Theoretical studies of chromosome folding as a consequence of molecular motor activity

In this thesis we develop theoretical models of how the molecular motor condensin folds the mitotic chromosome. In chapter 2 we simple theoretical models to discuss that the protein condensin must be a molecular motor. In chapter 3 we develop a theory for single condensin molecules. By fitting our theory to available experimental data, we predict the ATP hydrolysis rate and the magnitude of structural fluctuations in condensin. In chapter 4 we extend our single molecule theory of condensins, and show that multiple condensin molecules can dynamically come together to create fold the mitotic chromosome structure. We also use a data-driven model to show that mitotic chromosomes are helical with random helix perversions. In chapter 5, we provide a theoretical description of reverse stepping of condensin and predict that one of the states during the loop-extrusion process must involve a catch-bond like state. In chapter 6, we take a turn towards more a general polymer problem. We demonstrate that a deep neural network can approximately reconstruct the structure of a polymer from its contact map. Overall this thesis has derived theoretical and computational principles of chromosome folding.