A Review on Recent Computational Methods for Predicting Noncoding RNAs

Table 5

Methods to predict miRNA.

Name	URL	Feature	Prediction algorithm
CSHMM [38]		Structure only	A discriminant function based on likelihood score for a hidden Markov model
MiPred [54]		32 possible combinations of the middle nucleotide among the triplet elements, local contiguous structure sequence composition, MFE, and value of randomization test	Random forest
PlantMiRNAPred [55]	http://nclab.hit.edu.cn/PlantMiRNAPred/	115 features including (1) 17 primary sequence-related features, (2) 64 secondary structure-related features, and (3) 34 energy- and thermodynamics-related features	SVM
miRdentify [56]	http://www.ncrnalab.dk/#mirdentify/mirdentify.php	5′ heterogeneity, overhangs, negative numbers indicating 5′ overhang, thermodynamics, entropy, tailing, and multimapping	Mapping and seeking duplex-forming reads within 46-80nt distance with the guide strand
CID-miRNA [57]	https://github.com/alito/CID-miRNA	Secondary structure likelihood	Stochastic context-free grammar model, Chomsky normal form; Cocke-Young-Kasami algorithm, and Classification tree
miRank [58]	https://omictools.com/mirank-tool	36 global and local intrinsic features, including the normalized MFE of folding, the normalized base pairing propensities of both arms, and the normalized loop length	Belief propagation on a weighted graph, random walks-based ranking algorithm
miRCat [59]	http://srna-workbench.cmp.uea.ac.uk/tools/analysis-tools/mircat/	-value of alignment and MFE of secondary structure	Dynamic programming
mirTool [60]	http://centre.bioinformatics.zj.cn/mirtools/	miRNA/miRNA, absolute/relative reads count, and the most abundant tag	Folding the flanking genomic sequence using the miRDeep program
miRanalyzer [61]	http://bioinfo5.ugr.es/miRanalyzer/miRanalyzer.php	Number of bindings in read cluster sequence, normalized mean free energy of precursor sequence, number of bindings in precursor, length of read cluster, the corresponding putative mature star sequence, number of bindings in read cluster divided by the read cluster length, number of reads in read cluster, mean free energy of precursor sequence, degree of bulb asymmetry in precursor, and the number of bulbs in precursor secondary structure	Random forest
sRNAbench [62]	http://bioinfo5.ugr.es/sRNAbench/	Within cluster ratio, 5′ fluctuations, most frequent to all ratio, minimum number of hairpin bindings, minimum number of mature bindings, most frequent read, length interval, and minimum reads	Hierarchical clustering