Modeling Ensembles of Genomes: An Energy Landscape Perspective

Abstract

Abstract The three dimensional organization of the genome plays a critical role in gene regulation and cellular function. Unlike folded globular proteins, chromosomes exhibit high structural variability and must be described as ensembles of dynamic conformations rather than as single stable states. This chapter explores the physical principles of genome architecture through the lens of generalized Energy Landscape Theory. We describe the application of the maximum entropy principle to construct data driven, minimally biased restraint potentials from experimental observables such as Hi-C contact maps. Building on this framework, we discuss the Minimal Chromatin Model (MiChroM), which captures key physical interactions, including A/B compartmentalization and the ideal chromosome lengthwise compaction term, to simulate structural ensembles of chromosomes and their spatiotemporal dynamics. We also highlight recent advances that integrate artificial intelligence with polymer physics. Specifically, we discuss TECSAS, a deep learning Transformer architecture that predicts 3D structural annotations directly from 1D epigenetic marks, thereby bridging the gap between sequence and architecture without relying on experimental Hi-C data. Combined with high performance computational tools such as OpenMiChroM, these integrated multiscale approaches provide a comprehensive physical framework for investigating nucleome organization across cell types and cell cycle stages.