Identifying Genome-scale Interaction Effects in Human Traits and Diseases

Harvard School Of Public Health, Boston MA

Investigators

Linked publications & trials

Abstract

? DESCRIPTION (provided by applicant): Thousands of studies searching for gene-gene and gene-environment interaction effects on human traits and diseases have been conducted over the past 10 years. Almost all of them considered pairwise interaction only, i.e. testing for interaction effects between a single genetic variant and a single interacting factor (i.e. another genetic variant or an environmental exposure). Despite this huge effort, no pairwise interaction has been clearly established and replicated in independent data. In this proposal, we hypothesize that interaction effects take place at a higher structural level, involving scaling variables that affect the association between a phenotype and multiple genetic variants at the same time, and resulting in very low marginal pairwise interaction effects. We aim at developing and applying innovative strategies that to detect such association patterns. We especially target risk modifiers that can alter the global effect of genetic backgrounds, and vice versa genetic background that can alter the marginal association of identified genetic and non-genetic risk factors. We devised two complementary approaches to identify such effects. We will first develop and evaluate strategies for the identification of factors modifying the effect of genetic risk scores (GRS) built from SNPs found associated to human traits in genome-wide association scan (GWAS). These approaches will be used to test for interaction effects between the GRS of several human traits in the NHS and EGEA studies. Second, we will explore the possibility of interaction effects between established genetic and non-genetic risk factors and genetic background as measured by global and local ancestry in African-American admixed population using data from the COPDGene study. This analysis will focus on pulmonary phenotypes, which are known to be associated with the proportion of African ancestry. The proposed approach differs profoundly from existing ones and represents a pioneering work. Its validation has the potential to change the perspective on how genetic epidemiologists define strategies for the identification of interaction effects in human traits.