HiCRep.py: fast comparison of Hi-C contact matrices in Python

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Hi-C is a key assay for studying genome-wide 3D chromatin organization.
• Comparing Hi-C contact matrices is crucial for assessing experimental reproducibility and sample relationships.
• The existing R implementation of HiCRep faces performance limitations with high-resolution or large single-cell Hi-C data.

Purpose of the Study:

• To introduce a more efficient Python implementation of the HiCRep algorithm.
• To demonstrate the performance improvements and utility of the new Python implementation.

Main Methods:

• Developed HiCRep.py, a Python version of the HiCRep algorithm.
• Evaluated HiCRep.py's speed and memory consumption compared to the R implementation.
• Applied HiCRep.py to assess reproducibility and reveal sample structures in Hi-C datasets.

Main Results:

• HiCRep.py exhibits significantly faster run times and lower memory usage than the R version.
• The Python implementation effectively distinguishes between replicate and non-replicate Hi-C samples.
• HiCRep.py successfully identifies cell type structures within collections of Hi-C data.

Conclusions:

• HiCRep.py offers a computationally efficient solution for analyzing Hi-C data similarity.
• This implementation overcomes the limitations of the R version for high-resolution and single-cell Hi-C data.
• HiCRep.py is a valuable tool for robust analysis of 3D genome organization.