Recursively partitioned mixture model clustering of DNA methylation data using biologically informed correlation...

Area of Science:

• Epigenetics
• Genomics
• Computational Biology

Background:

• DNA methylation is a key epigenetic mechanism linked to diseases and exposures.
• Existing clustering methods for DNA methylation data often ignore biological relationships and make unrealistic assumptions.
• Advances in microarray resolution necessitate improved analytical approaches.

Purpose of the Study:

• To develop a modified recursively partitioned mixture model (RPMM) that incorporates genomic proximity of CpG loci.
• To enhance the modeling of correlation structures for DNA methylation data clustering.
• To improve the accuracy and biological relevance of clustering DNA methylation profiles.

Main Methods:

• Modification of the recursively partitioned mixture model (RPMM).
• Integration of genomic proximity information to define correlation structures.
• Application of the enhanced RPMM to simulated and real DNA methylation datasets.

Main Results:

• The modified RPMM demonstrated improved goodness-of-fit compared to existing methods.
• Clustering consistency was enhanced by integrating biologically informative correlation structures.
• The approach successfully detected biologically meaningful clusters in methylation data.

Conclusions:

• Integrating genomic proximity into clustering models significantly improves DNA methylation data analysis.
• The enhanced RPMM offers a more robust and biologically informed approach for identifying disease-associated epigenetic patterns.
• This method is crucial for leveraging high-resolution methylation data in biological and medical research.