A Bayesian method for rare variant analysis using functional annotations and its application to Autism

Abstract

Rare genetic variants make significant contributions to human diseases. Compared to common variants, rare variants have larger effect sizes and are generally free of linkage disequilibrium (LD), which makes it easier to identify causal variants. Numerous methods have been developed to analyze rare variants in a gene or region in association studies, with the goal of finding risk genes by aggregating information of all variants of a gene. These methods, however, often make unrealistic assumptions, e.g. all rare variants in a risk gene would have non-zero effects. In practice, current methods for gene-based analysis often fail to show any advantage over simple single-variant analysis. In this work, we develop a Bayesian method: MIxture model based Rare variant Analysis on GEnes (MIRAGE). MIRAGE captures the heterogeneity of variant effects by treating all variants of a gene as a mixture of risk and non-risk variants, and models the prior probabilities of being risk variants as function of external information of variants, such as allele frequencies and predicted deleterious effects. MIRAGE uses an empirical Bayes approach to estimate these prior probabilities by combining information across genes. We demonstrate in both simulations and analysis of an exome-sequencing dataset of Autism, that MIRAGE significantly outperforms current methods for rare variant analysis. In particular, the top genes identified by MIRAGE are highly enriched with known or plausible Autism risk genes. Our results highlight several novel Autism genes with high Bayesian posterior probabilities and functional connections with Autism. MIRAGE is [available].