Mixture Model and Empirical Bayes Analysis of Microarray Gene Expression Studies for Drug Target Identification

• Activity Number: 233
• Type: Invited
• Date/Time: Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section*
• Author(s): David Allison*+
• Affiliation(s): University of Alabama, Birmingham
• Keywords: microarrays ; gene experssion ; pharmacogenomics ; mixture models

Abstract

Microarrays have emerged as powerful tools allowing investigators to simultaneously assess the expression of thousands of genes in different tissues and organisms. Statistical treatment of the resulting data remains a substantial challenge. This is due to many factors, perhaps most notably a very small number of cases (e.g., < 10) and a very large number of variables (e.g., > 10,000) in a typical dataset. Such datasets create substantial problems of multiple testing and/or low power if treated in a conventional frequentist manner. Investigators using microarray expression studies may wish to answer questions about the statistical significance of differences in expression of any of the genes under study, avoiding false positive and false negative results. We have developed a sequence of procedures involving finite mixture modeling, bootstrap inference, and combinations of fully Bayes and empirical Bayes methods to address these issues. We illustrate the use of these techniques with a dataset involving calorically restricted mice.