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

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