Comparison of Pharmacophore Searching Methods for Potential Inhibitors of Tyrosine Kinases
Introduction
Tyrosine Kinases (TKs) are one of the most exploited groups of drug targets, whose known available inhibitors are mostly of the ATP-competitive type. Three pharmacophore searching methods were applied to finding potential ATP-competitive inhibitors of the Tyrosine Kinase family: i) manual pharmacophore modeling (MOE), ii) “fuzzy” pharmacophore models (SQUID), and iii) similarity searching based on correlation-vector representations of potential pharmacophore-points (CATS3D).
Performance comparison between these methods was based on retrospective virtual screening in a library containing ~5,000 drug-like molecules and different sets of known Tyrosine Kinase inhibitors as a reference. We determined the enrichment factors and the diversity in the retrieved molecular scaffolds (“chemotypes”) in the virtual hit lists.
Tyrosine Kinases (TKs) are one of the most exploited groups of drug targets, whose known available inhibitors are mostly of the ATP-competitive type. Three pharmacophore searching methods were applied to finding potential ATP-competitive inhibitors of the Tyrosine Kinase family: i) manual pharmacophore modeling (MOE), ii) “fuzzy” pharmacophore models (SQUID), and iii) similarity searching based on correlation-vector representations of potential pharmacophore-points (CATS3D).
Performance comparison between these methods was based on retrospective virtual screening in a library containing ~5,000 drug-like molecules and different sets of known Tyrosine Kinase inhibitors as a reference. We determined the enrichment factors and the diversity in the retrieved molecular scaffolds (“chemotypes”) in the virtual hit lists.
