[gmx-users] Re: Frozen atoms and energy group exclusion

[Janne Hirvi]

> Hello,
>      Thanks for the fast response.  I went through my dppc.itp file and
> chose
> atoms such that the group of frozen atoms has a net charge of 0 as does the
> rest of the molecule.  When running this through grompp, the error before
> about the atoms being in different charge groups is gone.  However, the
> warning (WARNING 1 [file "dppc.top", line 19]: Can not exclude the lattice
> Coulomb energy between energy groups) is still present.  What exactly does
> this warning mean?  Will this have a large affect on my simulation output?
> 
> Mike Tomasini

Lennard-Jones interactions and Coulombic short-range interactions (real space)
of groups defined in energygrp_excl are excluded but long-range interactions
(reciprocal space) can not be excluded in PME. I think this is not any problem
mathematically so that everything is calculated correctly but there might be
problem if you dont want these interactions to be present in your system.

Janne 

> On 5 Mar 2007, Mark Abraham wrote:
> > >      When I run this through grompp, I get the following error:
> > > WARNING 1 [file "dppc.top", line 19]:
> > >   Can not exclude the lattice Coulomb energy between energy groups
> > > Checking consistency between energy and charge groups...
> > > ------------------------------------------------------
> > > Program grompp, VERSION 3.3
> > > Source code file: grompp.c, line: 307
> > >
> > > Fatal error:
> > > atoms 3212 and 3213 in charge group 2179 are in different energy groups
> > > -------------------------------------------------------
> > >
> > > Am I doing this incorrectly?  Should I be excluding the interactions wi=
> > th
> > > the
> > > frozen groups in another way?
> > 
> > Have a look at the description of charge groups in 4.6.2 and 4.6.3 and th=
> > e
> > implementation of this in 5.5.1. grompp is evidently complaining that you=
> > r
> > charge group boundaries are not the same as the energy group boundaries,
> > and since non-bonded energies are calculated looping over charge groups,
> > this is a fine thing for it to be unhappy about! So have a look at the
> > charge groups in your DPPC molecule and see if you can come up with a
> > compromise that keeps charge group neutrality.
> > 
> > Mark