Rank-Based Multiple Testing for Detecting Differentially Expressed Genes with Brief Overview of Parametric Methods

• Activity Number: 692
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Hossein Mansouri and Bo Li*
• Companies: Texas Tech University and Texas Tech University
• Keywords: microarray ; rank tests ; bootstrap ; multiple testing ; SAM

Abstract

Two major issues of concern in microarray data analysis are violation of the assumption of normality and influence of outliers. Since usually a very large number of tests are simultaneously carried out on microarray data, a serious concern is to control the familywise error rate (FWER), otherwise researchers may wrongly claim dozens even hundreds of genes to be differentially expressed. Another difficulty to deal with is deriving the theoretical distribution of the test statistic and calculation of the p-value. Resampling techniques such as permutation method and bootstrapping are used in data analysis to achieve better approximation of the p-values. This article provides a brief review of some permutation and bootstrap methods for the analysis of differentially expressed genes with application to a real dataset.