Table 2 Methods and tools for predicting m7G modification
From: N7-methylguanosine modification in cancers: from mechanisms to therapeutic potential
Tool | Function | Website | References |
|---|---|---|---|
m7GHub | A comprehensive online platform for decoding m7G positions, regulatory mechanisms, and pathogenesis within mRNA | [87] | |
m7GDisAI | A tool for exploring m7G-associated diseases, providing a list of the top 20 predicted m7G sites related to 177 diseases, along with detailed information on specific m7G-disease association | [101] | |
iRNA-m7G | Identifies m7G modification sites in the human transcriptome using a feature fusion strategy that combines sequence-based and structure-based characteristics | [102] | |
XG-m7G | Differentiates m7G modification sites by applying the XGBoost algorithm in combination with six sequence encoding schemes | [103] | |
m7G-IFL | A machine learning-based tool that uses an iterative feature representation algorithm to accurately identify m7G modification sites | [104] | |
m7G-DPP | Encodes RNA sequences by leveraging dinucleotide physicochemical properties and extracts features like Pearson correlation, dynamic time warping, and distance correlation. Utilises LASSO feature selection and an SVM classifier to identify m7G modification sites | https://figshare.com/articles/online_resource/m7G-DPP/15000348 | [105] |
m7GPredictor | A machine learning model designed to predict internal m7G modification sites in RNA based on sequence features | [106] | |
THRONE | A computational predictor that integrates multiple sequence-based features with machine learning classifiers to accurately identify m7G sites in the human genome | [107] | |
SVM-based model | Utilises optimal sequence features to predict n7-methylguanosine sites in human RNA | [108] | |
m7GHub V2.0 | An updated resource collection for m7G modifications across various RNA types and 23 species, containing 430,898 predicted m7G sites identified using NGS and ONT technologies | [109] | |
TMSC-m7G | A prediction model based on variational autoencoders and contrastive learning for accurate identification of m7G sites | [110] | |
m7GRegpred | A bioinformatics framework integrating multiple features to predict substrates of m7G modification enzymes and readers | [111] | |
HN-CNN | A convolutional neural network-based model designed to predict disease associations with m7G sites within a heterogeneous network | – | [112] |
BRPCA | A computational method that predicts potential associations between m7G sites and various diseases using advanced predictive techniques | – | [113] |
BERT-m7G | A deep learning-based model using the BERT architecture to identify m7G modification sites in RNA sequences with high accuracy | – | [114] |
Moss-m7G | An interpretable, motif-based deep learning method for predicting m7G modification sites in RNA sequences | – | [115] |