Picking the Most Likely Candidates for Further Development: Novel Intersection-union Tests for Addressing Multicomponent Hypotheses in Comparative Genomics

• Activity Number: 326
• Type: Topic Contributed
• Date/Time: Wednesday, August 11, 2004 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): David B. Allison*+
• Companies: University of Alabama, Birmingham
• Keywords: microarrays ; target identification ; intersection union tests ; RNAi ; Bayesian ; Frequentist

Abstract

In selecting genomic targets and physiological pathways for future study or drug development, investigators select a subset of candidates from among many options. In prioritizing, they may seek potential targets that meet all of several criteria--e.g., genes that are differentially expressed in response to a particular stimulus in each of several species (i.e,. evolutionarily conserved responses), or genes that are differentially expressed in response to a stimulus and also produce a predicted phenotypic change when knocked down in with RNAi. Testing whether several (k) null hypotheses can all be rejected requires intersection-union tests (IUTs). A traditional IUT rejects the union of all k null hypotheses in favor of the intersection of all k alternative hypotheses if a test for every one of the separate k null hypotheses is rejected. This IUT is generally conservative. Moreover, it yields results classifiable as significant or not, but not a quantitative p-value. Herein, we examine an approach to frequentist testing to overcome these limitations. We then present a Bayesian approach to the problem that makes more effective use of data when many IUTs are conducted simultaneously.