RT Journal Article
SR Electronic
T1 Transethnic genetic correlation estimates from summary statistics support widespread non-additive effects
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 036657
DO 10.1101/036657
A1 Brielin C. Brown
A1 Asian Genetic Epidemiology Network-Type 2 Diabetes (AGEN-T2G) Consortium
A1 Chun Jimmie Ye
A1 Alkes L. Price
A1 Noah Zaitlen
YR 2016
UL http://biorxiv.org/content/early/2016/01/14/036657.abstract
AB The increasing number of genetic association studies conducted in multiple populations provides unprecedented opportunity to study how the genetic architecture of complex phenotypes varies between populations, a problem important for both medical and population genetics. Here we develop a method for estimating the transethnic genetic correlation; the correlation of causal variant effect sizes at SNPs common in populations. Unlike some prior approaches, we take advantage of the entire spectrum of SNP associations and utilize only summary-level GWAS data, thereby avoiding the computational costs and privacy concerns associated with genotype-level information while remaining scalable to hundreds of thousands of individuals and millions of SNPs. We apply our method to gene expression, rheumatoid arthritis, and type-two diabetes data and overwhelmingly find that the genetic correlation is significantly less than 1. We argue that this is evidence for the presence of non-additive or differential tagging effects that modify the marginal effect sizes at SNPs common in both populations. Our method is implemented in a python package called popcorn.