DiffVar: a new method for detecting differential variability with application to methylation in cancer and aging

Area of Science:

• Genomics
• Epigenetics
• Bioinformatics

Background:

• DNA methylation is crucial for development and significantly altered in cancer.
• The Illumina HumanMethylation450 BeadChip is widely used for genome-wide CpG site profiling.
• Existing methods may not adequately address differential variability analysis.

Purpose of the Study:

• To introduce DiffVar, a novel statistical method for testing differential variability in DNA methylation data.
• To provide a robust and flexible tool for analyzing complex experimental designs.
• To enhance the analysis of epigenomic data in cancer and aging research.

Main Methods:

• DiffVar utilizes an empirical Bayes model framework.
• The method is designed to be robust to outliers.
• It accommodates diverse experimental designs for methylation analysis.

Main Results:

• DiffVar was successfully applied to The Cancer Genome Atlas (TCGA) datasets.
• The method was also validated on an aging dataset.
• Demonstrated utility in identifying differential methylation variability across different biological contexts.

Conclusions:

• DiffVar offers a powerful new approach for analyzing differential DNA methylation variability.
• The method is implemented in the missMethyl Bioconductor R package for accessibility.
• Facilitates deeper insights into epigenomic alterations in disease and aging.