Predicting the impact of non-coding variants on DNA methylation

Area of Science:

• Genomics
• Epigenetics
• Computational Biology

Background:

• DNA methylation is vital for gene expression and biological regulation.
• Predicting the impact of genetic variations on DNA methylation is challenging.
• Non-coding genetic variations' functional consequences remain poorly understood.

Purpose of the Study:

• To develop a sequence-based framework, CpGenie, for predicting DNA methylation changes due to sequence variation.
• To enable allele-specific DNA methylation prediction with single-nucleotide resolution.
• To aid in the interpretation of functional non-coding genetic variants.

Main Methods:

• Utilized a deep convolutional neural network to learn the DNA methylation regulatory code.
• Developed a sequence-based framework (CpGenie) for predicting DNA methylation.
• Applied CpGenie to predict methylation quantitative trait loci (meQTLs) and assess variant impact.

Main Results:

• CpGenie accurately predicts allele-specific DNA methylation at CpG sites.
• The framework enables precise prediction of meQTLs.
• CpGenie successfully prioritized validated Genome-Wide Association Study (GWAS) single nucleotide polymorphisms (SNPs).
• CpGenie aids in identifying expression quantitative trait loci (eQTLs) and disease-associated mutations.

Conclusions:

• CpGenie provides a powerful tool for predicting the functional impact of DNA sequence variation on DNA methylation.
• The framework enhances the identification and interpretation of regulatory non-coding variants.
• CpGenie facilitates a deeper understanding of genetic variation's role in gene regulation and disease etiology.