Dataset Viewer
Auto-converted to Parquet Duplicate
cpg
stringlengths
8
15
⌀
median
float32
0.02
0.98
⌀
mean
float32
0.04
0.96
⌀
std
float32
0.08
0.38
⌀
selected
bool
2 classes
cg14521995
0.896949
0.687968
0.384798
true
rs5931272
0.479238
0.51273
0.36053
false
rs798149
0.422891
0.444953
0.359267
false
rs1416770
0.54002
0.471776
0.345965
false
rs5987737
0.462537
0.50947
0.338127
false
cg11784298
0.719684
0.549343
0.332274
false
cg15794987
0.302276
0.500315
0.330151
false
rs2804694
0.472474
0.532696
0.327752
false
cg23489384
0.467621
0.490014
0.326226
false
cg16464924
0.232716
0.469083
0.325691
false
rs1495031
0.470691
0.531537
0.325439
false
cg14112997
0.639642
0.422162
0.325082
false
cg11738485
0.561771
0.468034
0.323243
false
cg21272897
0.137865
0.387908
0.322453
false
rs6626309
0.476856
0.490889
0.320825
false
rs7746156
0.488173
0.528889
0.320215
false
cg06193597
0.686199
0.478724
0.319862
false
cg21211688
0.828828
0.594406
0.318377
false
cg10995422
0.439493
0.467588
0.318178
false
cg24007926
0.281246
0.479973
0.317863
false
cg14323910
0.800013
0.568794
0.317358
false
rs3818562
0.467944
0.47795
0.316936
false
rs877309
0.519117
0.472726
0.315675
false
rs1019916
0.473022
0.422895
0.315412
false
cg27230769
0.512312
0.500932
0.313035
false
rs9292570
0.504914
0.546278
0.312907
false
cg05248234
0.133696
0.329385
0.312627
false
rs939290
0.536008
0.548732
0.312405
false
cg08238375
0.550969
0.461448
0.312024
false
cg06293782
0.847817
0.604382
0.311482
true
rs1510480
0.443983
0.52158
0.311132
false
cg25174111
0.806329
0.550262
0.310238
false
cg16112880
0.520496
0.425171
0.310012
false
rs10774834
0.475071
0.530438
0.309692
false
cg19311470
0.495413
0.370691
0.308982
false
cg11008123
0.387097
0.543966
0.308793
false
rs2468330
0.490465
0.420813
0.308307
false
cg23803868
0.450222
0.461307
0.307451
false
cg10528424
0.546831
0.485409
0.306792
false
cg18339359
0.788266
0.597446
0.306056
false
cg00211215
0.823007
0.632577
0.305705
true
rs1945975
0.520859
0.59801
0.304413
false
rs2521373
0.462929
0.57947
0.30428
false
cg14651435
0.519038
0.534428
0.303723
false
rs11034952
0.480383
0.435376
0.303633
false
cg03278611
0.185858
0.327824
0.303412
false
cg19697575
0.820521
0.611537
0.302651
false
cg21498547
0.540091
0.453546
0.302457
false
rs11249206
0.48487
0.45151
0.302309
false
cg08146708
0.607915
0.596785
0.302013
false
cg13573375
0.28905
0.469001
0.301696
false
cg06961873
0.54091
0.555744
0.301585
false
rs3936238
0.432232
0.394984
0.301543
false
cg11420142
0.499484
0.460787
0.301324
false
cg19949776
0.549601
0.563926
0.30087
false
cg11956442
0.579568
0.435165
0.300709
false
cg11128983
0.822105
0.589794
0.300317
false
cg08103988
0.459187
0.501522
0.300183
false
rs4331560
0.476513
0.509641
0.299806
false
rs2385226
0.411795
0.433339
0.299582
false
rs5936512
0.09307
0.288755
0.299459
false
cg16167565
0.831891
0.634649
0.299336
false
cg10422744
0.182839
0.33356
0.298717
false
cg15567368
0.489715
0.425238
0.298635
false
cg13462557
0.52344
0.46335
0.298487
false
cg07878625
0.812136
0.644863
0.298135
false
rs1484127
0.535565
0.667209
0.298098
false
cg26864826
0.522017
0.359126
0.297537
false
rs2032088
0.467327
0.354286
0.297197
false
cg09993319
0.549883
0.478903
0.297012
false
cg11400162
0.483703
0.404425
0.296777
false
rs10796216
0.46683
0.463665
0.296719
false
cg09533869
0.564411
0.421588
0.296692
false
cg15365500
0.863297
0.622695
0.296213
false
cg03407524
0.656082
0.529324
0.296201
false
cg05023192
0.573279
0.521422
0.296156
false
cg21130926
0.470615
0.489482
0.295873
false
cg08522473
0.510507
0.450354
0.295773
false
cg09214175
0.832358
0.67167
0.295724
false
cg03673787
0.412874
0.559386
0.295632
false
rs4742386
0.52161
0.534941
0.29562
false
cg01891583
0.457313
0.430184
0.295366
false
rs472920
0.561191
0.573354
0.295217
false
cg21589417
0.558366
0.479865
0.295167
false
cg24470466
0.52704
0.484513
0.295163
false
cg04922606
0.350292
0.496696
0.295119
false
cg14179288
0.528902
0.385382
0.295067
false
cg01600516
0.897416
0.696232
0.29458
false
cg12381531
0.863169
0.66365
0.294437
false
cg02856402
0.80256
0.55934
0.294435
false
cg03122926
0.306163
0.354793
0.294304
false
cg04123498
0.544434
0.615542
0.294044
false
cg01105403
0.341659
0.524476
0.293717
false
cg00143986
0.325414
0.456897
0.293356
false
cg17348244
0.303055
0.491858
0.293355
false
rs1510189
0.489161
0.451876
0.292831
false
rs133860
0.913097
0.683067
0.292772
false
cg21463262
0.454957
0.445725
0.292535
false
cg20089799
0.469148
0.388972
0.292345
false
cg11418607
0.533137
0.48938
0.292325
false
End of preview. Expand in Data Studio

