TADpole

TADpole identifies and analyzes the hierarchical organization of topologically associating domains (TADs) within intra-chromosomal interaction matrices derived from Chromosome Conformation Capture (3C)-based techniques.

Key Features:

• Hierarchical TAD identification: Detects entire nested hierarchies of TADs rather than treating domains as isolated units.
• Input data: Operates on intra-chromosomal interaction matrices derived from Chromosome Conformation Capture (3C)-based techniques, including Capture Hi-C.
• Algorithms: Employs principal component analysis (PCA) combined with constrained hierarchical clustering to delineate significant levels of chromatin organization.
• Boundary enrichment detection: Identifies domain borders enriched for CTCF, components of the cohesin complex, and the histone mark H3K4me3.
• Functional state discrimination: Distinguishes activation states within domain bodies that are differentially enriched for H3K36me3 or H3K27me3.
• Robustness: Demonstrates robustness across data resolutions, normalization strategies, and sequencing depths.
• DiffT metric: Computes a DiffT score to detect significant topological differences between wild-type and genetically engineered organisms (e.g., mice) using Capture Hi-C maps.

Scientific Applications:

• Hierarchical chromatin architecture analysis: Map nested TAD organization in genomic regions from 3C-based interaction maps.
• TAD boundary characterization: Identify domain borders associated with architectural proteins (CTCF, cohesin) and the promoter-associated mark H3K4me3.
• Epigenetic state profiling within TADs: Differentiate domain bodies enriched for H3K36me3 versus H3K27me3 to infer activation or repression states.
• Comparative topology analysis: Detect topological differences between wild-type and genetically engineered organisms (e.g., mice) using the DiffT score on Capture Hi-C data.

Methodology:

Applies principal component analysis (PCA) followed by constrained hierarchical clustering to delineate TAD hierarchies and computes a DiffT score on Capture Hi-C maps to detect significant topological differences.