[gmx-developers] charge group size in grompp

[Berk Hess]

Hi,

I didn't think about broken charge groups when writing the check.
You have to do some effort to produce a configuration with broken charge 
groups.
But if grompp allows reading of such configurations, the checks should 
handle
this correctly. You can file an issue with low priority.

Berk

On 03/30/2011 03:47 AM, Jochen Hub wrote:
> Hi,
>
> when running grompp for an equilibrated lipid system in a skewed box 
> (for having a hexagonal prism), grompp seems to compute the size of of 
> the charge groups incorrectly, if the lipids are in a broken 
> representation:
>
> Largest charge group radii for Van der Waals: 9.377, 9.375 nm
> Largest charge group radii for Coulomb:       9.377, 9.375 nm
> WARNING 1 [file em.mdp]:
>   The sum of the two largest charge group radii (18.751873) is larger 
> than
>   rlist (1.000000)
>
> When I make the molecules whole by getting first a tpr and then using 
> trjconv -pbc mol:
>
> grompp -p topol_pop.top -c pop.pdb -f em.mdp -maxwarn 1
> trjconv -s topol.tpr -f pop.pdb -o popw.pdb -pbc mol
>
> grompp is doing ok.
> Largest charge group radii for Van der Waals: 0.246, 0.245 nm
> Largest charge group radii for Coulomb:       0.246, 0.245 nm
>
> The wrong charge group size is given in all possible until cell 
> representations (tric, rect, and compact). When starting
>
> Is that a known issue? I use version 4.5.4-dev-20110321-f8688a9.
>
> Should I submit a bug report?
>
> Thanks,
>
> Jochen
>