Abstract

The usual (frequentist) approach to assessing evidence for a population association between genetic variants and a phenotype of interest is to compute a p-value for the null hypothesis (H 0 ) of no association. Despite their widespread use, p-values have a striking and fundamental limitation Bayesian methods provide an alternative approach to assessing associations that alleviates the limitations of p-values at the cost of some additional modelling assumptions. For example, a Bayesian analysis requires explicit assumptions about effect sizes at truly associated SNPs. Because of computational constraints, Bayesian approaches were not widely used until about 15 years ago, since when they have become more prevalent in many areas of science, including genetics 4-9 . This advance is now extending to genetic association studies, as recent papers have shown practical and theoretical advantages of using Bayesian approaches for the assessment of association Many genetics researchers are currently unfamiliar with Bayesian methods, and some may be reluctant to adopt them because they fear that editors and reviewers will also be unfamiliar with them. However, we believe that the benefits of Bayesian methods will lead to their widespread use in future genetic association analyses. Bayesian methods compute measures of evidence that can be directly compared among SNPs within and across studies. In addition, they provide a rational and quantitative way to incorporate biological information, and they can allow for a range of possible genetic models in a single analysis. Moreover, Bayesian approaches allow a coherent approach to combining results across studies (meta-analysis), across SNPs in genes and across gene pathways, which will be increasingly important as we move from single-SNP analyses towards more integrative approaches. In this Review, we present a guide for newcomers to understanding and implementing a Bayesian analysis in some of the most common settings. We focus particularly on the additional modelling assumptions Abstract | Bayesian statistical methods have recently made great inroads into many areas of science, and this advance is now extending to the assessment of association between genetic variants and disease or other phenotypes. We review these methods, focusing on single-SNP tests in genome-wide association studies. We discuss the advantages of the Bayesian approach over classical (frequentist) approaches in this setting and provide a tutorial on basic analysis steps, including practical guidelines for appropriate prior specification. We demonstrate the use of Bayesian methods for fine mapping in candidate regions, discuss meta-analyses and provide guidance for refereeing manuscripts that contain Bayesian analyses.