The Complementarity Between Protein-Specific and General Pathogenicity Predictors for Amino Acid Substitutions.

摘要: The usage of next-generation sequencing with biomedical/clinical purposes has fuelled the demand for tools that assess functional impact sequence variants. For single amino acid variants, general methods (GM), based on biophysics/evolutionary principles and trained by pooling variants from many proteins, are already available. Until now, their accuracy range (∼80%) limited in clinical applications. In parallel, a series studies indicate protein-specific predictors (PSP), using only information protein interest, could frequently surpass performance GM. However, two reasons suggest this may not always be case: existence threshold affecting both GM PSP, effect training data scarcity. Here, we characterize relationship between approaches deriving 82 PSP comparing them several (PolyPhen-2, SIFT, PON-P2, MutationTaster2, CADD). We find complementary GM, no approach outperforming other. varies limiting situations, example, outperformed best GM; however, opposite happens when compare SIFT. Finally, explore how observed complementarity lead to increased success rates pathogenicity prediction.