RNAmars: identification of multivalent RNA motifs and cognate regulators of alternative splicing

Pre-mRNA alternative splicing (AS) is the process of exon choice that increases protein diversity of eukaryotic genes. AS is finely regulated by RNA-binding proteins (RBPs) that typically bind clusters of short degenerate sequences, known as multivalent motifs (Cereda et al, 2014) to enhance or repress exon inclusion in a position-dependent manner. In spite of the abundance of transcriptomic data, it remains challenging to identify the RBPs responsible for the AS of distinct exons that are differentially regulated under a phenotype of interest. Hence, the identification of trans-acting factors that bind these cis-acting elements is a non-trivial task when relying only on sequence information. Here we present an enhanced version of the RNAmotifs algorithm for the identification of RNA Motifs And cognate Regulators of Alternative Splicing (RNAmars). Learning from results of enhanced crosslinking immunoprecipitation (eCLIP) and RNA-sequencing experiments available from the ENCODE Project, RNAmars identifies multivalent RNA motifs that are enriched at specific positions around the enhanced and silenced exons and the RBPs that specifically bind these elements. Applied to 384 primary prostate cancer transcriptomes available from TCGA, RNAmars revealed the role of specific RBPs, such as HNRNPK and U2AF2, in controlling AS in tumors with high expression of the oncogene FOXA1. By identifying the link between trans- and cis-acting elements, RNAmars enables a more straightforward comparison of regulatory principles of different splicing factors.