An optimized prediction framework to assess the functional impact of pharmacogenetic variants.

ABSTRACT: Prediction of phenotypic consequences of mutations constitutes an important aspect of precision medicine. Current computational tools mostly rely on evolutionary conservation and have been calibrated on variants associated with disease, which poses conceptual problems for assessment of variants in poorly conserved pharmacogenes. Here, we evaluated the performance of 18 current functionality prediction methods leveraging experimental high-quality activity data from 337 variants in genes involved in drug metabolism and transport and found that these models only achieved probabilities of 0.1-50.6% to make informed conclusions. We therefore developed a functionality prediction framework optimized for pharmacogenetic assessments that significantly outperformed current algorithms. Our model achi

SUBMITTER: Zhou Y 

PROVIDER: S-EPMC6462826 | biostudies-literature | 2019 Apr

REPOSITORIES: biostudies-literature