GENE-counter: a computational and statistical pipeline for assessing RNA-Seq data for genome-wide expression differences

RNA-Sequencing is a transformative method that captures the quantitative dynamics of a transcriptome with exquisite sensitivity and single-base resolution. There are, however, few computational pipelines for RNA-Seq with statistical tests that evince sufficient robustness and power as demanded by the difficult combination of small sample sizes and high variability in sequence read counts. To this end, we developed GENE-counter, a complete software pipeline for analyzing RNA-Seq data for genome-wide expression differences between replicated treatment groups. One important component of GENE-counter is a statistical test based on the NBP parameterization of the negative binomial distribution for identifying differentially expressed genome features. We used GENE-counter to analyze RNA-Seq data derived from Arabidopsis thaliana infected with a strain of defense-eliciting bacteria. We identified 308 genes that were differentially induced. Using alternative methods, we provided support for the induced expression and biological relevance of a substantial proportion of the genes. These results suggest the NBP parameterization of the negative binomial distribution is well suited for explaining RNA-Seq data and the statistical test makes GENE-counter a powerful pipeline for studying genome-wide expression changes. GENE-counter is freely available at http://changlab.cgrb.oregonstate.edu/.