Table 3 M6A prediction methods

DeepM6ASeq [125]

It is based on miCLIP-Seq data at single-base resolution and detects m6A sites. It can recognize new reader FMR1.

https://github.com/rreybeyb/DeepM6ASeq

M6APred-EL [126]

It uses position-specific k-mer nucleotide propensity, physicochemical properties, and ring function hydrogen chemical properties to optimize m6A position recognition accuracy.

http://server.malab.cn/M6APred-EL/

M6AMRFS [127]

It uses dinucleotide binary encoding and local position-specific dinucleotide frequencies to encode RNA sequences. It can identify m6A sites in multiple species.

http://server.malab.cn/M6AMRFS/

SRAMP [128]

It identifies mammalian m6A sites at single-nucleotide resolution and builds m6A site predictors. SRAMP = sequence-based RNA adenosine methylation site predictor.

http://www.cuilab.cn/sramp/

iRNA-Methyl [129]

Identifying m6A sites by incorporating the global and long-range sequence pattern information of RNA via the pseudo k-tupler nucleotide composition (PseKNC) approach.

http://lin.uestc.edu.cn/server/iRNA-Methyl

iRNA (m6A)-PseDNC [130]

It uses the Euclidean distance-based method and pseudodinucleotide composition to identify m6A sites in the Saccharomyces cerevisiae (yeast) genome.

http://lin-group.cn/server/iRNA (m6A)-PseDNC.php

m6Acomet [131]

It is based on the RNA co-methylation network comprising 339,158 putative gene ontology functions associated with 1,446 identified human m6A sites.

http://www.xjtlu.edu.cn/biologicalsciences/m6acomet

WHISTLE [132]

It integrates 35 genome-derived and conventional sequence-derived features. It enable direct queries of predicted RNA-methylation sites, their putative functions, and their associations with other methylation sites or genes.

http://whistle-epitranscriptome.com

pRNAm-PC [133]

It predicts m6A sites in RNA sequences based on physicochemical properties. RNA sequence samples are expressed by pseudodinucleotide composition (PseDNC).

http://www.jci-bioinfo.cn/pRNAm-PC

TargetM6A [134]

It identifies m6A sites from RNA sequences via position-specific nucleotide propensities (PSNP) and a support vector machine (SVM).

http://csbio.njust.edu.cn/bioinf/TargetM6A

AthMethPre [135]

It trains the SVM classifier using the positional flanking nucleotide sequence and the position-independent k-mer nucleotide spectrum to predict m6A sites in Arabidopsis thaliana.

http://bioinfo.tsinghua.edu.cn/AthMethPre/index.html

RNAMethPre [136]

It predicts m6A sites by integrating multiple mRNA features and training the SVM classifier in mammalian mRNA sequences.

http://bioinfo.tsinghua.edu.cn/RNAMethPre/index.html