Combining P-Values for Gene Set Analysis

• Activity Number: 186
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Ziwen Wei*+ and Lynn Kuo
• Companies: Merck and University of Connecticut

Abstract

The classical way to analyze the high-throughput microarray gene expression data is to analyze the genes individually. However, several researchers have pointed out that it is advantageous to analyze the microarray data at the level of gene sets that share common biological function, chromosomal location, or regulation. Gene set analysis will ease the interpretation of a large-scale experiment by identifying important pathways and processes. An increasing number of gene set analysis methods are proposed and many of them are commonly used. In this paper, we propose a gene set method based on summarizing individual p-values within each gene set. In addition, we evaluate the proposed method by a simulation experiment and compare it with six existing methods (SAM-GS, global test, global ANCOVA, GSEA, GSA and random set) in terms of the false positive rate, false negative rate, false discovery rate, false non-discovery rate and receiver operating characteristic curve. An application to the real data is also provided.