Inferring Functional Epigenetic Modules by Integrative Analysis of Multiple Heterogeneous Networks

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Gene expression and methylation are fundamental biological processes crucial for understanding cancer.
• Existing methods often fail to fully capture the complex relationships and heterogeneity within these datasets.
• Integrating gene expression and methylation data is essential for uncovering cancer's underlying mechanisms.

Purpose of the Study:

• To define and discover epigenetic modules by integrating gene expression and methylation data.
• To develop a novel algorithm, Ep-jNMF, that addresses data heterogeneity.
• To improve the accuracy and biological relevance of identified modules for cancer research.

Main Methods:

• Constructed separate gene co-expression and co-methylation networks to handle data heterogeneity.
• Defined epigenetic modules as common modules across multiple networks.
• Developed and applied a non-negative matrix factorization-based algorithm (Ep-jNMF) for joint clustering.

Main Results:

• Ep-jNMF demonstrated higher accuracy compared to baseline methods on artificial data.
• The algorithm identified biologically meaningful epigenetic modules.
• These modules showed potential in predicting cancer subtypes.

Conclusions:

• Ep-jNMF is an efficient tool for integrating gene expression and methylation data.
• The identified epigenetic modules offer valuable insights into cancer biology.
• This approach advances the discovery of cancer-related genomic patterns.