Hi,
hm I'm a bit late on this, but in case you still want to know my $0.02:
> The V0 state includes the solute in water and the V1 state is only the
> water box. I defined the system like this:
>
>
> V0 => [anqWAT.top/rst] solvent "WAT" (common atoms) + solute "ANQ" (unique
> atoms)
>
> V1 => [anqBOX.top/rst] solvent "WAT" (common atoms) only
that seems good so far. The number and starting positions of the waters
are the same I assume.
> => V0.in
>
> ifsc=1,
> scmask=':ANQ',
> crgmask=':ANQ',
> noshakemask='',
> /
>
> => V1.in
> ifsc=1,
> scmask='',
> crgmask='',
> noshakemask='',
> /
>
>
> 1) Doing this I find that sander removes the charges from the solute from
> the very beginning (V0) and therefore the whole calculation is run as if
> the solute did not have any charges during all the lambda-values from 0.01
> to 0.99. Am I right? If yes,
You are right, you would not want to use crgmask on ANQ, unless you want
to calculate the free energy of removing the chargeless compound from
solvent.
> a) Should I just leave the "crgmask" in blank in V0 and V1 and sander will
> deal with the charge automatically?
yes, that is what you would do for a one-step solvation free energy
calculation.
> b) in the paper it is said that SHAKE algorithm is used to constraint
> hydrogens. Is SHAKE also used in the solute? I am asking this because I
> thought that shaking the part of the system that experiments the
> alchemical transformation was not a good idea since the partition function
> is modified (by constraining the stretchings), and therefore the free
> energy might be also affected for it.
well, you get the free energy for whatever potential energy function that
you actually use, in this case using SHAKE on solvent and solute. If you
believe that explicitly considering the hydrogen vibrations leads to a
very different solvation free energy, you might want to disable SHAKE for
the solute, but I doubt this makes a big difference for normal molecules
under typical conditions.
> 2) Are the other options in the input files OK for the calculation I am
> trying to run?
as far as I see yes. You seem to use custom values for scalpha and beta,
which may lead to convergence problems depending on what your solute looks
like (they may also work better than the defaults, this hasn't been
explored very much). The settings are probably fine, this is just
something to watch out for.
Thomas
Dr. Thomas Steinbrecher
formerly at the
BioMaps Institute
Rutgers University
610 Taylor Rd.
Piscataway, NJ 08854
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Apr 23 2013 - 05:00:03 PDT
