Abstract: There is much recent interest in statistical methods regarding the false discovery rate (FDR). The literature on this topic has two themes. In the first, authors propose sequential testing procedures that control the false discovery rate. In the second, authors study the procedures involving FDR in a univariate mixture model setting. While this work is useful for the selection of hypotheses, there is interest in estimation as well. We take an Empirical Bayes approach and propose estimators and associated confidence intervals in the multiple testing setting. Our framework is general; the proposed methodology is applied to data from a genome scan in Alzheimer's disease.
Key words and phrases: Estimation target, hypothesis testing, James-Stein estimation, multiple comparisons, simultaneous inference.