Skip to main content

Table 2 Comparison of cancer driver variant predictors on the EXPDRV+NEUT and EXPDRV+ALL benchmark data sets

From: Current cancer driver variant predictors learn to recognize driver genes instead of functional variants

  EXPDRV+ALL   EXPDRV+NEUT   EXPDRV+DEL  
Method AUC AUPRC AUC AUPRC AUC AUPRC
ParsSNP 86.9 30.0 90.1 61.9 82.6 33.9
CanDrA v+ 87.1 47.2 87.2 57.6 86.9 58.8
Chasm 3.1 91.1 54.5 94.0 73.7 87.3 58.8
CHASMplus 94.3 56.9 96.6 80.5 91.1 60.4
FATHMM cancer 91.0 39.1 93.2 67.3 87.7 43.6
TransFIC 68.0 6.0 80.6 22.0 51.0 7.9
Condel 64.3 4.5 83.9 24.3 38.5 5.6
DEOGEN2 68.8 4.6 91.3 43.7 38.9 5.2
CADD 73.7 6.4 88.5 26.5 54.0 7.9
M-CAP 54.0 4.9 83.1 37.9 38.9 5.4
MetaSVM 66.8 4.5 89.4 34.0 36.8 5.2
Dummy 98.1 46.6 98.7 67.9 97.4 59.7