I know that multiple MD simulations are good for statistics and increases the reliability of the results.
I have a protein which is tetrameric in natural state. i would like to know whether I should run 50ns MD for complete tetramer or four 50ns simulation for one of the chain? I think when working with complete tetramer i might get more realistic simulation BUT will it be statistically significant like four 50ns simulation.
@jarmila thanks for the reply. What duration of simulation you suggest to perform ?
Running MD with tetramer in your case is meaningful. Dynamic nature of each monomer is dependent on presence of others. The conformational and subunit interactions can only be studied with tetramer as a unit.
For publication purposes, the standard benchmark is now 100ns and this is increasing by the day due to the available computer power.
How long a simulation needs to run in order to provide physically meaningful results depends on the observable that you want to measure. Different observables have very different relaxation times, and generally these times grow with the spatial extent of your observable. I can imagine observables for your protein for which 100ns would be a big overkill, and other observables for which it would be way too short. The idea that there exists a single number beyond which you are safe is therefore misleading.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics