2L-piRNA: A Two-Layer Ensemble Classifier for Identifying Piwi-Interacting RNAs and Their Function

摘要: Involved with important cellular or gene functions and implicated many kinds of cancers, piRNAs, piwi-interacting RNAs, are small non-coding RNA around 19–33 nt in length. Given a molecule, can we predict whether it is piRNA according to its sequence information alone? Furthermore, there two types piRNA: one has the function instructing target mRNA deadenylation, other does not. Can discriminate from other? With avalanche sequences emerging postgenomic age, urgent address problems for both basic research drug development. Unfortunately, best our knowledge, so far no computational methods whatsoever could be used deal the second problem, let alone together. Here, by incorporating physicochemical properties nucleotides into pseudo K-tuple nucleotide composition (PseKNC), proposed powerful predictor called 2L-piRNA. It two-layer ensemble classifier, which first layer identifying query molecule non-piRNA, second without deadenylation. Rigorous cross-validations have indicated that success rates achieved quite high. For convenience most biologists development scientists, web server 2L-piRNA been established at http://bioinformatics.hitsz.edu.cn/2L-piRNA/, users easily get their desired results need go through mathematical details.