Inferring protein structure and dynamics from simulation and experiment
- Kyle A. Beauchamp.
- Aug. 2013.
- Physical description
- online resource (xii, 151pages) : illustrations (some color)
- Beauchamp, Kyle A.
- Das, Rhiju. thesis advisor (primary).
- Harbury, Pehr thesis advisor.
- Martinez, Todd J. (Todd Joseph), 1968- thesis advisor.
- Pande, Vijay. thesis advisor (primary).
- Stanford University. Biophysics Program.
- Stanford University. Committee on Graduate Studies. degree grantor.
- Includes bibliographical references (p. 133-151). 170 refs.
- An atomic-scale understanding of biological molecules remains a grand challenge for the physical and biological sciences. Here, I describe how molecular dynamics simulations can be used to directly connect to biophysical experiments. I first describe the use of Markov state models to connect simulated and measured protein kinetics, allowing studies of protein folding at the atomic scale. I then introduce the use of NMR measurements, such as chemical shifts and scalar couplings, for the evaluation of molecular dynamics force field quality. Finally, I propose a new statistical technique that can be used to combine both simulation and experiment into accurate models of conformational ensembles. Such models are shown to be free of force field bias and can be used to investigate the structural and equilibrium properties of biomolecules. In sum, the present work demonstrates how statistically-sound methods of inference can forge a direct connection between simulation and experiment.
- Publication date
- Submitted to the Department of Biophysics and the Committee on Graduate Studies of Stanford University.
- Thesis (Ph.D.)--Stanford University, 2013.