Many computational tools have been proposed during the two last decades for predicting piRNAs, which are molecules with important role in post-transcriptional gene regulation. However, these tools are mostly based on only one feature that is generally related to the sequence. Discoveries in the domain of piRNAs are still in their beginning stages, and recent publications have shown many new properties.
Predicts piRNA by integrating twenty-three discriminative features. piRNApredictor is constructing on a genetic algorithm-based weighted ensemble method (GA-WE). It is able to produce good performances in the cross-species prediction. This tool determines the weights for each base learner in a self-tune manner. It shows good robustness comparing with other state-of-the-art methods and can be very useful for transposon-derived piRNA prediction.
Allows users to predict piRNAs using piRNA-transposon interaction information. Piano classifies real piRNAs and pseudo piRNAs thank to a support vector machine (SVM) method. It was applied to detect piRNAs for an important rice pest, the rice striped stem borer, Chilo suppressalis. This tool is able to achieve a high sensitivity, specificity and accuracy of over 90 per cent. It is useful for large-scale piRNA prediction from small RNA sequencing data or for genome-wide annotation of piRNAs.
Provides a multiple kernel-based support vector machine (SVM) algorithm to predict piRNAs. piRPred combines heterogeneous types of piRNA features. It allows to use heterogeneous features, each kernel implementing a class of features. This tool provides specificities: (1) several kernels that represent heterogeneous feature sets are built and used, (2) a new type of feature is explored, (3) the characteristic of piRNAs to occur in clusters on the chromosome is coded in a kernel to use it in a supervised way.
Differentiates transposon-derived piRNAs from non-piRNAs. This method is based on the utilization of sequential and physicochemical features to proceed. The features concerned are spectrum profile, mismatch profile, subsequence profile, position-specific scoring matrix, pseudo dinucleotide composition, and local structure-sequence triplet elements. This tool shows high performances as well as good robustness.
Captures the real features of piRNA and other RNA sequences. Pibomd is based on a support vector machine (SVM) algorithm that allows motif discovery for piRNA. It enables prediction of piRNAs in mouse with high specificity and sensitivity when only using 258 motifs appearing frequently. This tool can only use variable-length motifs that frequently appear in RNA sequences as features. It offers a web interface that allows to submit a single small RNA sequence or multiple small RNA sequences.
Identifies Piwi-Interacting RNAs. 2L-piRNA is useful for genome analysis and drug development, particularly in those areas involved with non-coding RNAs. It integrates physicochemical properties of nucleotides into the pseudo K-tuple nucleotide composition (PseKNC). This tool shows good performance in accuracy (Acc) and Matthews correlation coefficient (MCC). It is able to yield the second-layer prediction.
Detects piwi-interacting RNAs (piRNAs) based on circonvolution neural network (CNN). piRNN is a deep learning program suited for detection within four organisms including drosophila melanogaster and rattus norvegicus. The application was developed to perform its analysis without requiring genome or epigenomic additional data and also provides the training protocol to train both new models for new species as well as to apply a novel training to the existing ones.