STarMir: a web server for prediction of microRNA binding sites.

William Rennie,C Steven Carmack,Chaochun Liu,Shaveta Kanoria,Ye Ding,Adam Wolenc,Jun Lu,Dang Long

doi:10.1093/nar/gku376

William Rennie, C Steven Carmack + Show 6 more

Open Access

https://doi.org/10.1093/nar/gku376

Copy DOI

Abstract

STarMir web server predicts microRNA (miRNA) binding sites on a target ribonucleic acid (RNA). STarMir is an implementation of logistic prediction models developed with miRNA binding data from crosslinking immunoprecipitation (CLIP) studies (Liu,C., Mallick, B., Long, D., Rennie, W.A., Wolenc, A., Carmack, C.S. and Ding, Y. (2013). CLIP-based prediction of mammalian microRNA binding sites. Nucleic Acids Res., 41(14), e138). In both intra-dataset and inter-dataset validations, the models showed major improvements over established algorithms in predictions of both seed and seedless sites. General applicability of the models was indicated by good performance in cross-species validations. The input data for STarMir is processed by the web server to perform prediction of miRNA binding sites, compute comprehensive sequence, thermodynamic and target structure features and a logistic probability as a measure of confidence for each predicted site. For each of seed and seedless sites and for all three regions of a mRNA (3′ UTR, CDS and 5′ UTR), STarMir output includes the computed binding site features, the logistic probability and a publication-quality diagram of the predicted miRNA:target hybrid. The prediction results are available through both an interactive viewer and downloadable text files. As an application module of the Sfold RNA package (http://sfold.wadsworth.org), STarMir is freely available to all at http://sfold.wadsworth.org/starmir.html.

Highlights

MicroRNAs are a class of small endogenous non-coding ribonucleic acid (RNA) of ∼22 nucleotides in length that have been found in plants, animals and viruses. miRNAs are involved in the post-transcriptional regulation of gene expression by binding to target messenger RNAs, leading to translational repression and/or mRNA destabilization [1]
Prediction and validation of miRNA targets are essential for understanding functions of miRNAs in gene regulation
Algorithms for the prediction of miRNA:target binding sites are typically based on the seed rule, i.e. the target forms Watson–Crick (WC) pairs with bases two through seven or eight, at the 5 end of the miRNA [2]

Summary

INTRODUCTION

MicroRNAs (miRNAs) are a class of small endogenous non-coding RNAs (ncRNAs) of ∼22 nucleotides (nts) in length that have been found in plants, animals and viruses. miRNAs are involved in the post-transcriptional regulation of gene expression by binding to target messenger RNAs (mRNAs), leading to translational repression and/or mRNA destabilization [1]. STarMir web server allows users to submit miRNA and mRNA sequences for prediction of binding sites by the models. A comprehensive list of sequence, thermodynamic and structure-based features are computed as previously described [9] These features are used by our logistic model with parameters specific for the site type (seed or seedless) and the target region (5 UTR, CDS or 3 UTR) to compute a logistic probability as a measure of confidence in the predicted site. The user needs to input sequence information for one or more miRNAs and a single target mRNA for job processing by the web server. The user selects the species for prediction This information will be used if the RefSeq ID is entered for the target mRNA sequence so that the server can utilize prestored evolutionary conservation information in the modeling computation. The output results are presented to the user through both an interactive viewer and downloadable files

C CCU C UU

CONCLUSION