Using motion planning to study protein folding pathways.

摘要: We present a framework for studying protein folding pathways and potential landscapes which is based on techniques recently developed in the robotics motion planning community. Our focus this work to study mechanism assuming we know native fold. That is, instead of performing fold prediction, aim issues related process, such as formation secondary tertiary structure, dependence pathway initial denatured conformation. uses probabilistic roadmap (PRM) have proven successful problems involving high-dimensional configuration spaces. A strength these methods their efficiency rapidly covering space without becoming trapped local minima. applied our PRM technique several small proteins (~60 residues) validated computed by comparing structure order paths known hydrogen exchange experimental results. An advantage over other simulation that it enables one easily efficiently compute from any starting state (known) This aspect makes approach ideal global properties protein's landscape, most are difficult simulate with methods. For example, study, different states sometimes share common portions when they close fold, moreover, appears largely independent Another feature distribution sampled conformations correlated and, particular, differentiate situations clearly forms first those obtained more directly. Overall, results applying very encouraging indicate promise not available.