DeepCpG: accurate prediction of single-cell DNA methylation states using deep learning

By Christof Angermueller, Heather J. Lee, Wolf Reik, Oliver Stegle. 2016

488 citations recorded by the source corpus.

Recent technological advances have enabled DNA methylation to be assayed at single-cell resolution. However, current protocols are limited by incomplete CpG coverage and hence methods to predict missing methylation states are critical to enable genome-wide analyses. We report DeepCpG, a computational approach based on deep neural networks to predict methylation states in single cells. We evaluate DeepCpG on single-cell methylation data from five cell types generated using alternative sequencing protocols. DeepCpG yields substantially more accurate predictions than previous methods. Additionally, we show that the model parameters can be interpreted, thereby providing insights into how sequence composition affects methylation variability.

Subfields: Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

Opportunities and obstacles for deep learning in biology and medicine (2017, background): “[216] trained a neural network to predict the .”
Deep Learning in Pharmacogenomics: From Gene Regulation to Patient Stratification (2018, background): “17 Other recent applications of deep learning models to prediction of regulatory elements and their interactions with the state-of-the-art performance include enhancer prediction [76-78]; classification of gene expression using histone modification data as input [60]; prediction of DNA methylation s”
Multimodal Large Language Models in Health Care: Applications, Challenges, and Future Outlook (2024, unknown): “Finally, omics encoders (eg, DeepVariant [62], Basenji [63], and DeepCpG [64]) focus on identifying genetic markers or patterns associated with specific diseases, aiding in the development of targeted therapies.”
Gene2vec: gene subsequence embedding for prediction of mammalian N6-methyladenosine sites from mRNA (2018, background): “…et al. 2016; Xu et al. 2017), drug discovery (Stephenson et al. 2018), protein subcellular localization (Almagro Armen-teros et al. 2017; Wei et al. 2018a), protein function prediction (Cao et al. 2017), and single-cell DNA methylation states (Angermueller et al. 2017; Xu and Zhou 2018).”
SBSM-Pro: support bio-sequence machine for proteins (2023, unknown): “Hybrid predictive models combining CNNs and RNNs are currently popular and have been applied in various computational biology areas, including DNA methylation [29], chromatin accessibility [30], and noncoding RNAs [31].”
DNA methylation aging clocks: challenges and recommendations (2019, background): “Computational techniques such as imputation may help with these absent values [175].”
DeePromoter: Robust Promoter Predictor Using Deep Learning (2019, methodology): “In addition, it was successfully applied in biological problems such as DeepBind (Alipanahi et al., 2015), DeepCpG (Angermueller et al., 2017), branch point selection (Nazari et al., 2018), alternative splicing sites prediction (Oubounyt et al., 2018), 2’-Omethylation sites prediction (Tahir et…”
A Survey of Deep Learning: Platforms, Applications and Emerging Research Trends (2018, methodology): “[15] applied deep learning methods to predict methylation states in individual cells.”
Applications of Deep Learning and Reinforcement Learning to Biological Data (2017, methodology): “used DNN based method to estimate DM state by predicting the changes in single nucleotides and uncovering sequence motifs [85].”
Deep learning in bioinformatics: introduction, application, and perspective in big data era (2019, methodology): “For recurrent neural networks, the more advanced architectures help people deal with the gradient vanishing or explosion issue and accelerate the execution of RNN.”