The genomes of higher eukaryotes are organized into complex structures spanning multiple length scales. At the largest scale, chromosomes can be easily observed under a standard light microscope. At the smallest scale, nucleosomes—comprising DNA wrapped around a core of eight histone proteins—represent the fundamental units of chromatin. These nucleosomes further organize into chromatin fibers, leading to a highly compact genome configuration. This hierarchical packaging tightly regulates the accessibility of regulatory DNA sequences (genes) by DNA-binding proteins (DBPs). The spatial positioning of genes within the nucleus is intricately linked to gene regulation, warranting detailed investigation to unravel the underlying mechanisms. However, this task is challenging due to the inherently multiscale nature of chromatin and our limited understanding of how chromatin organization influences gene recognition by transcription factors.

To address these challenges, we employ multiscale polymer simulations alongside advanced machine learning techniques. Our goal is to bridge the current knowledge gaps that hinder our understanding of the molecular mechanisms of gene regulation under nuclear constraints. We aim to decode the complexities of the 3D chromatin landscape associated with transcription, aspects of which have not been adequately captured—or in some cases, not captured at all. Our research seeks to provide novel insights into the structure-function relationships within the genome, with significant implications for fundamental biology and medical science, particularly in understanding developmental disorders and cancers arising from gene misexpression.