Solvation has been a long standing problem for structure-based drug design. We have implemented the generalized-Born (GB) model to DOCK. The GB model provides a good balance between computational speed and accuracy in solvationcalculations. We have derived a formula to calculate the binding energy. To improve the computational speed, the protein contribution to screening is pre-calculated and stored on a grid. The refinement of the ligand position is necessary to optimize the non-bonded interactions between ligand and receptor. Our implemented GB algorithm takes ~10 second per orientation on a Silicon Graphics Indigo2 workstation. The solvation scoring can be used as a post-DOCK database screening. Two systems have been tested: Dihydrofolate Reductase (4dfr), and trypsin. Our results show good improvement as compared with the results from current force field scoring. The Computer Graphics Laboratory facilities have been used for molecular graphic displays.
Showing the most recent 10 out of 508 publications