Protein folding and prediction using energy landscape concepts

Over the last 10--15 years a general understanding of the process of protein folding has emerged from statistical mechanics. The study of protein folding kinetics based on energy landscape ideas has been of direct benefit to protein structure prediction. The funneled energy landscape incorporates both the principle of minimal frustration, which produces a bias toward that native state and topographic roughness, which simulates local energy minima. This thesis takes advantage of these energy landscape concepts to predict protein structures. Results from recent blind predictions will be shown. The same ideas are extended to describe the progress of different structure prediction energy functions, and explore the foldability of naturally occurring protein sequences.