Assessing and Improving Computational Models of Protein Thermodynamics and Kinetics

Abstract

The purpose of this thesis is to rigorously assess and improve computational models of protein thermodynamics and kinetics. The first part consists of computational ddG prediction; we explore the performance of protocols which sample an increasing diversity of conformations and examine their abilities to recapitulate both changes in free-energy as well as changes in structure. Application of the improved ddG prediction protocol yields high performance on independent benchmarks as well as success in two blind applications. The second portion consists of assessing and improving discrete computational models of protein kinetics. The space accessed by a folding macromolecule is vast, and how to best project computer simulations of protein folding trajectories into an interpretable sequence of discrete states is an open research problem. There are numerous alternative ways of associating individual configurations into collective states, and in deciding on the number of such clustered states there is a trade-off between human interpretability (smaller number of states) and accuracy of representation (larger number of states). Here we introduce measure for assessing alternative discrete state models of protein folding and assess different methods of defining discrete states. Using the most predictive representation to study the folding transitions of the WW domain in very long molecular dynamics simulations we identify new states and transitions. The methods developed here should be generally useful for investigating the thermodynamics and kinetics of protein structure.