Molecular Dynamics Simulations of DNA Hybridization and Dynamic Force Spectroscopy

Abstract

A ubiquitous problem in chemistry is determining the energy change of moving from one state to another. Often these states are separated by a large energy barrier which can prevent the observation of the equilibrium distribution on a reasonable timescale. Techniques have been developed to accelerate the dynamics of these systems and determine the equilibrium energies. In this thesis, we examine two systems using non-equilibrium techniques. We first studied DNA hybridization which involves two single-stranded DNA interacting through hydrogen bonds between their bases. We investigated DNA hybridization using coarse-grained molecular dynamics simulations and by creating a Markov State Model from the simulation data. We found that the mechanism of hybridization to be nucleation and zipping up. We also found the transition states contain one or two bonds. If the strands have three or more bonds, they are very likely to proceed to full hybridization. We then studied dynamic force spectroscopy, a technique in which systems are biased out of equilibrium using external forces. We studied five models of dynamic force spectroscopy and found many differences between them. Overall we found the Hummer-Szabo intermediate model does the best. However these models tend to focus on the mean rupture force and largely do not account for the variance or distribution of forces. We found the variance to be less sensitive to the pulling protocol than the mean. We also found a relationship between the rupture force and the work performed on the system until rupture which requires furtherstudy.