So i have two structures, both ligand bound, and i need to predict the structure of the ligand free molecule. To my knowledge it is not possible but i am quite at the beginning of understanding options in MD and MM. Thanks for answers.
Danilo Boskovic
I think it depends on the differences between those conformations (ligand bound and ligand free). Will be hard to observe a large conformational change (if this is the case) in a "typical" MD simulation... Otherwise, if the differences are mainly in the disposition of side chains of the amino acids belonging to the binding site, it's perfectly possible to get that structure by using MD. I think, looking at other system similar to yours (for example a similar protein belonging to the same family) will give you an idea about the change in teh active site...
hope this helps!
The job can be done in a two-step way.
Firstly, run some conventional MD simulations to check whether conformational change happens within several hundred ns. If so, your goal is actually achieved. The rest of the work is just running longer to check whether this new conformation is stable.
If no conformational change is observed in several hundred ns simulations, use the last frame of conventional MD simulations and initiate replica-exchange MD simulations, which is more powerful to find new conformations. Do not use any reaction-coordinate based techniques such as umbrella sampling, as they have some biased on the possible conformational space to explore.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics