Bayesian methods for genetic association analysis with heterogeneous subgroups: From meta-analyses to gene-environment interactions

摘要

Genetic association analyses often involve data from multiplepotentially-heterogeneous subgroups. The expected amount of heterogeneity canvary from modest (e.g., a typical meta-analysis) to large (e.g., a stronggene--environment interaction). However, existing statistical tools are limitedin their ability to address such heterogeneity. Indeed, most geneticassociation meta-analyses use a "fixed effects" analysis, which assumes noheterogeneity. Here we develop and apply Bayesian association methods toaddress this problem. These methods are easy to apply (in the simplest case,requiring only a point estimate for the genetic effect and its standard error,from each subgroup) and effectively include standard frequentist meta-analysismethods, including the usual "fixed effects" analysis, as special cases. Weapply these tools to two large genetic association studies: one a meta-analysisof genome-wide association studies from the Global Lipids consortium, and thesecond a cross-population analysis for expression quantitative trait loci(eQTLs). In the Global Lipids data we find, perhaps surprisingly, that effectsare generally quite homogeneous across studies. In the eQTL study we find thateQTLs are generally shared among different continental groups, and discussconsequences of this for study design.