We have been awarded a 5-year Maximizing Investigators’ Research Award (MIRA) R35 grant from NIGMS. We proposed to develop computational methods to interpret coding and noncoding variation in genetic analysis, including risk gene discovery, population screening, and clinical genetic testing. A key idea is to conceptually separate the impact of genetic variation at molecular (protein function and expression) and population (selection coefficient) levels, and jointly estimate them using probalistic graphical models and deep learing. To do that, we will use large human population genome sequence data, protein structure, and functional genomics data. New postdoc positions to work on this project are available.