[gmx-users] Re: Targeted MD

[Mark Abraham]

Stanislav Bobritsky wrote:

>      > Well, but (as I know from Manual, Ch. 6, p. 120) there is
>     "approach to do targeted MD" in GMX.
> 
>     Did you add position restraint to your protein
>     topology?
> 
> 
> No, I did not.
> What group must be restrained therefore? Whole protein or whole protein 
> except loop?

 From "man grompp",

> When using position restraints a file with restraint coordinates can be
> supplied with -r, otherwise restraining will be done with respect to the
> conformation from the -c option. For free energy calculation the the
> coordinates for the B topology can be supplied with -rb, otherwise they will
> be equal to those of the A topology.

Thus I expect your earlier grompp command line didn't do the job you 
thought it did.

You could restrain just the loop if you knew the protein was stable, but 
easiest is to present conformation A as -c and B as -r. pdb2gmx will 
have supplied you with a posre.itp file, and you need -DPOSRES in the 
include .mdp directive so that the former works.

Your .mdp file as supplied earlier should not T-couple only your solute, 
and neither should it couple your counterions separately from the 
solvent. See http://wiki.gromacs.org/index.php/Thermostats

Based on the foregoing, I'd guess that you're fairly new to MD and/or 
GROMACS. I'd strongly recommend doing some (more) tutorials and 
background reading so that you have more idea what "normal" MD is, and 
the consequences of changes you might make. Otherwise the odds are 
excellent you're just using an complicated quasi-random number generator.

Mark