miES: predicting the essentiality of miRNAs with machine learning and sequence features.

Fei Song,Chunmei Cui,Lin Gao,Qinghua Cui

DOI: https://doi.org/10.1093/bioinformatics/bty738

IF: 5.8

2019-01-01

Bioinformatics

Abstract:Motivation MicroRNAs (miRNAs) are one class of small noncoding RNA molecules, which regulate gene expression at the post-transcriptional level and play important roles in health and disease. To dissect the critical miRNAs in miRNAome, it is needed to predict the essentiality of miRNAs, however, bioinformatics methods for this purpose are limited. Results Here we propose miES, a novel algorithm, for the prioritization of miRNA essentiality. miES implements a machine learning strategy based on learning from positive and unlabeled samples. miES uses sequence features of known essential miRNAs and performs miRNAome-wide searching for new essential miRNAs. miES achieves an AUC of 0.9 for 5-fold cross validation. Moreover, experiments further show that the miES score is significantly correlated with some established biological metrics for miRNA importance, such as miRNA conservation, miRNA disease spectrum width (DSW) and expression level. Availability and implementation The R source code is available at the download page of the web server, http://www.cuilab.cn/mies. Supplementary information Supplementary data are available at Bioinformatics online.