Bit-string methods for selective compound acquisition

Selective compound acquisition programs need to ensure that the compounds t hat are chosen do not contain undesirable functionality. This is easy to ac hieve if a supplier is prepared to provide unambiguous structure representa tions for the compounds that they have available: this paper discusses sele ction techniques that can be used when a supplier is prepared to make avail able only fragment bit-string representations for the compounds in their ca talog. Experiments with three databases and three types of bit-string show that a simple k-nearest-neighbor searching method provides a surprisingly e ffective, although far from perfect, way of selecting compounds when only b it-string representations are available. A second approach, based on the us e of a fragment weighting scheme analogous to those used in substructural a nalysis studies, proved to be noticeably less effective in operation.