Multi-scale phase separation by explosive percolation with single-chromatin loop resolution.

Area of Science:

• Genomics and Molecular Biology
• Biophysics and Computational Biology

Background:

• Human DNA (2m) is highly compacted within the 10μm nucleus via chromatin fiber folding.
• This folding creates hierarchical structures: chromosomal territories, compartments, Topologically Associating Domains (TADs)/Chromatin Contact Domains (CCDs), and loops.

Purpose of the Study:

• To propose dynamical models of human genome folding into hierarchical structures.
• To simulate genome folding trajectories in various human cell lines (lymphoblastoid, stem, fibroblast).

Main Methods:

• Modeling chromosomes as graphs with CTCF loops as edges, employing explosive percolation theory.
• Simulating folding by adding loops based on network properties, frequencies, compartmentalization, or epigenomic features.
• Integrating the Loop Extrusion Model (LEM) and Linear Discriminant Analysis for an order parameter.

Main Results:

• A biophysical pseudo-time genome folding model guided by a single order parameter.
• Chromatin phase separation into topological domains and compartments observed at critical contact numbers.
• Consistent finding: at least 80% of loops are necessary for 3D chromatin condensation across cell types.

Conclusions:

• The in-silico model integrates 3D genome interaction data with phase separation theory.
• It enables modeling of event-based dynamics for chromatin loop formation and folding trajectories.
• Provides a framework for understanding the physical principles governing genome organization.