Alkes Price

Common variants implicated by genome-wide association studies (GWAS) of complex diseases are known to be enriched for coding and regulatory variants.  However, the contribution of different classes of functional variation to the heritability of complex traits is currently unknown.  Here, we investigate two approaches for answering this question that account for linkage disequilibrium between functional categories: variance components methods using raw genotype and phenotype data, and LD Score regression using summary association statistics.  We evaluate both approaches via simulation and present results on empirical data sets containing tens (or hundreds) of thousands of samples, including autoimmune disease and schizophrenia data sets.  We describe overall and cell-type specific enrichments in heritability explained for many functional categories including DNase I Hypersensitivity sites (DHS), histone marks and predicted enhancers.  Our results highlight the value of analyzing components of heritability to unravel the functional architecture of common disease.