Fig. 4

Results of five benchmark datasets. A Bar graph of ROC-AUC score. B Table of ROC-AUC score. For each benchmark dataset, tools were ranked based on the ROC-AUC score and represented on the color scale: white signifies a lower rank, while shades of red, culminating in dark red, represent higher ranks. In Network&AA, we could not acquire prediction scores of variants in IARC TP53 and CGC-recurrent. In CHASMplus and CHASM, variants in CGC-recurrent did not have prediction scores