We present a virtual screening library to identify binding fragments. The idea is based on identifying patterns of interactions between ligand and receptor atoms for a database of complexes. This can be done by, first, applying Almost Delaunay tessellation to define interfacial atoms between ligands and their receptors. These interfacial atoms form the pattern of interaction for which both geometric and chemical features are memorized. Thus, given a particular binding pocket, we search the memorized patterns for ones have their receptor atoms matching (geometrically and chemically) any part of the particular binding pocket's atoms. For each match found, the ligand atoms corresponding to the matched receptor atoms in that pattern are identified. These ligand atoms form the binding fragments, which can then be used to search for molecules containing them. Or, they can be augmented using de novo drug design techniques to form new molecular entity. (Check http://www.unc.edu/~raed for Software). 

FragVLib: Fragment-based virtual screening library using geometric and chemical patterns of interactions at interface of ligand-receptor complex crystal structures

240th National Meeting (2010)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.