Docking flexible molecules: A case study of 3 proteins
- Sandia National Labs., Livermore, CA (United States)
- Sterling-Winthrop Inc., Collegeville, PA (United States)
One central problem in computational drug design is the prediction of binding geometries of candidate drugs in a protein binding site. This problem can be broken down into 3 well defined parts: (1) defining the molecule of interest; (2) modeling the protein and its interactions with solvent and the drug molecule; and (3) performing the conformational search to find low energy states of the system that (hopefully) correlate with the actual binding mode. This paper examines the conformational search issue, using a genetic algorithm (GA) for generating conformations. We present a case study analyzing the performance of our previously described method for docking flexible molecules into a protein binding site. We examine three proteins -- thermolysin, carboxypeptidase A, and dihydrofolate reductase and attempt to dock one or more ligands whose binding geometry is known crystallographically. These conformations give benchmarks against which we can measure rms fit vs molecular mechanics energy. Additionally, by minimizing the crystal inhibitor and watching its movement, we can get a measure of how unique the crystal ligand conformation is in our force field. For thermolysin, we used a single protein conformation, derived from the 5tmn crystal structure. From the Brookhaven Protein Data Bank and docked 8 ligands into it, each of which has a crystal structure bound with thermolysin. Cases whose binding energies span 6 orders of magnitude were examined.
- Research Organization:
- Sandia National Labs., Livermore, CA (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC04-94AL85000
- OSTI ID:
- 55293
- Report Number(s):
- SAND--95-8203; ON: DE95010693
- Country of Publication:
- United States
- Language:
- English
Similar Records
Comparative Assessment of Pose Prediction Accuracy in RNA–Ligand Docking