Stephen Rong, PhD

Topic: Moving from GWAS to Casual Genes and Variants

Project Summary: Genome-wide association studies have revealed many disease-associated variants, but identifying causal variants is challenging. This proposal aims to address this by developing deep learning models to predict regulatory effects of variants based on massively parallel reporter assay data. By exploring how these functional predictions relate to genomic signatures of natural selection across evolutionary timescales and disease categories, it will assess when evolutionary constraint can serve as an informative proxy for regulatory function in interpreting GWAS. The predictive models and constraint maps will then be applied to nominate likely causal variants from GWAS of autoimmune and metabolic diseases. This research will nominate candidate causal variants across diverse human ancestry groups, advancing understanding of the precise regulatory mechanisms disrupted in common human diseases. Overall, this proposal develops and applies integrative genomic methods to elucidate complex disease genetics.