Detecting TAD-like domains from RNA-associated interactions.

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Topologically associated domains (TADs) are fundamental units of chromatin organization.
• RNA-chromatin and RNA-RNA interactions (RAIs) are increasingly recognized for their role in genome architecture.
• RAIs may delineate TAD-like domains (TLDs), suggesting a link between RNA biology and chromatin structure.

Purpose of the Study:

• To develop and validate an algorithm, SuperTLD, for detecting hierarchical TAD-like domains (TLDs) from RAI data.
• To compare the TLDs identified from RAIs with traditional TADs derived from Hi-C data.
• To investigate the epigenetic characteristics and regulatory implications of novel TLDs.

Main Methods:

• Development of SuperTLD, a domain detection algorithm that incorporates imputation for analyzing RAI datasets.
• Application of SuperTLD to four RAI datasets.
• Comparative analysis of TLDs and TADs using metrics for hierarchical similarity and fine structures.
• Enrichment analysis of epigenetic marks (CTCF, H3) and Gene Ontology (GO) terms at TLD boundaries and within domains.

Main Results:

• SuperTLD successfully identified TLDs from RAI data.
• TLDs exhibited moderate similarity to TADs in hierarchy (≥ 0.5312) and fine structures (≥ 0.8295).
• Novel TLDs showed significant enrichment of CTCF (0.6245 fold change) and H3 histone marks, with GO analysis revealing links to gene regulation at novel boundaries.

Conclusions:

• RAI data can be used to identify TLDs, complementing Hi-C based TAD identification.
• SuperTLD provides a novel approach to decode chromatin organization from RNA-centric interactions.
• TLDs, particularly novel ones, are associated with specific epigenetic marks and play a role in gene regulation.