FastHiC: a fast and accurate algorithm to detect long-range chromosomal interactions from Hi-C data

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Three-dimensional (3D) chromatin folding is crucial for gene transcription regulation.
• Hi-C technologies provide genome-wide measurements of pairwise chromatin interactions.
• Accurate detection of biologically meaningful chromatin interactions (peak calling) from Hi-C data remains a challenge.

Purpose of the Study:

• To develop a computationally efficient and accurate method for peak calling in Hi-C data.
• To address the computational limitations of previous Bayesian methods for 3D chromatin interaction analysis.

Main Methods:

• Developed FastHiC, a novel approach utilizing simulated field approximation.
• Approximated the joint distribution of hidden peak status using independent random variables for tractable computation.
• Improved upon a previous hidden Markov random field (HMRF)-based Bayesian method by modeling spatial dependencies.

Main Results:

• FastHiC accelerates the original Bayesian method by over five times.
• Achieved higher peak calling accuracy compared to previous methods in real data analysis.
• Demonstrated improved robustness and statistical power in detecting chromatin interactions.

Conclusions:

• FastHiC offers a significant advancement in analyzing Hi-C data for 3D chromatin conformation.
• The method provides a faster and more accurate solution for identifying functionally relevant chromatin interactions.
• Enables more efficient exploration of the relationship between chromatin folding and gene regulation.