On Mon, Apr 4, 2011 at 7:35 AM, Christoph Malisi <
christoph.malisi.tuebingen.mpg.de> wrote:
> Hi Jason,
>
> thanks for your quick reply.
> >
> > You had me panicked here a little bit (I had thought that the values
> below
> > belonged to the total energy decomposition table).
> Sorry for the inconvenience!
>
An occasional sanity check is always good.
> The results you get
> > below are completely reasonable (and expected) for sidechain/backbone
> > pairwise energy decomposition. The reason is the way the energies are
> > decomposed in this case (sidechain/backbone). These potential terms are
> the
> > result of the sum of the interactions of the sidechain (or backbone)
> atoms
> > of Resid 1 with ALL of the atoms of Resid 2. Thus, the sidechain
> pairwise
> > interaction Residue A -> Residue B will be defined differently than the
> > pairwise interaction Residue B -> Residue A, so there's no reason to
> expect
> > them to be the same.
> >
> > This way of decomposing allows you to simply calculate the total as the
> sum
> > of the sidechain and backbone contributions (which is what is done in the
> > code). THESE values (the Total tables), should represent all-to-all in
> > terms of atomic pairwise potential terms between the two residues, so I
> > believe the values in those tables should match almost (if not) exactly.
> Do
> > they?
> >
> They do, here's the corresponding lines:
>
> Resid 1 | Resid 2 | Internal | van der Waals |
> Electrostatic | Polar Solvation | Non-Polar Solv. |
> TOTAL
>
> ASN 33 | ARG 41 | 0.000 +/- 0.000 | -1.200 +/- 0.317 | -4.847
> +/- 3.207 | 3.902 +/- 2.685 | -0.318 +/- 0.033 | -2.462 +/-
> 4.195
>
> ARG 41 | ASN 33 | 0.000 +/- 0.000 | -1.200 +/- 0.317 | -4.847
> +/- 3.207 | 3.902 +/- 2.685 | -0.316 +/- 0.033 | -2.460 +/-
> 4.195
>
> Seem to only have slight differences in the Non-Polar Solvation part. A
> quick look at some other residue pairs seems to confirm this.
> Thaks again!
>
Yea, the pairwise SA calculation is strange, and I don't understand how it's
decomposed. Furthermore, it's not strictly pairwise decomposable, so any
approach would have to be an approximation. In fact, even the GB energies
are not pairwise decomposable since they depend on the effective radii,
which are calculated via a volume integral over the entire system. The only
terms that *are* truly pairwise decomposable are the electrostatic and van
der waals interactions.
There are also known issues with SASA reproducibility using gbsa=2 (which is
what must be used for decomposition). See
http://archive.ambermd.org/201102/0227.html (and all messages in that
thread) and the continuation of that thread:
http://archive.ambermd.org/201102/0271.html.
All the best,
Jason
Best, Chris
>
> _______________________________________________
> 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 Mon Apr 04 2011 - 09:30:03 PDT
