sTarPicker: A Method for Efficient Prediction of Bacterial sRNA Targets Based on a Two-Step Model for Hybridization

Abstract

BackgroundBacterial sRNAs are a class of small regulatory RNAs involved in regulation of expression of a variety of genes. Most sRNAs act in trans via base-pairing with target mRNAs, leading to repression or activation of translation or mRNA degradation. To date, more than 1,000 sRNAs have been identified. However, direct targets have been identified for only approximately 50 of these sRNAs. Computational predictions can provide candidates for target validation, thereby increasing the speed of sRNA target identification. Although several methods have been developed, target prediction for bacterial sRNAs remains challenging.ResultsHere, we propose a novel method for sRNA target prediction, termed sTarPicker, which was based on a two-step model for hybridization between an sRNA and an mRNA target. This method first selects stable duplexes after screening all possible duplexes between the sRNA and the potential mRNA target. Next, hybridization between the sRNA and the target is extended to span the entire binding site. Finally, quantitative predictions are produced with an ensemble classifier generated using machine-learning methods. In calculations to determine the hybridization energies of seed regions and binding regions, both thermodynamic stability and site accessibility of the sRNAs and targets were considered. Comparisons with the existing methods showed that sTarPicker performed best in both performance of target prediction and accuracy of the predicted binding sites.ConclusionssTarPicker can predict bacterial sRNA targets with higher efficiency and determine the exact locations of the interactions with a higher accuracy than competing programs. sTarPicker is available at http://ccb.bmi.ac.cn/starpicker/.

Small RNA Deep-Sequencing Analyses Reveal a New Regulator of Virulence in Agrobacterium fabrum C58

“…mRNA target genes of sRNAs were predicted from the A. fabrum C58 genome (accession number NC_003062, NC_003063, NC_003064, and NC_003065) using IntaRNA (Busch et al 2008), sTarPicker (Ying et al 2011), and RNApredator (Eggenhofer et al 2011) algorithms. We arbitrarily selected the 350 most significant mRNA targets provided by each algorithm.…”

A novel plasmid‐transcribed regulatory sRNA, QfsR, controls chromosomal polycistronic gene expression in Agrobacterium fabrum

“…Target genes of QfsR were predicted from the A. fabrum C58 genome using IntaRNA (Busch et al ., ), sTarPicker (Ying et al ., ) and RNApredator (Eggenhofer et al ., ) algorithms. Targets of interest were selected as previously described (Dequivre et al ., ).…”

“…Predicting the mRNA targets of QfsR would provide clues for identifying its cellular function. We applied a combination of three algorithms (RNApredator, sTarPicker and IntaRNA) (Busch et al, 2008;Eggenhofer et al, 2011;Ying et al, 2011;Dequivre et al, 2015) and selected the 54 genes jointly identified by them. These candidates were homogeneously distributed among the four replicons, i.e., the circular and linear chromosomes and AtC58 and TiC58 plasmids (Supporting Information Table S2).…”

Bioinformatics of prokaryotic RNAs

“…An sRNA target prediction web server on top of RNAplex is implemented by the software RNApredator 115 . The web server sTarPicker combines ideas from accessibility-based and concatenation-based approaches 118 . Putative seed interactions are extended by computing a joint secondary structure of sRNA and mRNA.…”