|
Abstract:
|
The Benjamini-Hochberg procedure is a classical method for controlling the false discovery rate for multiple testing problems. This procedure was originally designed for continuous test statistics. However, in many applications, such as the analysis of next-generation sequencing data, the test statistics are discretely distributed. While it is well known that the Benjamini-Hochberg procedure still controls the false discovery rate in the discrete paradigm, it may be unnecessarily conservative. Thus, there is an urgent need to develop more powerful FDR procedures in the discrete paradigm. In this talk we aim to improve the Benjamini-Hochberg procedure in such settings by incorporating the discreteness of the p-value distributions. We investigate the performance of these approaches for empirical and simulated data.
|
