essHi-C: essential component analysis of Hi-C matrices

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Hi-C matrices are crucial for studying genome folding, including territorial organization, compartments, and topological domains.
• Hi-C assays capture a wide range of genomic distances, leading to stochastic background noise that convolves specific interaction patterns with non-specific ones.

Purpose of the Study:

• To introduce essHi-C, a novel computational method designed to separate specific genomic interaction signals from background noise in Hi-C data.
• To enhance the analysis of genome architecture by isolating the essential component of Hi-C matrices.

Main Methods:

• Development and application of the essHi-C algorithm to process Hi-C matrices.
• Systematic comparisons of essHi-C with existing methods to evaluate its performance.

Main Results:

• essHi-C effectively isolates specific interaction patterns, improving clarity and robustness against sequencing depth variations in identifying topologically associating domains.
• The method enables unsupervised clustering of Hi-C experiments across different cell lines and accurately recovers cell-cycle phasing from single-cell Hi-C data.

Conclusions:

• essHi-C provides a powerful tool for discerning significant biological and physical features within Hi-C data by effectively removing non-specific background noise.
• The software package is publicly available, facilitating broader adoption and advancement in the field of genome organization studies.