[gmx-users] grompp warning

[Mark Abraham]

sarbani chattopadhyay wrote:
> Hi,
> I have a 13 residue peptide which I have kept in a box using the 
> following command
> 
> editconf -f conf.gro -bt cubic -o box.gro -d 0.75
> the flag "-d 0.75" is given as per the ' minimum image convention'.
> 
> When I solvate the box, 6772 water molecules get added.
> 
> With such large no. of water molecules i get the following warning while 
> using the "grompp"
> command.
> 
> T coupling group is less than 10% of the total system.

So you're being warned that you may be doing something unsound. You're 
actually probably OK - the warning is designed to alert newbies who are 
(e.g.) erroneously T-coupling their three counter-ions separately.

> I have coupled the protein and solvent to the heat bath separately, 
> using the terms
> "protein " and "non-protein".
> 
> Moreover with such large no. of water molecules , the no.of degrees of 
> freedom increases
> very much, and while doing Replica exchange md, I need to have more 
> no.of replicas to
> maintain my desired acceptance ratio.
> 
> How can I avoid this problem?

That's a cost of doing business. There are a variety of approximations 
to REMD that are designed to get around this issue, but none are 
implemented in GROMACS, to my knowledge.

Mark