RepliSage: a stochastic graph-based framework for 3D chromatin modeling across the cell cycle

Understanding chromatin dynamics across the cell cycle is crucial, as the structural transitions of chromosomes are fundamental to processes including transcriptional regulation, DNA replication, and faithful chromosome segregation. Although chromatin undergoes extensive reorganization throughout the cell cycle, no existing biophysical model describes its transitions from G1 through S and G2 to the completion of mitosis. To address this limitation, we present RepliSage, a multi-scale framework that integrates three fundamental processes shaping chromatin architecture: DNA replication, loop extrusion, and compartmentalization. In our model, replication forks are modeled as dynamic barriers that interact with loop extrusion factors, altering chromatin architecture during S phase. The framework integrates three complementary components: (i) replication fork progression simulated from single-cell replication timing data, (ii) Monte Carlo modeling of loop extrusion and epigenetic state transitions, and (iii) 3D reconstruction in OpenMM. Unlike previous approaches, RepliSage captures chromatin dynamics across the full cell cycle. In G1, random loop extrusion dominates; during S phase, replication forks interact with extrusion factors; and in mitosis, condensins drive long-range loop formation, facilitating chromosome segregation and polymer compaction. Chromatin is represented as a dynamic graph whose node states and connectivity evolve continuously over time. By tuning parameters across phases, RepliSage reproduces known structural transitions and provides a mechanistic platform to investigate how replication stress perturbs genome organization. The model was extensively validated using both publicly available and proprietary datasets. To our knowledge, this is the first framework to dynamically couple DNA replication, loop extrusion, and compartmentalization throughout the entire cell cycle.