Accounting for Nuisance Covariates When Using RNA-Seq Data to Identify Differentially Expressed Genes

• Activity Number: 186
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Yet T. Nguyen*+ and Dan Nettleton
• Companies: Iowa State University and Iowa State University
• Keywords: Model Selection ; FDR ; RNA Sequencing

Abstract

High-throughput DNA sequencing technologies can be used to identify sequences of bases that occur in samples of RNA. This approach (known as RNA sequencing or RNA-seq for short) provides counts that serve as measures of transcript abundance in a biological sample for each of thousands of genes. The amount of messenger RNA (mRNA) produced by a gene is often referred to as a gene's expression level. Thus, RNA-seq provides expression level measurements for thousands of genes. When RNA-seq technology is applied to multiple independent samples of different types, researchers often want to determine which genes are differentially expressed, i.e., which genes have mean levels of expression that differ across sample types. Analyses can often be complicated by the presence of nuisance factors that arise due to experimental design limitations and heterogeneity of experimental units that can be seen in continuous covariates measured for each experimental unit and/or RNA sample. This talk will present examples of such nuisance covariates and describe inference strategies that account for their effects.