[gmx-users] Re:Re: how to calculate the non-bond interaction between solvent and solvent in pure solvent system (Justin A. Lemkul)

[Mark Abraham]

Jinyao Wang wrote:
>>> Hi Justin, Thank you for your reply. But I still don't know how to set the
>>> "energygrps" option in the rerun.mdp for calculating the interaction between
>>> solvent molecule and solvent molecule in pure solvent system. As you said
>>> that these terms should be obtained running the g_energy,could you give me
>>> some details about the "energygrps" option in rerun.mdp file or how to
>>> calculate the intermolecular interaction in pure solvent system. Thanks in
>>> advance. 
>>>
>> If you have a pure solvent system, there is no way to set separate energygrps, 
>> as you found out earlier.  There is also no need to re-run your simulation.  Use 
>> g_energy to analyze your original ener.edr file; you will see the LJ and 
>> Coulombic terms in the list of options.
>>
>> -Justin
> 
> Hi Justin,
>  My simulation system is pure benzaldehyde system. I set the option nrexcl = 3. 
> So I think the LJ and Coulombic terms include intermolecular interaction energy and intremolecular interaction eneryg (e.g. between atom and atom)

When you didn't state what your solvent was, we presumably all assumed
it was small enough that there were no intra-molecular interactions,
since that's normally the case with solvent molecules in biomolecular
MD. Please learn from this - we've all wasted a bunch of time and emails
because you didn't make a good enough description of the problem.

You're right that those total terms do include intra-molecular
contributions. To exclude them, do a rerun with a .tpr based on a .top
that uses a sufficiently high nrexcl, or manual [exclusions].

Energy group exclusions exclude sets of interactions between groups,
which would work here only if you defined a group for each molecule and
a pair between each molecule and itself (and that was legal... dunno
about that).

Mark