Rajarshi Guha wrote: > > On Jul 30, 2007, at 10:41 AM, Christoph Steinbeck wrote: > >> Peter, >> >> I don't think you are overlooking something. >> The code is very clean but the way the problem is implemented it >> catastrophic :-) >> (This is nothing against the author. At least, he did it!) >> >> The code performs a full substructure search for *every* possible path >> (length 2 to 6) instead of recording pathes while performing a >> breadth-first-search. > > The BFS approach would work OK for the chain type (which are contiguous > paths) Chi descriptors, but will not be applicable to the Path, > Path-Cluster types, where you need to consider rings and branches.
The branched substructures could also be detected via (modified) BFS starting from each atom. Admittedly, I don't know if there will be a performance increase. > I should also note that the latest version of Molconn-Z has simplified > the definition of Path and Path-Clusters, which might make detection of > these substructures easier - the current CDK implementation uses the > original definition of these descriptors as used in older versions of > Molconn-Z and ADAPT which are more complex substructures Is this documented somewhere? Cheers, Chris -- PD Dr. Christoph Steinbeck Lecturer in Chemoinformatics Univ. Tuebingen, WSI-RA, Sand 1, D-72076 Tuebingen, Germany Phone: (+49/0) 7071-29-78978 Fax: (+49/0) 7071-29-5091 What is man but that lofty spirit - that sense of enterprise. ... Kirk, "I, Mudd," stardate 4513.3.. ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Cdk-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/cdk-user
