[gmx-users] Complete list of all interaction energies

[Mark Abraham]

jgoetze at uni-potsdam.de wrote:
> Hello everyone,
> 
> I would like to know if there is a tool for grabbing all individual 
> interactions of a system (analogous to the tinker tool "analyze"). I 
> skimmed through the available tools, however, I was unable to find one 
> which performs that function.
> 
> Technically, it should be enough to have the .gro ad the .top file as 
> input?

Well, not really, since even if it has the box and velocities, the .gro 
file is a reduced-precision snapshot of the state.

> The output should look somewhat like this:
> 
> Interaction type (bonds, angles, dihedrals, coulomb, vdW... whatever), 
> participating atoms (2, 3 or 4), potential energy contribution
> 
> Is there any way to obtain this information? Did I just miss the 
> appropriate gromacs tool?
> 
> The aim is to use that information for implementation of a QM/MM scheme.

You can use the energygrps run parameter to set up groups for which you 
would like the total intra- and inter-group nonbonded interactions. 
There is no similar feature for bonded interactions.

In GROMACS, the use of an MD "rerun" is the way to approach this 
problem. You take the same trajectory and apply a potentially different 
run input file to it. To isolate a subset of bonded interactions, you 
would need to create a modified version of the topology with other 
interactions zeroed out, construct your run input file suitably, and 
then do the rerun. Now only the interactions you preserved are still 
there. You won't get a bond-by-bond breakdown unless you script it by 
hand or hack the source code.

Mark