RIsearch: fast RNA–RNA interaction search using a simplified nearest-neighbor energy model

Abstract

Motivation: Regulatory, non-coding RNAs often function by forming a duplex with other RNAs. It is therefore of interest to predict putative RNA–RNA duplexes in silico on a genome-wide scale. Current computational methods for predicting these interactions range from fast complementary-based searches to those that take intramolecular binding into account. Together these methods constitute a trade-off between speed and accuracy, while leaving room for improvement within the context of genome-wide screens. A fast pre-filtering of putative duplexes would therefore be desirable.Results: We present RIsearch, an implementation of a simplified Turner energy model for fast computation of hybridization, which significantly reduces runtime while maintaining accuracy. Its time complexity for sequences of lengths m and n is with a much smaller pre-factor than other tools. We show that this energy model is an accurate approximation of the full energy model for near-complementary RNA–RNA duplexes. RIsearch uses a Smith–Waterman-like algorithm using a dinucleotide scoring matrix which approximates the Turner nearest-neighbor energies. We show in benchmarks that we achieve a speed improvement of at least 2.4× compared with RNAplex, the currently fastest method for searching near-complementary regions. RIsearch shows a prediction accuracy similar to RNAplex on two datasets of known bacterial short RNA (sRNA)–messenger RNA (mRNA) and eukaryotic microRNA (miRNA)–mRNA interactions. Using RIsearch as a pre-filter in genome-wide screens reduces the number of binding site candidates reported by miRNA target prediction programs, such as TargetScanS and miRanda, by up to 70%. Likewise, substantial filtering was performed on bacterial RNA–RNA interaction data.Availability: The source code for RIsearch is available at: http://rth.dk/resources/risearch.Contact: gorodkin@rth.dkSupplementary information: Supplementary data are available at Bioinformatics online.

Small RNA Transcriptome of the Oral Microbiome during Periodontitis Progression

“…When dealing with a large number of genomes, as in the present case, in silico prediction of sRNA targets is computationally intensive; thus, we decided to use RIsearch under highly stringent energy interaction conditions. This software identifies putative duplexes of RNAs based on an implementation of a simplified Turner energy model, significantly reducing run-time, while at the same time maintaining accuracy (36). Among the enriched GO terms for these targets was pathogenesis, which confirms the hypothesis that sRNAs play a key role on the regulation of bacterial activity during periodontitis progression.…”

“…To better characterize the role that the DE sRNAs could have in regulating metabolic activities during periodontitis progression, we generated a set of target predictions using RIsearch with a cutoff value of Ϫ45 ⌬G (kcal/mol), which maximizes true positives according to the ROC curve presented by the authors of this software (36). A total of 148,987 genes were identified as putative targets for the 12,097 DE sRNAs in the database.…”

“…The energy cutoff was Ϫ45 ⌬G (kcal/mol), which is based on the receiver-operating characteristic (ROC) curve presented by the authors to maximize true positives and minimize the rate of true negatives (36). Target identification was performed against the ORFs of the different genomes, using only the sections of IGRs that aligned with the query sequences (see Table S1 in the supplemental material).…”

IRBIS: a systematic search for conserved complementarity

“…Besides this technical limitation, there is a fundamental problem that the additive model is insufficient to describe entropy contribution of loops in molecules with pseudoknots and that important steric and topological limitations also need to be taken into account (Pervouchine 2004). Consequently, RRI prediction methods avoid intramolecular interactions to be computationally efficient (Mückstein et al 2006;Wenzel et al 2012). The best current trade-off approach uses precomputed accessibility profiles in addition to free energy scoring of exposed binding sites (Mückstein et al 2006;Tafer et al 2011).…”

RAID: a comprehensive resource for human RNA-associated (RNA–RNA/RNA–protein) interaction

“…Specifically, RAID adopts the predicted binding sites and scores by miRanda for a miRNA and its targets (John et al 2004), while containing the predicted binding sites and score by RIsearch for the RNA-RNA interactions ( Fig. 3; Wenzel et al 2012). For RNAprotein interactions, bindN (Wang and Brown 2006), bindN+ , Pprint (Kumar et al 2008), and RNAbindR (Terribilini et al 2007) are commonly used tools to predict RNA-binding residues in proteins (Puton et al 2012).…”