Xin Chen, Laurent Alquier, and Charles H. Reynolds,
Computer Assisted Drug Discovery, R. W. Johnson Pharmaceutical
Research Institute, Raritan, NJ 08869, xchen3@prius.jnj.com
223rd ACS National Meeting
Orlando, FL
April 7 – 11, 2002
Abstract
An in-house similarity searching system, PRISM (PRI Similarity Miner), has
been developed for the general purpose of lead identification for drug discovery.
Its algorithm is based on the fusion of 2D linear fragment descriptors of
different sizes, which reflect different degrees of fuzziness. The details of this
algorithm will be explained. Some validation results, using the NCI AIDS and
MDDR data sets, will be presented. We will also discuss application of this
method to fast HTS follow-up.