COMP 45 |
| Utilizing knowledge from preexisting crystallographic structures is routinely done in a wide variety of docking protocols, including those implemented during pose generation as well as post-processing filters. Kinases provide a unique focal area for knowledge-based efforts due to the high conservation of interactions present in the hinge region of the ATP binding site as well as the multitude of data available, both of which are routinely exploited. Implementation of this knowledge in a general kinase protocol dissipates the bias that may arise when utilizing data from a congeneric series of ligands or within single kinases. As opposed to filtering poses using interactions that are common to a large number of co-crystal structures, which may narrow the search space, utilizing structural information from all kinase co-crystal structures provides the possibility to increase the range of poses generated. Core substructures common to kinase ligands have been defined and the placement of these substructures over all kinase co-crystal structures has been determined. Utilization of these core placements as starting poses in the docking protocol results in greater reproducibility of the crystallographic pose without limiting the possibility for less frequently observed poses. Additionally, coupling this docking protocol with a scoring function that is optimal for the description of kinase-ligand interactions improves the ability to retrieve well-docked poses. |
|
Drug Discovery
1:00 PM-5:05 PM, Sunday, August 19, 2007 BCEC -- 161, Oral
Division of Computers in Chemistry |