|
Stochastic Roadmap Simulation (SRS) is a tool designed to efficiently estimate ensemble properties and other key properties (e.g., order of events) of molecular motion. Unlike previous approaches, SRS does not perform individual simulation runs. Instead, it pre-computes a roadmap by sampling the molecular conformation space at random and uses tools from Markov Chain Theory to compute properties over the many molecular pathways encoded in the roadmap. Our previous work has demonstrated the efficiency of this approach on protein folding and ligand-protein binding examples. But roadmaps were pre-computed by sampling conformation spaces uniformly. Our recent and current research is aimed at extending SRS to conformation spaces of higher dimensionality (>40-50) by designing and testing various non-uniform sampling strategies that put more samples in “interesting” subsets of the conformation spaces. This work is done in cooperation with Prof. Doug Brutlag (Biochemistry, Stanford U.) and Prof. David Hsu (Computer Science, Nat. Univ. of Singapore).