R code for Bayesian FDR control

This repository contains R code to perfrom Robust Bayesian FDR control. The statistical methods are described in this manuscript.

License

Software distributed under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. See LICENSE for more details.

R functions

• EBF_fdr implements the EBF procedure, which estimates upper-bound of pi_0 using sample mean. The input to the function consists of a vector of pre-computed Bayes factors and the alpha-level for FDR control (if ignored, alpha is set to 0.05)

• BF_fdr implements a generic Bayes factor-based Bayesian FDR control procedure. Comparing to the EBF_fdr fucntion, it requires an additional parameter pi_0 to be specified. Typically, this function is used when pi_0 is estimated based on sample quantiles (i.e., the QBF procedure).

Reference and Citation

Wen, X. Robust Bayesian FDR Control using Bayes Factors, with Applications to Multi-tissue eQTL Discovery. Statistics in Biosciences (In press).