MethylMix: an R package for identifying DNA methylation-driven genes

Area of Science:

• Epigenetics
• Genomics
• Computational Biology

Background:

• DNA methylation regulates gene transcription and is implicated in carcinogenesis.
• Aberrant DNA methylation (hyper- and hypomethylation) can deregulate gene expression in various diseases.
• High-throughput assays generate extensive genome-wide methylation data, but tools for functional analysis are limited.

Purpose of the Study:

• To develop an algorithm for identifying disease-specific, transcriptionally relevant DNA methylation changes.
• To address the lack of computational tools for analyzing genome-wide methylation data in disease contexts.

Main Methods:

• Developed MethylMix, an R package utilizing a beta mixture model to identify methylation states.
• Introduced a novel 'Differential Methylation value' (DM-value) to quantify methylation state differences from normal.
• Integrated gene expression data to identify methylation changes predictive of transcriptional alterations.

Main Results:

• MethylMix algorithm successfully identifies disease-specific hyper- and hypomethylated genes.
• The DM-value provides a metric for assessing methylation state deviations.
• Analysis incorporating gene expression data pinpoints methylation changes that functionally impact gene transcription.

Conclusions:

• MethylMix is a valuable tool for identifying functionally relevant DNA methylation patterns in diseases.
• The algorithm aids in understanding the role of epigenetic alterations in disease pathogenesis.
• MethylMix facilitates the discovery of transcriptionally predictive methylation states from high-throughput data.