RNA-protein interactions (RPIs) have critical roles in numerous fundamental biological processes, such as post-transcriptional gene regulation, viral assembly, cellular defence and protein synthesis. As the number of available RNA-protein binding experimental data has increased rapidly due to high-throughput sequencing methods, it is now possible to measure and understand RNA-protein interactions by computational methods. In this study, we integrate a sequence-based derived kernel with regularized least squares to perform prediction. The derived kernel exploits the contextual information around an amino acid or a nucleic acid as well as the repetitive conserved motif information. We propose a novel machine learning method, called RPiRLS to predict the interaction between any RNA and protein of known sequences. For the RPiRLS classifier, each protein sequence comprises up to 20 diverse amino acids but for the RPiRLS-7G classifier, each protein sequence is represented by using 7-letter reduced alphabets based on their physiochemical properties. We evaluated both methods on a number of benchmark data sets and compared their performances with two newly developed and state-of-the-art methods, RPI-Pred and IPMiner. On the non-redundant benchmark test sets extracted from the PRIDB, the RPiRLS method outperformed RPI-Pred and IPMiner in terms of accuracy, specificity and sensitivity. Further, RPiRLS achieved an accuracy of 92% on the prediction of lncRNA-protein interactions. The proposed method can also be extended to construct RNA-protein interaction networks. The RPiRLS web server is freely available at http://bmc.med.stu.edu.cn/RPiRLS.

Gene post-transcription regulation involves several critical regulators such as microRNAs (miRNAs) and RNA-binding proteins (RBPs). Accumulated experimental evidences have shown that miRNAs and RBPs can competitively regulate the shared targeting transcripts. Although this establishes a novel post-transcription regulation mechanism, there are currently no computational tools to scan for the possible competing miRNA and RBP pairs. Here, we developed a novel computational pipeline—RBPvsMIR—that enables us to statistically evaluate the competing relationship between miRNAs and RBPs. RBPvsMIR first combines with previously successful miRNAs and RBP motifs discovery applications to search for overlapping or adjacent binding sites along a given RNA sequence. Then a permutation test is performed to select the miRNA and RBP pairs with the significantly enriched binding sites. As an example, we used RBPvsMIR to identify 235 competing RBP-miRNA pairs for long non-coding RNA (lncRNA) MALAT1. Wet lab experiments verified that splicing factor SRSF2 competes with miR-383, miR-502 and miR-101 to regulate MALAT1 in esophageal squamous carcinoma cells. Our study also revealed the global mutual exclusive pattern for miRNAs and RBP to regulate human lncRNAs. In addition, we provided a convenient web server (http://bmc.med.stu.edu.cn/RBPvsMIR), which should accelerate the exploration of competing miRNAs and RBP pairs regulating the shared targeting transcripts.

Recent analyses show that transcriptome sequencing can be utilized as a diagnostic tool for rare Mendelian diseases. The third generation sequencing de novo detects long reads of thousands of base pairs, thus greatly expanding the isoform discovery and identification of novel long noncoding RNAs. In this study, we developed TGStools, a bioinformatics suite to facilitate routine tasks such as characterizing full-length transcripts, detecting shifted types of alternative splicing, and long noncoding RNAs (lncRNAs) identification in transcriptome analysis. It also prioritizes the transcripts with a visualization framework that automatically integrates rich annotation with known genomic features. TGStools is a Python package freely available at Github.