[gmx-users] Bond constraints during the equilibration for membrane-protein system

[Justin A. Lemkul]

li lv wrote:
> Dear All
> 
>  
> 
> I am conducting a MD simulation for membrane protein with GMX 4.0.5. I 
> met lots of LINCS warnings (almost for angle change of N-H and C-H bond) 
> when doing equilibration for the membrane-protein system under NPT 
> condition. I had doing energy minimization (emtol = 500) before the 
> equilibration. To solve the problem, I was trying to turn off the bond 
> constraints, i.e. constraints = none, and restart 2-ns equilibration. It 
> seemed that everything is OK after the new equilibration (no bond was 
> broken and the protein backbone changed structurally slightly). Next I 
> continued 1-ns equilibration adding the bond constraints again and no 
> LINCS warning happened.
> 
>  

Bonds do not break or form during MD.  If you're seeing such visualization 
artifacts, it just means that your system was shearing apart and atomic 
positions were becoming unnaturally long or short.

Adding flexibility in the early stage of equilibration may have allowed some bad 
contacts to relax, but I cannot say if there is any negative impact from doing 
things this way.  If there were bad contacts to start, then the dynamics may be 
somewhat influenced by this fact.  Minimizing without constraints may help, as 
would be using several iterations of minimizations using different algorithms.

> 
> Before conducting my production MD with the files from the above 
> procedure, I am eager to clarify some confusion for my simulation, that is:
> 
> 1.       Whether my operations (especially turning off the bond 
> constraints during equilibration) are accepted theoretically for a good 
> MD simulation?
> 

If the constraints are failing, that does not mean you should disable them.  The 
problem is not the constraints themselves, just that it's the first algorithm to 
fail when your system becomes unstable.

http://www.gromacs.org/Documentation/Errors#LINCS.2fSETTLE.2fSHAKE_warnings

> 2.       If not, should I repeat energy minimization for the system? Be 
> honest, emtol = 500 is where the system can reach until now with my best 
> try. Maybe I have ignored other applicable EM or equilibration 
> operations besides the GMX manual?
> 

No one knows.  You haven't provided any relevant input information.  Generally, 
Fmax < 500 should be stable, but without being able to assess any of your other 
.mdp settings, we can't draw any conclusion.

> 3.       If yes, what should I do to guarantee the correct foundation of 
> the current membrane-protein structure, like checking the bond change, 
> or any other means?
> 

You haven't provided enough information, like what the system is, how you built 
it, what your input settings are, etc.

-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

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