Random folding drives the emergence of topologically associating domains in chromatin three-dimensional structure

Domains are units of genome organization. Due to the seemingly irreconcilable difference between topologically associating domains (TADs) revealed by population-based biochemical studies and domains (sTADs) by single-cell imaging, finding a mechanism that simultaneously shape TADs and sTADs is challenging. Here we propose that TADs and sTADs are underlied by random folding of chromatin fiber heterogeneous in DNA density. On the hypothesis, we develop a model, termed RCHC, to yield chromatin structure ensemble from chromatin accessibility data. Calculated ensemble enables our hypothesis to be validated by population-based and single-cell experiments. Our simulation confirms the independence between domain and compartment structures in genome organization and shows that RCHC can predict the chromatin reorganization during differentiation. We mechanistically prove that genome is organized randomly with biases introduced by DNA-encoded information. One-Sentence Summary The TADs emergence is underlied by random folding of heterogeneous chromatin fiber carrying nucleosome occupancy information.