Optimizing potentials for protein structure prediction, inverse protein folding and protein folding

Under appropriate physiological conditions, proteins fold into their biologically active native conformation. This dissertation has touched upon two problems related to protein folding: the protein structure prediction problem and the protein folding problem.;An accurate potential is the key to success in the protein structure prediction problem. For a single protein, the potential function can be optimized for predictive success by maximizing the energy gap between the correct structure and the ensemble of random structures relative to the distribution of the energies of these random structures (Z-score). Different averaging procedures have been proposed to deal with an ensemble of database proteins. Here a new approach, maximizing the average probability of success, is demonstrated to carry out this task. Even though the optimized potentials perform better than previously-used averaging procedures, the results show that the Z-score based optimization methods tend to underestimate the repulsive interactions due to an inherent tendency to over-stabilize the high-energy states. To lessen the bias, the distribution is weighted to suppress the high-energy state contribution to the Z-score calculation. Using a lattice model, the improved optimal potentials are both more accurate and more successful in predicting protein structures than those obtained by all other previous potential derivation on methods.;An alternative to protein structure prediction is the related "inverse protein folding" where one identifies the sequences in the database that fold into a given structure. Here a non-physical potential scheme is developed to optimize for this purpose. Using a lattice model of proteins, it is shown that the optimal potentials actually work better than the real potential.;To address the protein folding problem, protein folding has been modeled as diffusion on a free-energy landscape. This allows the diffusion equation to be used to study the impact of energy parameters on the folding dynamics. For marginally stable proteins, fastest folding is achieved when the non-specific interactions favoring compaction are strong, resulting in a high folding temperature. Such proteins fold by rapid collapse followed by slower accumulation of correct contacts.