Adaptive Procedures Controlling False Discoveries Under Dependence

• Activity Number: 642
• Type: Contributed
• Date/Time: Thursday, August 8, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Li He*+ and Sanat K. Sarkar
• Companies: Merck Research Laboratories and Temple University
• Keywords: False discovery rate ; multiple testing ; adaptive BH procedures ; positive dependence ; pairwise correlations

Abstract

Although Blanchard and Roquain (2009) presented for the first time methods of adapting the Benjamini-Hochberg (BH) method to the data through an estimate of the proportion of true null hypotheses that continue to control the false discovery rate (FDR) under positive dependence, they are often too conservative. To improve the performances of these methods in some specific positive dependence situations without losing the FDR control, we propose to use their alternative versions under some additional assumptions by explicitly utilizing pairwise correlations whenever they are available. We offer numerical evidence of improved performances of the proposed alternatives under some positive dependence scenarios, including those assumed in this talk.