These days simulations are performed for over 100ns easily. Will a 30ns data still be valid?
Hi Revathi
Total time of MD simulation is depend on what you want to see. If you want domain movement study then 100 ns MD simulation will be fine or more long MD simulation will required. If you want to see small confrontational changes then 30-40 ns is enough and moreover if RMSD plot of MD simulation is not fluctuating more then you can stop and analyze your data.
Revathi,
Its not a compulsion that you should go upto 100ns or more based on current scenario. The simulation time mainly depends upon your power to analyse the result which you want to show. With the trajectory of 30ns, if you are able to show the distinct conformational changes in your protein of interest, then you can start analysing with these results. If you think some more time scale should be extended for better understanding of results,then continue the simulation.
Well I suggest u can try till 50ns, which can be beneficial. Try to show your results distinctly what u want to prove. Good luck
Revathi,
I recommend you to continue the simulation process because it may generate further more varied conformation, Good luck
Hi Revathi
Total time of MD simulation is depend on what you want to see. If you want domain movement study then 100 ns MD simulation will be fine or more long MD simulation will required. If you want to see small confrontational changes then 30-40 ns is enough and moreover if RMSD plot of MD simulation is not fluctuating more then you can stop and analyze your data.
Hi
A long simulation is not sufficient. One should have to perform several independent simulations for better statistics. If you think that you have observed conformational changes, why not run another two/three 30 ns simulations to capture that similar conformational changes. If similar conformational changes will be observed, then statistics will improve, or, a different type of conformational fluctuations will be observed.
How is your system different from literature your are referring to? Smaller size? condense phase or vacuum. Which type of transition are you interested in? First order or second? There are many other things that can affect the transition time in MD simulation (Inter molecular long range interactions and even initial structure). If you want to make sure that you are not seeing an artifact I suggest trying different simulations with different initial structures.
Dear all,
Thank you so much fro your interest, I am really glad to see so much support in this forum.. I am intending to study the effect of a single disulphide bond in the stability of my protein. The change is visualised in RMSD within 30 ns from RMSD plot. I am not sure when I go for a publication if they will consider this short simulation or is it better to extend it further.
@Abolfazl Noorjahan : My system consists of 300000 atoms overall and it is much larger than systems in literature and fully hydrated system.
Thank you all once again.
Revathi
@Revathi
Have you checked the convergence in RMSD and energy during simulations? The system is large so try to determine the time in which these parameters are converged.
Dear Sir,
I extended the simulations for about 150 ns and now my RMSD graph looks like this: I am s=not sure how much forther should i keep extending and the refernce structure that I used to calcuate the structure is the structure that was obtained after solvating the vacuum minimised protein.
Kindly provide your valuable suggestions
The RMSD is not converged. what kind of RMSD are you calculating?
AS mentioned by other long simulation (100ns or more) is required for protein folding and conformation analysis while shorter one sat (30-50ns) is good for ligand interaction analysis. Moreover, multiple MD simulation increases the trust level in your analysis.
AS you are studying disulphide bond in the stability of my protein i would say go for at least 50ns simulation even if not more and try to do it 3 or more times. then you cam merge the trajectories and do the analysis. for each simulation try to gave different protein conformation so that you can cover maximum space by MD simulation
@Nitin: The simulation trajectory corresponds to 150ns. I am really sorry about the graph. This graph corresponds to a protein-ligand system that was run for 150 ns. Do you think I should extend it further? Kindly provide your valuable suggestions.
Thank you in advance.
the longer simulation is always better BUT your simulation is already long enough so in my opinion no need to extend. better repeat is once more and the merge the two simulations
@Revathi
It seems that RMSD is not converged. There are jumps in RMSD of ~1 A, I would not expect such type of jump (as I dont know time gap between frames). How you have calculated RMSD? What atom group you have used? Have you aligned structures before calculation? RMSD depends on the alignment (C-alpha, backbone and protein-H), choose carefully reference frame. Your interpretation about RMSD would be depend on this alignment and reference frame . If you have used both protein-ligand, then try to calculate for only protein. What of ligand RMSD with respect to the protein? Try to find, which part of protein's RMSD is not converging? Try to calculate RMSD of different regions separately.
Try to do 2-3 independent simulations, and then compare, which part of the protein is deviating, it is same region or something different.
- Badges
- Science topic