NEW MOLECULAR SHAPE DESCRIPTORS - APPLICATION IN DATABASE SCREENING

Geometric descriptors are becoming popular tools for encoding molecula r shape, for use in database screening and clustering calculations. Th ey provide condensed representations of complex objects and, as a cons equence, can usually be compared quite rapidly. Here we present a numb er of new descriptors and methods for the quantification of molecular shape similarity. The techniques are tested using two different biolog ical systems, with particular emphasis on their potential utility as m ethods for prescreening shape-based database searches. Results are com pared with data sets produced using the DOCK program. We find that suc h similarity evaluations are useful for finding molecules with complem entary shape, and that they contain an enriched number of potential DO CK kits-when compared to the original databases. Significant limitatio ns in the utility of such DOCK prescreens are discussed, and potential solutions are considered.