File S1 - CanDrA: Cancer-Specific Driver Missense Mutation Annotation with Optimized Features

Supplementary Tables and Figures. Figure S1. Definition of CanDrA score, category and significance. Figure S2. Evaluation of single descriptors by ROC AUC on GBM.S1 and GBM.S2. Shown are descriptors with |AUC-0.5|>0.1 and P value (Bonferroni corrected) <0.05. Figure S3. Evaluation of single descriptors by ROC AUC on OVC.S1 and OVC.S2. Shown are descriptors with |AUC-0.5|>0.1 and P value (Bonferroni corrected) <0.05. Table S1. Missense mutations in set GBM.S1. Table S2. Missense mutations in set GBM.S2. Table S3. Missense mutations in set OVC.S1. Table S4. Missense mutations in set OVC.S2. Table S5. Ninety-five (95) features involved in CanDrA modeling. Table S6. Feature combination selected for GBM in CanDrA. Table S7. Feature combination selected for OVC in CanDrA. Table S8. GBM missense mutations for testing correlation between CanDrA scores and mutation prevalence. Table S9. OVC missense mutations for testing correlation between CanDrA scores and mutation prevalence. Table S10. Rare missense mutations from GBM and OVC related genes. Table S11. Comparison of algorithms for predicting rare driver mutations. Table S12. Missense mutations used for evaluating CanDrA and CHASM in predicting cancer-type-specific drivers.
(XLS)