the workspace data and produce predictions.csv with one row per assayed variant and two columns: construct_id, and impact_score — a number that is larger when that variant's subcellular distribution differs more from its own wild-type reference construct. Work only from the data in /app: do not retrieve external information of any kind, and if you recognise the underlying study, do not look it up. Rank the full set; do not pre-filter to the ones you are most confident about.
tasks/lacoste24-mislocalisation/evaluator/score.py ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ """EP for lacoste-24-mislocalisation.
3
+
4
+ AUROC of the submitted impact_score against the planted truth. Chance = 0.500.
5
+ Pass threshold = 0.8169, the midpoint of the gap between the lowest legitimate
6
+ arm (0.8374, mad+L2+median+dropna+all_cols) and the highest broken arm
7
+ (0.7963, pooled/P/none+spearman+morph_only) over a 1800-arm full
8
+ factorial. Every legitimate arm scores above it and every broken arm below it;
9
+ under a paired bootstrap the ordering holds in 99.6% of resamples.
10
+
11
+ Usage: score.py <predictions.csv> -> JSON on stdout
12
+ """
13
+ import json, os, sys
14
+ from pathlib import Path
15
+ import numpy as np
16
+ import pandas as pd
17
+
18
+ THRESHOLD = 0.8169
19
+ CHANCE = 0.50
20
+ MIN_SCORED = 50 # per class; below this the submission is not evaluable
21
+ TRUTH = Path(os.environ.get("PLANTED_TRUTH", "/tests/planted_truth.csv"))
22
+
23
+
24
+ def fail(reason, **extra):
25
+ print(json.dumps(dict(passes=False, valid_submission=False, ep=0.0,
26
+ chance=CHANCE, threshold=THRESHOLD, reason=reason,
27
+ **extra), indent=2, sort_keys=True))
28
+ sys.exit(0)
29
+
30
+
31
+ def auroc(scores, labels):
32
+ labels = np.asarray(labels, bool)
33
+ ranks = pd.Series(np.asarray(scores, float)).rank().to_numpy()
34
+ n_pos, n_neg = int(labels.sum()), int((~labels).sum())
35
+ return float((ranks[labels].sum() - n_pos * (n_pos + 1) / 2) / (n_pos * n_neg))
36
+
37
+
38
+ def main(path):
39
+ p = Path(path)
40
+ if not p.exists():
41
+ fail("predictions.csv not found")
42
+ try:
43
+ pred = pd.read_csv(p)
44
+ except Exception as e:
45
+ fail(f"predictions.csv unreadable: {e}")
46
+
47
+ cols = {c.strip().lower(): c for c in pred.columns}
48
+ if "construct_id" not in cols or "impact_score" not in cols:
49
+ fail("predictions.csv must have columns construct_id and impact_score",
50
+ columns_found=list(pred.columns))
51
+ pred = pred.rename(columns={cols["construct_id"]: "construct_id",
52
+ cols["impact_score"]: "impact_score"})
53
+ pred["impact_score"] = pd.to_numeric(pred["impact_score"], errors="coerce")
54
+ pred = pred.dropna(subset=["construct_id", "impact_score"])
55
+ pred = pred.drop_duplicates(subset="construct_id", keep="first")
56
+
57
+ # "null" is a default NA string for read_csv, which would silently drop every
58
+ # null row; the label is spelled true_null and NA conversion is turned off.
59
+ truth = pd.read_csv(TRUTH, keep_default_na=False, na_values=[])
60
+ m = truth.merge(pred, on="construct_id", how="left")
61
+ scored = m.dropna(subset=["impact_score"])
62
+ pos = scored.loc[scored.role == "positive", "impact_score"].to_numpy(float)
63
+ nul = scored.loc[scored.role == "true_null", "impact_score"].to_numpy(float)
64
+
65
+ if len(pos) < MIN_SCORED or len(nul) < MIN_SCORED:
66
+ fail("too few planted constructs scored; rank the full set rather than "
67
+ "a confident subset",
68
+ n_positives_scored=int(len(pos)), n_nulls_scored=int(len(nul)),
69
+ n_required_each=MIN_SCORED,
70
+ n_rows_submitted=int(len(pred)))
71
+
72
+ ep = auroc(np.r_[pos, nul],
73
+ np.r_[np.ones(len(pos), bool), np.zeros(len(nul), bool)])
74
+ print(json.dumps(dict(
75
+ passes=bool(ep >= THRESHOLD), valid_submission=True, ep=ep,
76
+ chance=CHANCE, threshold=THRESHOLD,
77
