Fig. 2

CAPICE outperforms other predictors in discriminating pathogenic variants and neutral variants. a True/false classification for all predictors tested against the full benchmark set that contains all types of variants. Top bar shows the breakdown of the test set. Other bars show the classification performance for each method. Purple blocks represent correct classification of pathogenic variants. Dark-blue blocks represent neutral variants. Pink and light-blue blocks denote false classifications. Gray blocks represent variants that were not classified by the predictor tested. Threshold selection methods are described in the “Methods” section. b Receiver operating characteristic (ROC) curves of CAPICE with AUC values for a subset of the benchmark data that only contains non-synonymous variants (the ROC curve for the full dataset can be found in Additional File 1: Fig. S2). Each ROC curve is for a subset of variants displaying a specific molecular consequence. AUC values for the different methods are listed in the figure legend