Please login first
Finding Biological Active Compounds in Large Databases;
* , * , *
1  Novartis AG, Postfach, CH-4002 Basel, Switzerland

Abstract: A method is proposed, and shown to be able to find active compounds in a Large Database by iterative selection/screening cycles. The method is derived from genetic algorithms. Starting from an initial small parent population the next compounds to screen are selected by similarity search based on the structural features of two high rating parents. The performance of the method is demonstrated using 20000 compounds with biological data from the NCI1 database.
Keywords: n/a