[gmx-users] how to calculate internal energy for energygroups?

[Mark Abraham]

Marc Baaden wrote:
> Hi Robert,
> 
> thanks for the hint.
> Just to make sure I get this right. So I'd use tpbconv with the subselection
> I'm interested in to make a new tpr file, and I also probably trjcat the
> trajectory that I want to analyze to write out the trajectory of the
> subselection. Then I do an mdrun -rerun with this tpr and xtc file and
> should be able to use g_energy to extract bond, angle and dihedral terms
> for my subselection.
> 
> Someone correct me if this is not the way to proceed. Thanks,

I'm sure mdrun will require a .tpr that has the same number of atoms as 
the trajectory you are analyzing. The way to get a subset of the bonded 
interactions is to only define in the .top file that set in which you 
are interested. You don't mess with the atom definitions however. For 
speed, turn off evaluation of all the non-bonded interactions somehow.

Mark