Modeling complex patterns of differential DNA methylation that associate with gene expression changes.

Area of Science:

• Genomics
• Epigenetics
• Bioinformatics

Background:

• DNA methylation alterations are implicated in various diseases, but their precise impact on gene expression remains unclear.
• Understanding the relationship between methylation changes and gene expression is crucial for disease mechanism elucidation.

Purpose of the Study:

• To develop and validate an integrative analysis tool, Methylation-based Gene Expression Classification (ME-Class), for predicting gene expression changes based on DNA methylation.
• To investigate disease-specific methylation patterns and their association with gene expression variations.

Main Methods:

• Developed ME-Class, a computational model capturing promoter methylation complexity to link methylation variation with expression change.
• Utilized whole-genome bisulfite sequencing and RNA-seq data from the Roadmap Epigenomics and Blueprint Epigenome Projects.
• Applied ME-Class to normal-tumor pairs from The Cancer Genome Atlas (TCGA).

Main Results:

• ME-Class significantly outperformed standard methods in predicting differential gene expression from methylation data.
• Analysis of hematopoietic cell types revealed that expression-associated methylation changes are more pronounced between distantly related lineages.
• Cancer-specific methylation changes associated with expression differed from tissue-specific changes, and ME-Class identified functionally relevant, reversible cancer-specific changes.

Conclusions:

• ME-Class is a powerful tool for identifying genes dysregulated by DNA methylation in disease.
• The findings suggest that transcriptional program changes may precede associated methylation changes during cell differentiation.
• ME-Class facilitates the discovery of disease-specific epigenetic alterations impacting gene expression.