[gmx-users] Energy calculation problem with molecule leaving the box

[Mark Abraham]

On 4/02/2011 1:05 AM, Gordan Horvat wrote:
> I'm doing molecule dynamics of a calixarene in a acetonitrile box with 
> pbc, energy groups defined and npT constant. When I extract 
> interaction energies with g_energy from edr file I expect them to be 
> pretty much constant because I see no significant conformational 
> changes or changes in distance between energy groups and that is true 
> for one part of simulation when the molecule is completely in the box. 
> But when a molecule partially leaves the box energies of interaction 
> for that part which is out of the is box is not calculated (bonded or  
> with other energy groups) and interaction energy sometimes drops to 
> 0.000 for a period of time even though the distance between groups 
> doesn't change.
> I tried extracting energies through rerunning simulation on 
> trajectories which were converted with -pbc nojump or -pbc mol -center 
> but that gave zero energies all the time.
>
> How can I fix this energy calculation problem?

So, having run your original simulation with pre-defined energy groups, 
starting from the mdrun output files, what commands reproduce your 
observations?

Mark