ChemAxon's Chemical Fingerprints-Based Clustering to Assess AurSCOPE Databases Chemical Diversity

presentation · 16 years ago
by Ismail Ijjaali (Aureus Pharma)

Among the knowledge databases developed at Aureus Pharma, AurSCOPE Kinase and AurSCOPE Ion Channels are focused on drugs described as kinase inhibitors or ion channel blockers, openers or activators. Both chemical and biological data are structured into a unique knowledge management system where chemical structures of ligands as well as precise description of their targets and biological activities have been collected from published articles and patents. The present work is aimed to assess the chemical and biological diversity of both databases. A specific query was constructed to retrieve most active non-peptide ligands. 2D molecular descriptors were computed and their distributions analyzed. Afterward, to encode the chemical space for the whole active dataset (9897 and 8371 molecules for ion channels and kinases, respectively), ChemAxon’s chemical hashed fingerprints were generated and optimized. These fingerprints were then employed to cluster these datasets using Jarp algorithm as implemented in JKlustor module. Clusters and singletons distributions at different similarity thresholds were analyzed and linked to corresponding biological activity. The findings show how highly structured databases combined with ChemAxon solutions can be an efficient tool to assist and inspire medicinal chemists, biologists, and computational chemists involved in a drug discovery project.