metilene³: identifying DMRs across multiple conditions with auto-classification.

Area of Science:

• Epigenetics and Genomics
• Computational Biology
• Bioinformatics

Background:

• DNA methylation is a key epigenetic regulator of genome function.
• Identifying differentially methylated regions (DMRs) is crucial for understanding gene regulation.
• Existing DMR detection methods often require predefined sample conditions, limiting discovery in unlabeled or complex datasets.

Purpose of the Study:

• To develop a rapid, multi-condition method for detecting DMRs in DNA methylation data.
• To enable both supervised and unsupervised analysis of epigenetic patterns.
• To facilitate the discovery of new regulatory elements and sample stratifications.

Main Methods:

• Introduced metilene3, a novel method for DMR detection.
• Implemented both supervised (user-defined labels) and unsupervised (autonomous clustering) modes.
• Utilized genome segmentation based on multiple pairwise methylation difference signals for analysis.

Main Results:

• metilene3 accurately detects DMRs and robustly clusters samples across simulated and human datasets.
• The method demonstrated potential in revealing new regulatory elements and sample stratifications.
• In a pancreatic tissue dataset, metilene3 identified DMRs linked to key transcription factors in pancreatic cancer development.

Conclusions:

• metilene3 offers a fast and interpretable framework for exploring heterogeneous methylomes.
• The method supports the discovery of novel epigenetic patterns in complex biological and clinical data.
• metilene3 can aid in understanding epigenetic relationships and sample heterogeneity.