CINF 75 |
| Virtual screening using 3D pharmacophores has evolved into an important and successful method for drug discovery over the last decades. We recently presented an efficient alignment method for super-positioning shared chemical features of pharmacophores and/or molecules in 3D space. Although efficient super-positioning techniques are of utmost importance to guarantee high throughput in virtual screening technologies, there is a need for automatically assessing the relevance and quality of a specific alignment for processing large data sets. Being aware of the problems of scoring functions in docking approaches the presented ranking approach has a different scope, since the position in 3D space is already defined by the single alignment solution coming from the alignment algorithm. The presented scoring function is therefore designed not to select poses of one single molecule, but to select those molecules, which better fit to a pharmacophore (or a shared feature pharmacophore hypothesis) compared to others. Geometric, steric and energetical contributions have been used for implementation and parametrization and applied to a diverse set of H1 antagonists. We used a pseudo-structure-based approach using a homology model and docked a data set of selected, active H receptor ligands using GOLD, and compared this to a ligand-based approach using multiple conformations generated by OMEGA within the LigandScout framework. |
|
General Papers
8:30 AM-11:50 AM, Thursday, August 23, 2007 BCEC -- 252 A, Oral
Sci-Mix
Division of Chemical Information |