gms | German Medical Science

Advances in high-throughput genotyping technology lead to the realization of genome-wide association studies (GWAS) which helped identify new susceptibility loci of complex traits. Such studies usually start with genotyping fixed arrays of genetic markers in an initial sample. Out of these markers, some are selected which will be further genotyped in independent samples. Due to the very low a priori probability of a true positive association, the vast majority of all marker signals will turn out to be false positive.

Some of the true and robust marker associations derived from classical candidate gene approaches, however, will not be among those markers with the smallest unadjusted p-values. Consequently, alternative methods to sort marker data have been proposed. In this talk I will introduce and evaluate some properties of the "q-values" [1], the "false positive report probability" [2], and the "Bayesian false-discovery probability" [3] as examples of such alternative ranking methods. In particular, I performed simulation studies based on a population genetic coalescent model and simulations using real data for a genomic region derived from GWAS data sets. Furthermore, I applied all ranking methods to a GWAS case-control data set. For the purpose of selecting markers from a GWAS and within the limits of the simulation, I show the benefits of both the ranking by p-values and the application of a full Bayesian approach for which the explored "Bayesian false-discovery probability" is a first approximation.