Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.605591
Title: Molecular dynamics and complexity analysis of molecular systems
Author: Jensen, C. H.
Awarding Body: University of Cambridge
Current Institution: University of Cambridge
Date of Award: 2010
Availability of Full Text:
Full text unavailable from EThOS. Please contact the current institution’s library for further details.
Abstract:
In this thesis, Complexity Analysis, which is defined as the use of Markov models and Computational Mechanics, is applied to Molecular Dynamics simulations of peptides. To achieve this, the trajectories from the Molecular Dynamics simulations are clustered into conformational states and by investigating the time series of these states, statistical models are constructed. A basic property of a Markov model is that the probability distribution of the subsequent states depends only on the current state and not the history. This has previously been used to develop a method for testing the model which is based on calculating and comparing eigenvalues for Markov models constructed at different time steps. Here, the method is applied to a simulation of the four residue peptide VPAL and it is found that the Markov model is accurate at a minimum time step of 100ps. The determination of the time step using this test is, however, subjective, so I have developed a method which is based on Computational Mechanics to determine the minimum time step at which the dynamics are Markovian. An important part of the application of a Markov model is the clustering of the Molecular Dynamics simulation into conformational states. The effect of varying the clustering of the simulation is investigated by calculating the mean first passage times between conformational states as the cluster boundaries are varied. It is found that the mean first passage times are sensitive to specific clustering, and to reduce the model sensitivity to variations in clustering, it is especially important to exclude sparsely populated states from the model. Finally, it is demonstrated that the folding time of a slow folding protein can be very sensitive to changes in the Markov model transition matrix. This implies that folding times calculated using Molecular Dynamics cannot meaningfully be compared to folding times obtained from experiments.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.605591  DOI: Not available
Share: