Assessing the Significance of a Gene Set

• Activity Number: 299
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 PM
• Sponsor: Biometrics Section
• Author(s): Huey-Miin Hsueh*+ and Chen-An Tsai and Da-Wei Zhuo
• Companies: National Chengchi University and China Medical University and National Chengchi University
• Keywords: Gene set analysis ; permutation ; p-value ; random forest

Abstract

In DNA microarray studies, a gene-set analysis (GSA) is used to evaluate the association between the expression of biological pathways, or a priori defined gene sets, and a particular phenotype. Two types of differentially expressed testing are of research interest: the competitive testing and the self-contained testing. The competitive test is to determine whether the specific gene set is relatively differentially expressed when compared to other gene sets. The self-contained test is interested in finding whether the gene set alone is differentially expressed. The two tests involve different null distributions. To take consideration on the interaction or correlation within the gene set, we consider assessing the significance of the gene set by the performance of a classifier developed upon the gene set. In this study, the Random Forest classification is applied. For each of the two tests, the corresponding empirical P-value of an observed out-of-bag (OOB) error rate of the classifier is introduced by using adequate resampling method. Several real examples are analyzed for comparison. A simulation study is conducted for verification.