📝 SummarXiv

Abstract

Understanding chromatin dynamics across multiple spatiotemporal scales requires models that reconcile biological specificity with physics-based interactions and computational tractability. We present a modular, recognition-enabled ultra-coarse-grained (UCG) framework that captures both histone-DNA and histone-histone interactions using site-specific, off-center Recognition potentials. These recognition sites, combined with generic attractive and repulsive terms, encode directional and stoichiometrically faithful assembly rules. Benchmark simulations demonstrate that this scheme robustly drives the self-assembly of geometrically correct histone octamers and enables stable nucleosome formation. The model also supports tunable resolution, allowing simplification of intra-octamer, nucleosomal, or fiber-level structures depending on the biological question. This flexibility is especially useful for exploring chromatin reorganization driven by epigenetic regulation. While developed with chromatin in mind, our framework generalizes to other multivalent assemblies governed by molecular recognition.