MODEL-BASED DISTANCE EMBEDDING WITH APPLICATIONS TO CHROMOSOMAL CONFORMATION BIOLOGY.

Area of Science:

• Genomics
• Biotechnology
• Computational Biology

Background:

• High-throughput chromosome conformation capture (Hi-C) technologies provide genome-wide interaction data.
• These interaction signals are inherently noisy, complicating the elucidation of true chromosomal organization.
• Accurate 3D genome structures are crucial for understanding gene regulation and function.

Purpose of the Study:

• To introduce a novel model-based distance embedding (MDE) framework.
• To efficiently recover accurate Euclidean distance matrices from noisy Hi-C observations.
• To reveal the spatial organization of chromosomes with improved fidelity.

Main Methods:

• Developed a general probabilistic modeling framework linking data properties to distance recovery.
• Applied MDE to simulated data mimicking chromosomal helix structures and random movements.
• Validated MDE performance on real Hi-C data from human and mouse cells.

Main Results:

• MDE framework successfully recovers underlying chromosomal conformations from noisy data.
• Numerical experiments demonstrate MDE's effectiveness in reconstructing known structures.
• Applications to real Hi-C data show practical utility and validate against benchmarks.

Conclusions:

• The MDE framework offers a robust method for elucidating 3D genome organization from noisy Hi-C data.
• This approach enhances our ability to understand chromosomal spatial structures and their biological implications.
• MDE provides a valuable tool for analyzing high-throughput conformation capture data.