Inferring 3D chromatin structure using a multiscale approach based on quaternions.

Area of Science:

• Genomics
• Molecular Biology
• Biophysics

Background:

• Understanding chromatin spatial organization is key to deciphering DNA regulation.
• Chromosome Conformation Capture (3C) techniques generate high-throughput contact data for DNA fragments.
• Existing methods attempt 3D genome reconstruction from 3C data, often by converting contacts to distances.

Purpose of the Study:

• To develop a novel 3D chromatin reconstruction technique that bypasses the distance conversion step.
• To integrate biochemical and biological constraints into a scalable geometrical chromatin model.
• To generate plausible chromatin configurations from experimental contact data.

Main Methods:

• Developed a reconstruction technique avoiding contact-to-distance conversion.
• Introduced a scalable geometrical chromatin chain model with biochemical/biological constraints.
• Employed simulated annealing with quaternion operators for efficient solution space exploration.

Main Results:

• Generated 3D chromatin configurations from human chromosome 1 Hi-C data.
• The method successfully integrates experimental contact data with prior biological knowledge.
• Reconstructions are compatible with both the input data and known biological features.

Conclusions:

• The novel method provides plausible 3D chromatin structures.
• Structural differences in reconstructions correlate with known functional genomic regions.
• This approach offers a robust way to study genome organization and function.