Utilizing networks for differential analysis of chromatin interactions

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Chromatin conformation capture with high-throughput sequencing (Hi-C) is crucial for mapping 3D genome structure.
• Detecting changes in chromatin interactions between different conditions is essential for understanding genome regulation.
• Existing methods for analyzing differential interactions in Hi-C data have limitations.

Purpose of the Study:

• To develop novel network-based computational approaches for identifying differentially interacting genomic regions from Hi-C data.
• To propose a normalization strategy for improving comparability across multiple Hi-C experiments.
• To evaluate the performance of new methods against existing ones and assess their utility with real and simulated data.

Main Methods:

• Development of two network-based methods (local and global connectivity) to detect differential chromatin interactions.
• Implementation of a normalization strategy based on network topological properties for multi-experiment Hi-C data.
• Comparative analysis of the proposed methods against existing approaches using simulated and real Hi-C datasets.

Main Results:

• The proposed local and global network methods significantly outperform two existing methods for detecting differential genomic interactions.
• A normalization strategy based on network topology improves the performance of the proposed methods.
• The local method is more effective for simulated data, while the global method performs better on real Hi-C data.

Conclusions:

• The novel network-based methods provide a robust framework for identifying differentially interacting genomic regions in Hi-C data.
• The proposed normalization strategy enhances data comparability and analytical power.
• These methods offer valuable tools for researchers studying genome organization and regulation.