Dynamic Bayesian testing of sets of variants in complex diseases.

ABSTRACT: Rare genetic variants have recently been studied for genome-wide associations with human complex diseases. Existing rare variant methods are based on the hypothesis-testing framework that predefined variant sets need to be tested separately. The power of those methods is contingent upon accurate selection of variants for testing, and frequently, common variants are left out for separate testing. In this article, we present a novel Bayesian method for simultaneous testing of all genome-wide variants across the whole frequency range. The method allows for much more flexible grouping of variants and dynamically combines them for joint testing. The method accounts for correlation among variant sets, such that only direct associations with the disease are reported, whereas indirect associations

SUBMITTER: Zhang Y 

PROVIDER: S-EPMC4224176 | biostudies-literature | 2014 Nov

REPOSITORIES: biostudies-literature