Greg Grant
Researcher
Center for Bioinformatics
Department of Computer & Information Science
Methods in differential expression analysis of microarray data
The desire to use microarray data to find the differentially expressed genes between two experimental conditions has presented statisticians with several new challenges. We will discuss the evolution of methods for the differential expression problem from the classical approach of controlling the family-wise error rate, to the newer trend towards controlling the false discovery rate. We will then discuss how, up to this point, control of the false discovery rate in practice has involved making numerous assumptions which, ideally, should be removed, and how recent developments allow weakening of several of these assumptions.