The BFDP combines significance level, study power, and cost of false discovery and non-discovery into consideration. The detailed procedures of this method are described in Wakefield, 200734. Essentially, the approximate Bayes Factor (ABF) which BFDP uses reflects how much the prior odds change in the light of the observed data (i.e. relative probability of the observed estimates under the null versus alternative hypothesis). Given the nature of GWA studies, we applied a flat prior for all variants at prior probability of 10−6 and 10−8 to demonstrate the range of BFDP.
