Building and Testing Expression Models

When analyzing complex traits among populations, statistical methods such as genome-wide association studies (GWAS) are used to find associations with complex traits. Despite the growing number of published GWAS, non-European only comprise 19% of all studied individuals. Our project will utilize TOPMed (Trans-Omics for Precision Medicine) data, a set of diverse populations, to build multiple transcriptome predictive models. We will test the predictive performance of the models, then we will apply these models to PrediXcan to analyze complex traits in diverse populations