Genetic interactions drive heterogeneity in causal variant effect sizes for gene expression and complex traits
Authors: Roshni Patel 1; Shaila Musharoff 1,2; Jeffrey Spence 1; Harold Pimentel 3; Catherine Tcheandjieu 1,2; Hakhamanesh Mostafavi 1; Nasa Sinnott-Armstrong 1,2; Shoa Clarke 1,2; Courtney Smith 1; Peter Durda 4; Kent Taylor 5; Russell Tracy 4; Yongmei Liu 6; Craig Johnson 7; Francois Aguet 8; Kristin Ardlie 8; Stacey Gabriel 8; Josh Smith 7; Stephen Rich 9; Jerome Rotter 5; Philip Tsao 1,2; Themistocles Assimes 1,2; Jonathan Pritchard 1; VA Million Veteran Program
Affiliations: 1) Stanford University, Stanford, CA; 2) VA Palo Alto Health Care System, Palo Alto, CA; 3) University of California Los Angeles, Los Angeles, CA; 4) University of Vermont, Burlington, VT; 5) Lundquist Institute for Biomedical Innovation, Torrance, CA; 6) Duke University, Durham, NC; 7) University of Washington, Seattle, WA; 8) Broad Institute, Cambridge, MA; 9) University of Virginia, Charlottesville, VA
Keywords: Complex traits
Despite the growing number of genome-wide association studies (GWAS), it remains unclear to what extent gene-by-gene and gene-by-environment interactions influence complex traits in humans. The magnitude of genetic interactions in complex traits has been difficult to quantify because GWAS are generally underpowered to detect individual interactions of small effect. Thus, despite the widespread use of the additive model in quantitative genetics, its applicability to human traits is yet to be determined. Here, we develop a method to test for genetic interactions that aggregates information across all trait-associated loci. Specifically, we test whether SNPs in regions of European ancestry shared between European and admixed African-American individuals have the same causal effect size. We hypothesize that, in African-Americans, the presence of genetic interactions will drive the causal effect sizes of SNPs in regions of European ancestry to be more similar to those of SNPs in regions of African ancestry. Because we focus on comparing regions of shared European ancestry in two different populations, our analysis is not biased by differences in LD structure between European and African ancestries. We apply our method to two traits: gene expression of 319 African-Americans and 499 Europeans in the Multi-Ethnic Study of Atherosclerosis (MESA) and low-density lipoprotein cholesterol (LDL-C) of 72K African-Americans and 298K Europeans in the Million Veteran Program (MVP). We find significant evidence for genetic interactions in our analysis of gene expression; for LDL-C, we observe a similar point estimate although this is not significant, likely due to lower statistical power. These results underscore the role of genetic interactions in human complex traits and highlight the limitations of the additive model.