| Abstract
| - This paper describes a novel clustering methodology for classifying over 700 conformations of a flexibleanalogue of GBR 12909, a dopamine reuptake inhibitor that has completed phase I clinical trials as a treatmentfor cocaine abuse. The major aspect of the clustering methodology includes an efficient data-conditioningscheme where a systematic feature extraction procedure based on the structural properties of the moleculewas used to reduce the associated feature space. This allowed region-specific clustering that focused onindividual pharmacophore elements of the molecule. For clustering of the reduced feature set, the fuzzyclustering partitional method was utilized. Due to the relational nature of the feature data, fuzzy relationalclustering was employed, and it successfully detected natural groups defined by rotational minima aroundN(sp3)−C(sp3), O(sp3)−C(sp3), and C(sp3)−C(sp2) bonds. The proposed clustering methodology alsoemployed several cluster validity measures, which corroborated the partitions produced by the clusteringtechnique and agreed with the results of hierarchical clustering using the XCluster program. Representativestructures which exhibited a reasonable spread of energies and showed good spatial coverage of theconformational space were identified for use as putative bioactive conformations in a future ComparativeMolecular Field Analysis of GBR 12909 analogues. The clustering methodology developed here is capableof handling other computational chemistry problems, and the feature extraction technique can be easilygeneralized to other molecules.
|