Marvz model store (research prototype)

Weights and reference tables for the Marvz core module. Research use only. Not a diagnostic or screening tool.

Methylation cohorts (each trained separately against its own controls)

cohort GEO status cross-validated AUC age/sex/smoking-only AUC note
ra_wholeblood_gse42861_v1 GSE42861 usable 0.869 0.546
crohn_wholeblood_gse87648_v1 GSE87648 error ValueError: no samples match filters {'cell_type': 'wh blood'}. Example values: [{'cell_type': 'Whole blood', 'sex': 'F'
uc_wholeblood_gse87648_v1 GSE87648 error ValueError: no samples match filters {'cell_type': 'wh blood'}. Example values: [{'cell_type': 'Whole blood', 'sex': 'F'
sle_sorted_cells_gse59250_v1 GSE59250 skipped switched off in the config
t1d_daisy_gse142512_v1 GSE142512 skipped switched off in the config

Only cohorts with status usable are in this repository. The gate: cross-validated AUC of at least 0.70 and at least 0.05 above the covariates-only baseline. Reference tables hold GEO sample ids, the label and the embedding only (no age, sex or smoking).

HLA

No HLA-derived file is included: the IPD-IMGT/HLA database is CC BY-ND. The app rebuilds the HLA index from the official download at start-up (method pca, 16 dimensions); see manifest.json. IPD-IMGT/HLA release downloaded 2026-11-09 (sha256 d5da882fca64...). Please cite Barker DJ et al., Nucleic Acids Res 2026;54(D1):D1152-D1158.

Provenance and limits

GEO GSE42861 (Liu Y et al., Nat Biotechnol 2013) and GSE87648 (Ventham NT et al., Nat Commun 2016): public, de-identified data; please cite the original studies. Each model detects one disease versus controls in whole blood. Blood cell composition was not adjusted for. Cross-validated within one cohort; not validated elsewhere.

Downloads last month
67