Deep Learning in Epigenomics Literature

DNA Methylation Prediction

  • 39742481 (Zhou, Han, 2025 Sci Adv) - INTERACT: CNN+Transformer to predict DNAm levels from 2k sequence, and identify in silico variants that have high effects in prediction

  • 40215314 (Yan, Guo, 2025 Sci Adv) - DiffuCpG: Diffusion model to predict DNA methylation levels

  • 39986279 (Zhou, Han, 2025 Cell Genomics) - scMeFormer: A transformer-based deep learning model for imputing DNAm states at each CpG site in single cells

  • 38961062 (Stanojevic, Sikic, 2024 Nat Comm) - Rockfish: Transformer-based model for nanopore sequencing data

  • 38365920 (Ahsan, Wang, 2024 Nat Comm) - DeepMod2: Bidirectional long short-term memory (BiLSTM) model and a Transformer model for methylation detection in Nanopore sequencing data

  • 39196755 (Ma, Chen, 2024 Bioinformatics) - Deep5hmC: Integrates both the DNA sequence and epigenetic features such as histone modification and chromatin accessibility to predict genome-wide 5hmC modification

  • 37489753 (Zeng, Huson, 2023 Gigascience) - MuLan-Methyl: Transformer based method to predict N6-adenine, N4-cytosine, and 5-hydroxymethylcytosine

  • 37647650 (Deng, Fan, 2023 Bioinformatics) - GraphCpG: Transformer based method to predict N6-adenine, N4-cytosine, and 5-hydroxymethylcytosine

  • 34718418 (Waele, Waegeman, 2022 Bioinformatics) - CpG Transformer: Transformer imputation of single-cell methylomes

  • 30994904 (Ni, Wang, 2019 Bioinformatics) - DeepSignal: CNN to detect DNA methylation states from Nanopore sequencing reads

  • 31164644 (Liu, Wang, 2019 Nat Comm) - DeepMod: Bidirectional recurrent neural network to detect DNA methylation 5mC and 6mA

  • 28395661 (Christof, Stegle, 2017 Genome Biology) - DeepCpG: Bidirectional gated recurrent network to predict single cell DNAm levels

  • 28334830 (Zeng, Gifford, 2017 Nuleic Acids Res) - CpGenie: Deep convolutional neural network predicts non-coding variants that modulate DNA methylation and the methylation status of a CpG site from the flanking sequence at a single-nucleotide sensitivity

  • 26797014 (Wang, Wang, 2016 Sci Rep) - DeepMethyl: Stacked denoising autoencoders from Hi-C and sequence patterns to predict methylation state

DNA Methylation Biomarkers

  • 39824848 (Jeong, Lutsik, 2025 Nat Comm) - MethylBERT: Bidirectional Encoder Representations from Transformers to derive tumour purity estimate

  • 334245239 (Li, Wang, 2021 Brief Bioinform) - DISMIR: Predict the source of individual reads in plasma cfDNA WGBS

  • 32183722 (Levy, Christensen, 2020 BMC Bioinformatics) - MethylNet: Encoder-Decoder for EWAS traits

  • 28785873 (Korfiatis, Ercikson, 2017 J Digit Imaging) - Predicting MGMT methylation status using ResNet from MRI imaging

Foundation Models

  • 39574641 (Ying, Gladyshev, 2024 bioRxiv) - MethylGPT: A transformer-based foundation model trained on 226,555 human methylation profiles

  • Not available (Camillow, Wang, 2024 bioRxiv) - CpGPT: A transformer-based foundation model trained on 1,500 DNA methylation datasets