[gmx-users] distance constraints

[David Mobley]

All,

Since we're on this subject, I would just like to gripe about this
requirement that, in order to apply restraints (distance, angle,
dihedral) between atoms, the atoms must be in the same 'molecule'
definition in a GROMACS topology file. It's rather inconvenient. For
example, I ran some calculations where I found some ordered waters got
stuck in a protein during MD. I wanted to compute the free energy of
removing those ordered waters from within the protein, in order to
figure out whether it *really* was favorable for them to be in the
protein, or whether they were just "stuck" because I hadn't run long
enough. Currently, there is no way to do this in GROMACS without (a)
renaming the waters so they're no longer recognized as solvent, (b)
listing the waters as part of the protein 'molecule' (i.e. at the end
each of the sections in the molecule definition for the protein,
adding the relevant water parameters), and (c) editing the coordinate
file (.gro) to  place the waters in question consecutively after the
protein. I now have scripts that do this, of course, but it's a big
pain. And I think it probably wouldn't be too difficult to cook up
cases where you want to apply restraints between things that you
simply CAN'T DO with the current code.

It seems to me that the source of the problem is that topology files
don't necessarily have unique atom numbering; rather, the atom
numbering only runs within the same molecule definition. Hence if I
specify restraints between atoms 11 and 73, GROMACS requires that they
 be within the same molecule definition, otherwise it wouldn't be able
to sort out which atom corresponds to atom 11.

Probably there is a straightforward work around. I can think of two
possibilities off the top of my head:
(1) Add an additional restraint option which allows use of absolute
atom numbering (i.e., the numbering that shows up in the .gro file).
There's no ambiguity here, since in the .gro file, all of the atoms
have unique numbers.
(2) Add an option to specify the molecule(s) in which the atoms to be
restrained are. Again, numbering from the .gro file would probably
have to be used.

Anyway, just a suggestion.

Thanks,
David



On 2/21/06, Mark Abraham <Mark.Abraham at anu.edu.au> wrote:
> Uwe Richter wrote:
> > Hi Mark,
> >
> > I assume Michal knows how to define distance restraints. That's described
> > in the manual.
>
> It would be good to be able to assume that :-) Unfortunately my
> impression is that many people who post haven't done enough work
> themselves on the problem... to the point that they haven't read the
> manual (well enough). One could certainly interpret Michal's question in
> the way you did, though!
>
>  > The problem is you can't do this between different
> > 'molecules',
> > as you know. I think 'molecule' might not be the best expression in this
> > context.
>
> Indeed, I carefully used the word 'topologies' - probably not perfect
> given the specific GROMACS usage, but something like 'assemblies' or
> 'systems' aren't quite right either... 'structures'?
>
> > You can very well have many molecules in one topology section, which has
> > the
> > advantage that it is possible to define distance restraints between
> > them. Of
> > course this setup also has its disadvantages. It really depends on the
> > project.
>
> That you have to merge the topologies to be able to have the option of
> restraints is a good point - one I hadn't encountered before. Thanks.
>
> Mark
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>