As Carlos mentioned, you should try the "idecomp" flag to get what you
want. The problem with decomposing into pairwise interactions within a
system is that no potential energy functions that Amber uses are truly
pairwise decomposable for the electrostatic contributions.
In implicit solvent, the dielectric boundary is defined by every atom in
the system, and since the solvation term depends on the definition of this
boundary (to distinguish between a 'buried' and 'exposed' residue),
interactions between 'pairs' of atoms or residues in this context will
include contributions from all other atoms in the system via the
calculation of the effective Born radii (GB) or the dielectric switching
function (PB). Furthermore, by using an implicit solvent model, you're
including the screening effects of the solvent molecules surrounding your
protein in your pairwise interaction.
This type of decomposition is available through MMPBSA.py on a single
structure via a 'stability' calculation. If you want to know how two
protein residues interact, this is the route I suggest. ante-MMPBSA.py
will set up stripped topology files for you to run MMPBSA.py with.
Explicit solvent (PME) carries with it a completely different set of
limitations. In Ewald sums, the electrostatic interaction is divided into
a short-range term and a long-range term (the first summed via a direct
sum-over-nearby-pairs and the latter summed in reciprocal space using a
FFT). The reciprocal space term in general is not pairwise decomposable
(I've been told you can play tricks, but it's not straightforward nor is it
implemented in Amber), and the direct-space sum is meaningless without it
(since it is artificially dampened with a neutralizing gaussian counterion
to reduce the required cutoff, and the effect of this artificial counterion
is reversed in the reciprocal space part). A quick glance at the code
suggests there is no decomposition accumulation in the reciprocal space
part, so I'm not even sure how you would interpret decomposed energies in
explicit solvent...
Perhaps someone can clear up that last point for me if there is a
meaningful interpretation. In any case, I assert (perhaps incorrectly)
that pairwise-decomposed energies can be qualitatively helpful, but should
be taken with a grain of salt.
HTH,
Jason
On Thu, Jul 19, 2012 at 6:21 PM, Irene Newhouse <einew.hotmail.com> wrote:
>
> I'm EXTREMELY interested in being able to do this myself! Someone please
> tell! Irene Newhouse
> > Date: Thu, 19 Jul 2012 18:47:25 -0300
> > From: bracht.iq.ufrj.br
> > To: amber.ambermd.org
> > Subject: [AMBER] Interaction energy between residue pairs
> >
> > Hello. If one had a molecular dynamics simulation of a solvated
> > protein (in this case, water is the solvent), and was interested in
> > retrieving VdW and Coulomb interaction energies between pairs of
> > residues (very specific pairs of residues. Just in case you are
> > wondering, there are no ligands, just protein residues) for this
> > particular simulation. What would be the best way to do this in Amber?
> > The idea here is see the effect that a mutated residue has in one(or
> > more) of the catalytic site residues.
> > Thank you
> > Fabrício Bracht
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jul 19 2012 - 19:00:03 PDT
