scanMiR: A biochemically based toolkit for versatile and efficient microRNA target prediction

Abstract

Motivation: MicroRNAs are important post-transcriptional regulators of gene expression, but the identification of functionally relevant targets is still challenging. Recent research has shown improved prediction of microRNA-mediated repression using a biochemical model combined with empirically-derived k-mer affinity predictions; however, these findings are not easily applicable. Results: We translate this approach into a flexible and user-friendly bioconductor package, scanMiR, also available through a web interface. Using lightweight linear models, scanMiR efficiently scans for binding sites, estimates their affinity and predicts aggregated transcript repression. Moreover, flexible 3′-supplementary alignment enables the prediction of unconventional interactions, such as bindings potentially leading to target-directed microRNA degradation or slicing. We showcase scanMiR through a systematic scan for such unconventional sites on neuronal transcripts, including lncRNAs and circRNAs. Finally, in addition to the main bioconductor package implementing these functions, we provide a user-friendly web application enabling the scanning of sequences, the visualization of predicted bindings and the browsing of predicted target repression.