Sequence-based multiscale modeling for high-throughput chromosome conformation capture (Hi-C) data analysis.

Area of Science:

• Genomics and Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Understanding chromosome hierarchical structures is crucial for deciphering gene regulation.
• Existing methods for analyzing chromosome conformation, such as Hi-C data, often struggle to differentiate between global and local interaction patterns.
• Distinguishing between spatial and sequential genomic distances is key to accurate modeling.

Purpose of the Study:

• To introduce sequence-based multiscale models (SeqMMs) for advanced biomolecular data analysis.
• To investigate the distinct roles of global and local clustering in modeling chromosome structures.
• To enhance the analysis of genomic compartments and topological associated domains (TADs).

Main Methods:

• Employed spectral clustering for sequence-based multiscale modeling.
• Differentiated between global scale clustering (optimizing spatial distances) and local scale clustering (optimizing genomic distances).
• Proposed and applied two SeqMMs to Hi-C data for analyzing chromosome hierarchical structures.

Main Results:

• Genomic compartments are primarily determined by global scale information, showing resilience to the removal of local interactions up to 10 Mb.
• TAD boundary prediction is sensitive to small-scale variations at smaller sequential scales but stabilizes above a size-related threshold.
• SeqMMs demonstrate greater robustness and reliability compared to previous spectral clustering models.

Conclusions:

• SeqMMs provide a more accurate and reliable framework for analyzing chromosome hierarchical structures.
• The findings clarify the distinct contributions of global and local genomic interactions to chromosome organization.
• TAD boundaries identified by both methods align with local minima of a TAD summation function.