[gmx-users] Long range interaction energy of charged particles with PME - GMX 3.3.1

[David Mobley]

Ran,

> I previously asked how to calculate the long range interactions between
> charged particles simulated with PME:
> http://www.gromacs.org/pipermail/gmx-users/2004-October/012627.html
>
> I repeat the question here. If I want to calculate the LR Coulomb
> interaction energy between say a protein and a ligand the normal
> procedure would involve several mdrun -rerun runs: (1) with atoms
> charged (2) with the charges on the protein neutralised (3) with the
> charges on the ligand neutralised (4) with the charges on the solvent
> neutralised. When the protein is charged, there will be a net charge in
> the simulation shell in case (2). This net charged is corrected in PME
> as implemented in GMX. In GMX 3.2.1, I could not account for the
> correction, so the calculation wouldn't be exact (see the message
> above). What's the situation in GMX 3.3.1?

Systems with net charge are run with a uniform neutralizing background
charge sincce they need to be neutral for PME. I don't have any idea
how to properly account for this, although I would like to. This isn't
just an issue for calculating interaction energies between, say, two
molecules -- one also has to deal with similar issues if one wants to
compute, say, the free energy of inserting or removing an ion.

The problem, as I see it, is that I can't figure out any way to
calculate the contribution to the energy (or free energy) of turning
on the neutralizing background charge, which one would need to
subtract off to get this right.

I guess what makes this tricky is that you are really talking about
the same problem I would have if I want to get an ion hydration free
energy: You want to be able to evaluate the long range electrostatic
interactions in a system that has net charge (a protein+ligand, where
the protein+ligand isn't neutral; it's only neutral when you add the
ions), so you really need to be able to do this with the neutralizing
background charge (which will let you do the sum) and then somehow
subtract off the energy of turning on the neutralizing background
charge, which you can't compute directly since your sum won't converge
if you turn it off.

Maybe there is some way to get this out of the right combination of
mdrun -rerun options. Can you be a bit more explicit about how you are
thinking you would get this if you did (1) through (4)? And what do
you mean by saying that this is the "normal procedure"? (Do you have a
reference on it?)

Please continue the discussion. I'm happy to help in any way I can,
but I just need to understand what you're thinking a bit better, and
I'm also not sure anyone knows how to do this yet (at least, I'm not
aware of anything that explains this).

One reference that seems useful that I don't completely understand yet
is this one:

Kastenholz, M. & Hunenberger, P.H.
Computation of methodology-independent ionic solvation free energies from
molecular simulations: II. The hydration free energy of the sodium cation.
J. Chem. Phys., 124, 224501 (2006).

David Mobley
UCSF


>
> Thanks,
> Ran.
>
> --
> ------------------------------------------------------
> Ran Friedman
> http://bioinfo.tau.ac.il/~ran
> Laser Laboratory for Fast Reactions in Biology
> Department of Biochemistry
> Faculty of Life Sciences
> Tel-Aviv University
> Tel. +972-3-6409824
> Fax. +972-3-6409875
> ------------------------------------------------------
>
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>