Additional file 6: Table S6. of Machine learning classifier for identification of damaging missense mutations exclusive to human mitochondrial DNA-encoded polypeptides

Training set after oversampling with SMOTE. 1666 mutations are synthetic samples. Each mutation is described by its phenotype of the mutation (Phenotype), numerical scores for the three features (Feature_1, Feature_2, Feature_3), gene (GENE), amino acid position within the polypeptide (AA position), wild type amino acid (WT AA), mutant amino acid (M AA) and domain (Domain). (XLS 427 kb)