+ n_positives_scored=int(len(pos)), n_nulls_scored=int(len(nul)),
78
+ n_planted_total=int(len(truth)), n_rows_submitted=int(len(pred)),
79
+ coverage=float(len(scored) / len(truth)),
80
+ ), indent=2, sort_keys=True))
81
+
82
+
83
+ if __name__ == "__main__":
84
+ main(sys.argv[1] if len(sys.argv) > 1 else "/app/predictions.csv")
tasks/lazar24-proxbias/data/crispr_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3324c6e699b96d4b5d1d5c5ff95019d507c664cc641d9b9c0b6833502d8d5031
3
+ size 1590772517
tasks/rohban17-gene-pairs/data/expression_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a62e1fc40039aef483403ae0e06c773392a8844d82ed9219c4da28b135391062
3
+ size 181324525
tasks/rohban17-gene-pairs/data/morphology_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91efd22ff659f5c2a37f001cb1d1f00f1038fbb4b8cfad98801b847d79823116
3
+ size 51657518
tasks/rohban17-gene-pairs/evaluator/accepted_endpoints/reference_05.csv ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ anchor_gene_id,candidate_gene_id,relationship_sign,module_id
2
+ GENE_0100,GENE_0088,positive,positive
3
+ GENE_0100,GENE_0134,negative,module_01
4
+ GENE_0100,GENE_0167,negative,module_01
5
+ GENE_0100,GENE_0040,negative,module_02
6
+ GENE_0100,GENE_0068,negative,module_02
7
+ GENE_0100,GENE_0015,negative,module_03
8
+ GENE_0100,GENE_0093,negative,module_03
9
+ GENE_0100,GENE_0122,negative,module_04
10
+ GENE_0100,GENE_0174,negative,module_04
11
+ GENE_0100,GENE_0137,negative,module_05
12
+ GENE_0100,GENE_0146,negative,module_05
13
+ GENE_0100,GENE_0082,negative,module_06
14
+ GENE_0100,GENE_0163,negative,module_06
15
+ GENE_0100,GENE_0063,negative,module_07
16
+ GENE_0100,GENE_0074,negative,module_07
tasks/rohban22-mapk14/data/compound_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5d9d82f243be686b79a3ec35ebfcc4001742d1bff3457b2f3919b69e91a82b4
3
+ size 1364982270
tasks/rohban22-mapk14/data/compound_samples.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0dc0edf05bd34aabf5241ddb0e2153d0d22833f3a9721ca4757d30d4e73bc26
3
+ size 13388663
tasks/rohban22-mapk14/data/compounds.csv ADDED
The diff for this file is too large to render. See raw diff
 
tasks/rohban22-mapk14/data/gene_profiles.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a0a3742bb213f39bb3f9d78c0f5c025a3de1b575c0a936b04dc2acfb0c20caa
3
+ size 35233426
tasks/rohban22-mapk14/evaluator/GT/CP/tolerance.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "1.0",
3
+ "endpoint": "CP",
4
+ "status": "final",
5
+ "checkpoint": "processed_profile_integrity",
6
+ "submission_directory": true,
7
+ "required_files": [
8
+ "candidate_rankings.csv",
9
+ "compound_scores.csv",
10
+ "processed_compound_profiles.parquet",
11
+ "processed_control_profiles.parquet",
12
+ "processed_anchor_profile.parquet",
13
+ "analysis_config.json",
14
+ "report.md"
15
+ ],
16
+ "allowed_similarity_metrics": [
17
+ "pearson",
18
+ "spearman",
19
+ "cosine"
20
+ ],
21
+ "k_per_direction": 15,
22
+ "score_reproduction_absolute_tolerance": 1e-5,
23
+ "feature_namespace": {
24
+ "canonical_channels": [
25
+ "DNA",
26
+ "ER",
27
+ "RNA",
28
+ "AGP",
29
+ "Mito"
30
+ ],
31
+ "accepted_gene_side_aliases": {
32
+ "Hoechst": "DNA",
33
+ "ERSyto": "ER",
34
+ "ERSytoBleed": "RNA",
35
+ "PhGolgi": "AGP"
36
+ }
37
+ },
38
+ "components": {
39
+ "feature_scale": {
40
+ "statistic": "min(q90(log(feature_sd))-q10(log(feature_sd)), q90(log(feature_iqr/1.349))-q10(log(feature_iqr/1.349)))",
41
+ "bound": 0.05
42
+ },
43
+ "plate_control_center": {
44
+ "statistic": "median_feature(std_across_plates(plate_DMSO_median)/compound_feature_sd)",
45
+ "bound": 0.5
46
+ },
47
+ "feature_coverage": {
48
+ "statistic": "minimum_feature_count/observed_feature_count",
49
+ "minimum_feature_count": 1088,
50
+ "canonical_feature_count": 1450
51
+ }
52
+ },
53
+ "aggregate": "maximum_normalized_component",
54
