|
Activity Number:
|
102
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #310037 |
|
Title:
|
Conservatively Optimal Multiple Testing for Comparative Microarray Experiments
|
|
Author(s):
|
James Signorovitch*+
|
|
Companies:
|
Harvard University
|
|
Address:
|
Dept. of Biostatistics, Boston, MA, 02115,
|
|
Keywords:
|
gene expression ; optimal multiple testing ; empirical Bayes ; false discovery rate
|
|
Abstract:
|
Optimality theory for multiple hypothesis testing is considered in the context of gene expression microarray experiments. We show that optimal testing procedures are generally not identifiable but by conditioning on certain sufficient statistics we can define near-optimal procedures that are identifiable and conservative. In simulations and in real data, estimates of these near-optimal procedures are shown to substantially outperform the estimated Optimal Discovery Procedure (Storey et al. 2005) and the Empirical Alternative Hypothesis (Signorovitch 2006) by exploiting strong patterns in the data that are generally ignored.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
