Testing for Monotonic Changes in Multivariate Gene Expression Distributions

• Activity Number: 427
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 2:00 PM to 3:50 PM
• Sponsor: IMS
• Author(s): Heng Wang*+ and Dan Nettleton
• Companies: Iowa State University and Iowa State University
• Keywords: Gene sets ; Order-restricted alternatives ; small sample sizes ; high-dimensional distributions

Abstract

In this study, four procedures are proposed for detecting monotonic changes in multivariate gene expression distributions. We consider cases where the levels of a treatment factor are quantitative or can be ordered in some intuitive manner. The testing procedures are applied to detect gene sets that express monotonically in the sense that each dimension of the multivariate gene expression distribution is either increasing or decreasing with increasing treatment level. Our procedures are able to deal with small sample sizes and high-dimensional distributions. Our simulations show that the proposed methods for detecting monotonic changes perform better than two other popular testing methods which are designed for detecting more general changes in multivariate distributions.