AMBER Archive (2006)

Subject: Re: AMBER: Fully QM periodic runs

From: Gustavo Seabra (gustavo.seabra_at_gmail.com)
Date: Wed Jun 14 2006 - 12:44:37 CDT


AFAIK, qmcut has nothing to do with the size of the quantum region. It
only determines the MM atoms that are being included in the QM
_interaction list_. From the manual:

" [...] for QM atoms any MM atom that is within qmcut of ANY QM atom
is included in the interaction list for all QM atoms. This means that
the value of qmcut essentially specifies a shell around the QM region
rather than a spherical shell around each individual QM atom".
(p. 147)

Gustavo.

On 6/14/06, Wei Zhang <zweig_at_scripps.edu> wrote:
> Hi Ross,
>
> I also have this question. There is a parameter "qmcut" in "qmmm"
> namelist, I tried to
> reduce its value before and I do get a smaller quantum region, Is it
> possible to get
> a full quantum molecule by simply set "qmcut" to zero?
>
> Sincerely,
>
> Wei
>
> Ross Walker wrote:
>
> >Hi Steven,
> >
> >Unfortunately in Amber 9 one cannot do a pure QM periodic system. You can
> >only do QM/MM with periodic boundaries and this has the restriction that the
> >QM region dimension + cutoff must not extend beyond the box boundaries. I.e.
> >it is designed for modelling say an active site in a protein that is then
> >solvated.
> >
> >The implementation is also designed for relatively small QM systems of the
> >order of 200 atoms max. You can go larger than this using divide and conquer
> >but even them the time per step is prohibatively large to do MD with. Thus
> >periodic boundaries for pure QM systems is not very useful on this scale
> >since if you were to have say just a box of 200 waters and treat this
> >periodically I think you would see artifacts due to the enforcing of
> >periodicity on such small systems.
> >
> >Such an option may be added to Amber 10 but there are no distinct plans at
> >present.
> >
> >All the best
> >Ross
> >
> >/\
> >\/
> >|\oss Walker
> >
> >| HPC Consultant and Staff Scientist |
> >| San Diego Supercomputer Center |
> >| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
> >| http://www.rosswalker.co.uk | PGP Key available on request |
> >
> >Note: Electronic Mail is not secure, has no guarantee of delivery, may not
> >be read every day, and should not be used for urgent or sensitive issues.
> >
> >
> >
> >>-----Original Message-----
> >>From: owner-amber_at_scripps.edu
> >>[mailto:owner-amber_at_scripps.edu] On Behalf Of Steven Winfield
> >>Sent: Wednesday, June 14, 2006 09:04
> >>To: amber_at_scripps.edu
> >>Subject: AMBER: Fully QM periodic runs
> >>
> >>Dear Amber users,
> >>
> >>I am trying to perform a QM/MM run on a box of water, treating
> >>everything quantum mechanically, but I'm running into some problems.
> >>If I use, for example, the TIP3PBOX unit as my input structure then
> >>sander fails because the size of the QM region is bigger than the
> >>periodic box. This is due to some atoms protruding out of the primary
> >>simulation cell. Even if I turn on iwrap=1 the problem is
> >>still present
> >>because the wrapping is done to the whole molecule only when
> >>the first
> >>atom of that molecule moves out of the primary cell. Also,
> >>introducing
> >>some vacuum around the structure still fails when molecules cross the
> >>new cell boundary.
> >>
> >>So my question is: Can a fully QM periodic run be done in the current
> >>AMBER, and if so how? (Wrapping atoms back into the primary cell
> >>individually rather than with the rest of the molecule for example?)
> >>
> >>Thanks,
> >>
> >>Steve Winfield.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu