[gmx-users] equilbium convergence and RMSD structure

[Justin A. Lemkul]

larif sofiene wrote:
> Greeting 
> I wonder if my MD simulation of a monomeric enzyme in WT and 3 mutated 
> forms has converged to equilibrium and simulation is suitable for analysis.
> The 4 MD simulations are done in cubic water volume, 500 ps of steepest 
> descent minimization , with 100 ps of NVT ensemble and 100 ps of NPT 
> ensemble , reaching for all of them their desirable value and being 
> stable over time
> When doing an MD production for 10 ns with those parameter : 
> * constraint_algorithm = lincs
> * constraints    = all-bonds
> * coulombtype    = PME
> * tcoupl        = V-rescale
> * pcoupl        = Parrinello-Rahman
> * pcoupltype    = isotropic
> 
> leading to those RMSD
> http://imageshack.us/photo/my-images/840/rmsd.png/
> the first one on the upper left is wild type other are mutants.
> my question is are my system in equilibrium despite those 2 spikes on 7 
> and 9 ns in WT RMSD , knowing that all the 4 MD had stable radius of 
> gyration , stable potential energy , stable H bond forming and 
> dissociation over time and reasonable H bonds under structural analysis, 
> and if so, does those spikes mean structural changes in flexible regions 
> that i must check for with RMSF ?
> Thanks in advance for responding :)
> 

Several general points:

1. I doubt that's the full .mdp file; in fact, it can't be.  Please do not post 
snippets of such information, post the full thing.  Details that may not strike 
you as significant can sometimes be.

2. 10 ns of time is relatively insignificant to study almost anything related to 
proteins.  You haven't stated your goals for these simulations, but I'd argue 
that you've got no more than 5 ns of usable data in some of the simulations, 
often less.

3. You cannot justify convergence of any phenomenon using a single simulation, 
especially of such short length.

4. It appears from the figure in the bottom left that the RMSD is still 
systematically trending upward, indicating that this simulation is not nearly 
done.  Block averaging will tell for sure whether or not I've passed this 
eyeball test.

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================