IRcall and IRclassifier: two methods for flexible detection of intron retention events from RNA-Seq data

The emergence of next-generation RNA sequencing (RNA-Seq) provides tremendous opportunities for researchers to analyze alternative splicing on a genome-wide scale. However, accurate detection of intron retention (IR) events from RNA-Seq data has remained an unresolved challenge in next-generation sequencing (NGS) studies.

We propose two new methods: IRcall and IRclassifier to detect IR events from RNA-Seq data. Our methods combine together gene expression information, read coverage within an intron, and read counts (within introns, within flanking exons, supporting splice junctions, and overlapping with 5' splice site/ 3' splice site), employing ranking strategy and classifiers to detect IR events. We applied our approaches to one published RNA-Seq data on contrasting skip mutant and wild-type in Arabidopsis thaliana. Compared with three state-of-the-art methods, IRcall and IRclassifier could effectively filter out false positives, and predict more accurate IR events.

The data and codes of IRcall and IRclassifier are available at http://mlg.hit.edu.cn/ybai/IR/IRcallAndIRclass.html.