Using RNA-seq Data to Detect Differentially Expressed Genes

RNA-sequencing (RNA-seq) technology has become a major choice in detecting differentially expressed genes across different biological conditions. Although microarray technology is used for the same purpose, statistical methods available for identifying differential expression for microarray data are generally not readily applicable to the analysis of RNA-seq data, as RNA-seq data comprise discrete counts of reads mapped to particular genes. In this chapter, we review statistical methods uniquely developed for detecting differential expression among different populations of RNA-seq data as well as techniques designed originally for the analysis of microarray data that have been modified for the analysis of RNA-seq data. We include a very brief description of the normalization of RNA-seq data and then elaborate on parametric and nonparametric testing procedures, as well as empirical and fully Bayesian methods. We include a brief review of software available for the analysis of differential expression and summarize the results of a recent comprehensive simulation study comparing existing methods. 2.