A length bias corrected likelihood ratio test for the detection of differentially expressed pathways in RNA-Seq data

RNA-Seq has become an important alternative to microarrays in transcriptomic studies. The unique features of count-based RNA-Seq data pose new challenges for pathway analysis and call for new computational tools. In this study, we developed a likelihood ratio test to identify differentially expressed pathways in RNA-Seq data. The proposed method takes into account the coherent gene expression patterns by considering a common variance component shared by genes within a pathway. Additionally, we implemented a method to correct the length bias existing in the differential expression analysis using RNA-Seq data, where longer transcripts are more likely to be identified as differentially expressed. We demonstrated the ability of the proposed method using both synthetic and real data. We found that the top differentially expressed pathways between liver and kidney tissue samples identified using our method are associated with organ-specific functions.