Polymer physics reveals a combinatorial code linking 3D chromosome architecture to 1D chromosome features.

The mammalian genome has a complex, far from random three-dimensional (3D) organization, intimately linked to vital functional processes. To understand the physical mechanisms underlying chromosome folding, polymer models from statistical physics and a variety of computational methods have been proposed. However, they usually cannot explain data at the length scale of full chromosomes. In this talk, I will show how our approach, that combines Machine Learning and Polymer Physics, can be easily extended genome-wide in order to obtain a set of polymer models that correctly explain 3D folding data of the whole genome. The obtained polymer models are validated by making predictions on the changes of the 3D structure caused by disease-linked genomic mutations and our predictions are confirmed by independent data from cells carrying such mutations. Finally, by exploiting such a complex organization, we developed a code to predict de novo the 3D structure of an independent set of chromosomes from only their 1D epigenetic marks. Overall, our results shed light on how 3D information is encrypted in 1D via the specific combinatorial arrangement of chromosome features.