All Times EDT
Keywords: Gene environment interaction, Complex traits, Rare variants, Exome sequencing, Pharmacogenomics
Studying genotype-by-environment interaction (GEI) is fundamental in understanding complex trait variations. Identifying genetic variants with GEI effects is challenging because GEI tests typically have low power. Here we propose Multi-trait Analysis of Gene-ENvironmenT-wide Association (MAGENTA) to test GEI on multiple traits in large-scale datasets, such as the UK Biobank. Our method is motivated by the fact that the presence of GEI implies the heterogeneity of the genetic effects across different environmental groups. MAGENTA combines the summary statistics for the association of variants in a gene with multiple traits under different environmental groups and produces a GEI test and a joint test for testing both main and GEI effects. MAGENTA is an omnibus method that is robust to a wide spectrum of genetic architectures over multiple traits and variants. Our simulation studies demonstrate that MAGENTA GEI and joint tests properly control type I error and yield higher and more robust power than existing single-trait-based GEI tests across a wide range of scenarios. We performed GEI analysis on UK Biobank Whole Exome Sequencing data with four environmental variables. For each environmental variable, we tested its interaction with genes on three lipid traits. MAGENTA GEI test identified a significant APOE gene-by-sex interaction, which was missed by the single-trait-based GEI test. Furthermore, MAGENTA joint test identified 41% more significant genes than the single-trait-based joint test.