On Sat, Feb 13, 2010 at 10:28:38PM -0500, case wrote:
> On Sat, Feb 13, 2010, M. L. Dodson wrote:
> >
> > In the simulations
> > showing the H-dissoc, qmshake=0 AND qmcut > cut (in the &cntrl section,
> > where ntc=ntf=2, also).
>
> Thanks for the detailed debugging. I will note that when ntf=2, there is no
> bond potential. Hence, if shake gets confused, there is nothing to prevent
> the protons from flying off, as you observe.
>
> I'm not sure that there are places where one needs qmcut to be greater than
> cut, since the long-range qm-mm electrostatic interactions should be handled
> by PME. But I need to think about this, and we need to try to track down the
> real reason why the mm bonds in question are not be shaken.
>
> As a quick test, you might take a case that failed, and set ntf=1 to see if
> that avoids dissociation.
>
> ...thanks...dave
>
Thanks for looking into this. I'll set up a ntf=1 test Monday and
let you know the results.
I'm not at all worried about this, having qmcut != cut was a blunder
on my part caused by insufficient attention to the settings in the
copied input file. I had no rational basis for wanting qmcut > cut.
Just irritated with myself for losing quite a bit of simulation time.
The observation has something to do with spatial interaction of the
groups involved. None of the other arg residues in the protein did
this, only the one immediately adjacent to the qm region. There is an
arg on the other side of the active site, only slightly further way,
and it did not dissociate. Also, I did not see it for any other
residues than arg (although I did not hunt in great detail).
Maybe an additional sentence should be added to the qmshake
documentation to the effect that you should have a really good reason
to set qmcut > cut (qmcut != cut?). IIRC, there are mild warnings
like that in the ntf/ntc sections.
If someone needs test cases, I can furnish all the files, just let me
know. Or, if you want to just let it drop, I can understand: this is
a rather extreme corner case. You might even say it was hiding behind
the lamp in the corner. ;-)
Thanks again,
Bud
PS, Ross knows about this. I am investigating whether this is related
to SCF failures I observed in some of these same runs, and he has been
helping with that.
--
M. L. Dodson
Business Email: activesitedynamics-at-comcast-dot-net
Personal Email: mldodson-at-comcast-dot-net
Phone: eight_three_two-56_three-386_one
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Feb 14 2010 - 06:30:02 PST
