Title: "Identifying Differentially Expressed Genes in DNA Microarray Experiments"
Speaker: Dr. John Storey
Place: SMITH HALL 108; Tuesday, 4:30pm


It is now common to obtain DNA microarrays over two or more biological conditions in an attempt to determine which genes show a statistically significant change in expression across these conditions. In this talk, a new multivariate frequentist approach to identifying differentially expressed genes is presented. We extend the Neyman-Pearson paradigm to the multivariate case in terms of the false discovery rate criterion, which is the expected proportion of false positives among all genes called differentially expressed. A new nonparametric method is presented that fits into this multivariate Neyman-Pearson paradigm and uses recently developed methods for estimating false discovery rates. Comparisons to other proposed frequentist and Bayesian methods are also made.