|
Abstract:
|
Genome-wide association studies (GWAS) have identified thousands of genetic loci associated with hundreds of complex traits. However, the biological mechanisms underlying these associations are often poorly understood, particularly for non-coding associations. Large regulatory genomics datasets (e.g., from ENCODE/Roadmap and GTEx) present new opportunities to gain insight into complex trait etiology, as well as new computational and statistical challenges. We present a novel Bayesian approach to identify molecular traits (e.g., gene expression, protein functionality) potentially underlying genetic associations with complex traits by integrating GWAS and functional genomics data. We use an approximate E-M algorithm for efficient empirical Bayes estimation. We discuss an application to lipid and diabetes-related traits using publicly available GWAS association statistics.
|
