A systematic approach to RNA-associated motif discovery

• DOI: 10.1186/s12864-018-4528-x
• Publication Date: 2018-02-14T12:44:07Z
• Authors: Tian Gao; Jiang Shu; Juan Cui
• Keywords: 570; 610; QH426-470; Exosomes; 03 medical and health sciences; Cell Line, Tumor; Genetics; Humans; RNA, Messenger; Nucleotide Motifs; Graph algorithms; Internet; 0303 health sciences; Binding Sites; Short sequences; Base Sequence; Computer Sciences; Life Sciences; Computational Biology; High-Throughput Nucleotide Sequencing; Reproducibility of Results; Genetics and Genomics; microRNAs; Motif finding; MicroRNAs; Exosomal RNAs; Other Computer Sciences; TP248.13-248.65; Biotechnology; Research Article

Sequencing-based large screening of RNA-protein and RNA-RNA interactions has enabled the mechanistic study of post-transcriptional RNA processing and sorting, including exosome-mediated RNA secretion. The downstream analysis of RNA binding sites has encouraged the investigation of novel sequence motifs, which resulted in exceptional new challenges for identifying motifs from very short sequences (e.g., small non-coding RNAs or truncated messenger RNAs), where conventional methods tend to be ineffective. To address these challenges, we propose a novel motif-finding method and validate it on a wide range of RNA applications.We first perform motif analysis on microRNAs and longer RNA fragments from various cellular and exosomal sources, and then validate our prediction through literature search and experimental test. For example, a 4 bp-long motif, GUUG, was detected to be responsible for microRNA loading in exosomes involved in human colon cancer (SW620). Additional performance comparisons in various case studies have shown that this new approach outperforms several existing state-of-the-art methods in detecting motifs with exceptional high coverage and explicitness.In this work, we have demonstrated the promising performance of a new motif discovery approach that is particularly effective in current RNA applications. Important discoveries resulting from this work include the identification of possible RNA-loading motifs in a variety of exosomes, as well as novel insights in sequence features of RNA cargos, i.e., short non-coding RNAs and messenger RNAs may share similar loading mechanism into exosomes. This method has been implemented and deployed as a new webserver named MDS2 which is accessible at http://sbbi-panda.unl.edu/MDS2/ , along with a standalone package available for download at https://github.com/sbbi/MDS2 .