Corrected False-Discovery Rate for Removing the Gene-Set-Level Bias of RNA-Seq

• Activity Number: 185
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Korean International Statistical Society
• Author(s): Seongmun Jeong*+ and Tae Young Yang
• Companies: Myongji University and Myongji University
• Keywords: Corrected false discovery rate ; empirical selection probability ; FDRseq ; Gene-level bias ; Gene-set-level bias ; RNA-seq

Abstract

In recent years, RNA-seq has become a very competitive alternative to microarrays. In RNA-seq experiments, the expected read count for a gene is proportional to its expression level multiplied by its transcript length. The characteristics of RNA-seq experiments create a gene-level bias such that the proportion of significantly differentially expressed genes increases with the transcript length, but such bias is not present in microarray data. In the gene set analysis of RNA-seq, the gene-level bias consequently yields the gene-set-level bias. Because gene expression is not related to its transcript length, any gene set containing long genes is not of biologically greater interest than gene sets with shorter genes. Accordingly, the gene-set-level bias should be removed to allow accurate calculation of the statistical significance of each gene-set enrichment in the RNA-seq. Controlling the false-discovery rate that we provided should give a similar number of enriched gene sets in each category so that the gene sets containing long genes are not more enriched than those with shorter genes. Therefore, the gene-set-level bias can be properly removed.