Hi,
The error you show is most probably related to the diffusion problem: the QM
zone is becoming too big. This diffusion is actually natural. The only way
to avoid it, for now, is to impose restraints to the QM water molecules, but
I have no idea what effect this will have in the end.
About the DFTB convergence, well... This is not an "Amber" problem:
convergence difficulties are well known in the DFTB users community. In
Amber 10, there is a way to improve DFTB convergence, but is not guaranteed
to work: Amber 10 has a new namelist keyword, 'dftb_telec', which sets an
'electronic temperature', which improves convergence. Notice that you do
*not* want to give it a very high number. Usually, 'dftb_telec=100' does
solve the problem, and higher numbers don't seem to make much difference.
You definitely don't want it to be higher than your system temperature.
HTH,
Gustavo.
> -----Original Message-----
> From: owner-amber.scripps.edu
> [mailto:owner-amber.scripps.edu] On Behalf Of M. L. Dodson
> Sent: Monday, July 30, 2007 2:23 PM
> To: amber.scripps.edu
> Subject: AMBER: QMMM to follow reaction involving water
>
> Hello Ambers,
>
> I have been doing QMMM (using a development version of
> Amber10 sander furnished by Ross Walker). The basic design
> is steered QMMM with the steered coordinate along a proposed
> scissile C-N bond. I have been having a lot of trouble with
> this calculation when I incorporate close waters into the QM
> region. These problems seems to be of two kinds: (1) the QM
> water(s) diffuse away from the enzyme active site, (2) the QM
> calculation (SCC-DFTB w/o dispersion corrections) has great
> difficulty achieving SCF.
>
> I can understand (I think) the reasons for both of these problems.
> The diffusion problem is obvious, the SCF problem is probably
> secondary to the diffusion, the waters are too far away from
> the remainder of the QM region.
>
> My question is whether anyone has suggestions to bypass these
> issues? (I already know I can restrain the waters, but I
> have some questions in my mind about the shape of the
> restraint. These question involve the restraint distance and
> the "steepness" of the parabolic restraint component. I
> don't want to bias the calculation whose whole purpose is to
> reveal possible reaction
> pathways.)
>
> I would appreciate any comments people might have. I have
> included the "in" file and the dist.RST file for my last run.
> It bombed with:
>
> ****************************************************
> ERROR: QM region + cutoff larger than box dimension:
> QM-MM Cutoff = 9.0000
> Coord Lower Upper Size Box(a,b,c)
> X -27.285 46.641 73.926 73.802
> Y 21.004 67.921 46.917 78.348
> Z 9.986 54.744 44.758 70.532
> ****************************************************
> SANDER BOMB in subroutine QM_CHECK_PERIODIC<qm_mm.f>
> QM region + cutoff larger than box
> cannot continue, need larger box.
>
> I think this is related to the diffusion problem.
>
> Bud Dodson
> --
> M. L. Dodson
> Email: mldodson-at-houston-dot-rr-dot-com
> Phone: eight_three_two-56_three-386_one
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Aug 01 2007 - 06:07:27 PDT
