Skip to main content
. Author manuscript; available in PMC: 2020 Sep 1.
Published in final edited form as: Hum Mutat. 2019 Aug 7;40(9):1579–1592. doi: 10.1002/humu.23826

Table 7.

Pathogenicity-prediction performance comparison (2-class evaluations) among MLR (multi-class logistic regression), CLM (cumulative logit model), and our WSR (weakly supervised regression) variants.

ClinVar Test Set CAGI Test Set
Model Binary AUC RMSD Binary AUC RMSD
MLR 0.751 ± 0.004 0.251 ± 0.005 0.720 ± 0.011 0.277 ± 0.003
CLM 0.853 ± 0.006 0.220 ± 0.004 0.854 ± 0.008 0.213 ± 0.005
WSR1 0.801 ± 0.010 0.212 ± 0.012 0.780 ± 0.005 0.234 ± 0.010
WSR2 0.971 ± 0.001 0.201 ± 0.004 0.961 ± 0.003 0.198 ± 0.004
WSR3 (Kernelized WSR2) 0.982 ± 0.003 0.131 ± 0.004 0.968 ± 0.002 0.161 ± 0.003