metilene3

metilene3 identifies Differentially Methylated Regions (DMRs) across multiple conditions in both supervised and unsupervised modes, enabling Differentially Methylated Tree construction for epigenetic relationship inference and sample clustering.

Key Features:

• Multi-condition DMR detection: Detects DMRs across multiple conditions simultaneously, unlike methods limited to pairwise comparisons.
• Supervised and unsupervised operation: Operates with user-provided sample labels or autonomously clusters unlabeled samples.
• Auto-classification: Classifies samples without predefined group identities by segmenting the genome based on multiple pairwise methylation difference signals.
• Differentially Methylated Tree construction: Enables DMR-anchored inference of epigenetic relationships and sample clustering.
• Genome segmentation: Segments the genome based on multiple pairwise methylation difference signals to identify DMRs.

Scientific Applications:

• Epigenetic pattern discovery: Reveals new epigenetic patterns and regulatory elements in heterogeneous methylomes, particularly when group identities are unknown or uncertain.
• Sample stratification: Robustly clusters samples and identifies sample stratifications in complex biological and clinical datasets.
• Cancer epigenetics: Identifies DMRs enriched for key transcription factors involved in cancer development, such as NFKB-NFAT regulatory programs in pancreatic tissue.
• Clinical epigenetics: Applicable to clinical settings where sample group identities are unknown or uncertain.

Methodology:

metilene3 segments the genome based on multiple pairwise methylation difference signals to detect DMRs and enable sample classification.