+ "checkpoint_pass_threshold": 1.0,
55
+ "pass_requires": [
56
+ "valid_output_schema",
57
+ "scores_reproduced_from_processed_profiles",
58
+ "candidate_ranking_reproduced",
59
+ "checkpoint_pass"
60
+ ]
61
+ }
62
+
tasks/rohban22-mapk14/evaluator/GT/EP/ground_truth.csv ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ broad_sample,direction
2
+ BRD-K30737452-001-01-7,oppose
3
+ BRD-K25645002-001-07-1,oppose
4
+ BRD-A88548664-001-01-5,oppose
5
+ BRD-K00526921-001-01-3,oppose
6
+ BRD-K69389910-001-05-8,oppose
7
+ BRD-K37527074-001-01-3,oppose
8
+ BRD-K22485649-001-01-7,oppose
9
+ BRD-K69678391-001-05-2,oppose
10
+ BRD-K16925239-001-01-6,oppose
11
+ BRD-K76565902-001-01-3,oppose
12
+ BRD-K48074802-001-04-6,oppose
13
+ BRD-K61416283-001-01-5,oppose
14
+ BRD-K92398088-001-01-5,oppose
15
+ BRD-K37167897-001-01-3,oppose
16
+ BRD-A73635700-001-03-9,oppose
17
+ BRD-K45937204-001-02-5,phenocopy
18
+ BRD-K83751480-001-05-1,phenocopy
19
+ BRD-K42367391-001-01-3,phenocopy
20
+ BRD-K73463737-001-01-5,phenocopy
21
+ BRD-K94506757-001-05-2,phenocopy
22
+ BRD-K10929790-001-07-2,phenocopy
23
+ BRD-K91513229-001-01-0,phenocopy
24
+ BRD-A71838263-001-05-5,phenocopy
25
+ BRD-A42288717-001-06-9,phenocopy
26
+ BRD-K28079131-001-06-6,phenocopy
27
+ BRD-K20110031-001-01-9,phenocopy
28
+ BRD-K40892926-001-01-6,phenocopy
29
+ BRD-K59843356-001-01-6,phenocopy
30
+ BRD-K83842983-001-05-0,phenocopy
31
+ BRD-K73701301-001-01-4,phenocopy
32
+
tasks/rohban22-mapk14/evaluator/evaluator/reference_endpoint.csv ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ broad_sample,direction,reference_score,published_score
2
+ BRD-K30737452-001-01-7,oppose,-0.559734810303398,-0.3354124392300225
3
+ BRD-K25645002-001-07-1,oppose,-0.5609276690384902,-0.20096139421237502
4
+ BRD-A88548664-001-01-5,oppose,-0.5639490199841891,-0.3518352882391145
5
+ BRD-K00526921-001-01-3,oppose,-0.566040681457666,-0.30999291253961647
6
+ BRD-K69389910-001-05-8,oppose,-0.5673820098987625,-0.1598779312780427
7
+ BRD-K37527074-001-01-3,oppose,-0.5705528179537337,-0.212606489739602
8
+ BRD-K22485649-001-01-7,oppose,-0.5715037795292311,-0.2048168892432595
9
+ BRD-K69678391-001-05-2,oppose,-0.5763954822674139,-0.2782643132050925
10
+ BRD-K16925239-001-01-6,oppose,-0.5793275235932095,-0.20895192864095402
11
+ BRD-K76565902-001-01-3,oppose,-0.5856904576879421,-0.1379886432101356
12
+ BRD-K48074802-001-04-6,oppose,-0.5952443106411512,-0.204344572897538
13
+ BRD-K61416283-001-01-5,oppose,-0.5988241710119823,-0.24025463378803302
14
+ BRD-K92398088-001-01-5,oppose,-0.6173388158884527,-0.3732429044084785
15
+ BRD-K37167897-001-01-3,oppose,-0.6634642105650854,-0.3151681759748035
16
+ BRD-A73635700-001-03-9,oppose,-0.6658555282853033,-0.335722463565985
17
+ BRD-K45937204-001-02-5,phenocopy,0.6534926899971233,0.484353778483633
18
+ BRD-K83751480-001-05-1,phenocopy,0.6519195278116234,0.521990497159284
19
+ BRD-K42367391-001-01-3,phenocopy,0.6412231741990793,0.532611268900154
20
+ BRD-K73463737-001-01-5,phenocopy,0.619876933979572,0.48155263630314904
21
+ BRD-K94506757-001-05-2,phenocopy,0.6138376054322582,0.443813788657334
22
+ BRD-K10929790-001-07-2,phenocopy,0.6126355583984774,0.5132195010459305
23
+ BRD-K91513229-001-01-0,phenocopy,0.6090339122812332,0.4654999910417395
24
+ BRD-A71838263-001-05-5,phenocopy,0.6082483383545371,0.4969946254666665
25
+ BRD-A42288717-001-06-9,phenocopy,0.6052229679387332,0.3698859334158945
26
+ BRD-K28079131-001-06-6,phenocopy,0.6042279676680227,0.5309574870075675
27
+ BRD-K20110031-001-01-9,phenocopy,0.6015406289256316,0.5211204280022235
28
+ BRD-K40892926-001-01-6,phenocopy,0.6002379149157401,0.37410552347821646
29
+ BRD-K59843356-001-01-6,phenocopy,0.59893708967152,0.520692532695842
30
+ BRD-K83842983-001-05-0,phenocopy,0.5978412812378928,0.47104812955007047
31
+ BRD-K73701301-001-01-4,phenocopy,0.5978146017282406,0.45084402442029053