Hello,
My comments are below:
On Fri, Dec 17, 2010 at 3:01 PM, Andreas Goetz <agoetz.sdsc.edu> wrote:
> Dear Dmitry,
>
> while it is true that total energies are irrelevant and only relative
> energies have significance, you should always use the total energy from the
> Amber output (Etot) to obtain relative energies, otherwise you would be
> using an energy expression that is not clearly defined.
>
> Now, if I understand correctly, the question you are posing is geared
> towards a different problem, which is how to deal with the noise on reaction
> path energies that is due to conformational changes in the environment. This
> is not directly related to QM/MM but also holds for pure QM or pure MM
> calculation. You probably have to do something like umbrella sampling along
> the reaction path. Maybe someone with more experience can comment on this.
>
> I hope this helps.
>
> All the best,
> Andy
>
> On Dec 17, 2010, at 4:31 AM, Dmitry Nilov wrote:
>
> > Dear Ross, dear Andreas! Could you be so kind to answer my question
> > associated with the topic!
> > As I suppose, it is not relevant to use TOTAL potential energy of
> > complex protein-solvent QM/MM system (where QM is the active site) to
> > compare two different energy states of the active site, because
> > motions in regions distant to the active site may significantly affect
> > the energy profile.
>
You are right in that distant motions may have a large impact on your
potential energy. I would disagree that you should only consider the SCF
energy, though, since as Andreas pointed out, any 'partial energy' term is
not well-defined. This is because the potential terms are non-orthogonal
(they all depend on each other). In fact, the ESCF depends on the distant
conformations as well inasmuch as they change the electric field felt at the
various positions in the quantum region.
For this reason, you can only really consider the total potential energy
when comparing two states. However, for flexible molecules, potential
energy is strongly conformation dependent and this "noise" is a fact of
life. What you have to do is average over these conformations in a
statistical mechanically correct way, which is to say you have to sample
from the appropriate boltzmann distribution. Thus, if you want to find the
free energy along a reaction pathway, you will average over all other
degrees of freedom by building a proper ensemble at each point along the
path (you can use either monte carlo or molecular dynamics for this). This
is called a potential of mean force (PMF), and can be obtained, as Andreas
suggested, with umbrella sampling, though other methods can be used as well.
Hope this helps,
Jason
> So, is it reasonable in this case to use just DFTBESCF energies to
> > reconstruct the energy (enthalpy) profile of a reaction while QM-MM
> > VDWAALS energies are not included in DFTBESCF?
> >
> > On Fri, Dec 17, 2010 at 1:18 AM, Andreas Goetz <agoetz.sdsc.edu> wrote:
> >> Dear Markus,
> >>
> >> The answer to all your questions below is 'yes'. You got everything
> right.
> >> Ross will send a more detailed email for anybody that didn't understand
> it so far.
> >>
> >> All the best,
> >> Andy
> >>
> >> For reference the links to earlier posts:
> >> http://archive.ambermd.org/201010/0503.html
> >> http://archive.ambermd.org/201011/0098.html
> >>
> >> On Dec 16, 2010, at 7:49 AM, Markus Kaukonen wrote:
> >>
> >>> Dear Amber,
> >>>
> >>> This issue has already been dealt with on this list, but I did not
> >>> quite get it, so here we go again:
> >>>
> >>> In QM/MM calculation with DFTB as QM engine one gets output like (with
> Amber10)
> >>>
> >>> NSTEP ENERGY RMS GMAX NAME NUMBER
> >>> 1 -2.1816E+04 1.7957E+01 1.4342E+02 C36 12502
> >>>
> >>> BOND = 2156.1278 ANGLE = 6029.2390 DIHED =
> 9228.7109
> >>> VDWAALS = -7560.4126 EEL = -52399.6187 EGB =
> -5475.9033
> >>> 1-4 VDW = 2123.9140 1-4 EEL = 37427.0797 RESTRAINT =
> 0.0000
> >>> DFTBESCF= -13345.6223
> >>>
> >>> Question: What is included in the term 'DFTBESCF'
> >>> Is it only the QM energy of the QM region (terminated with H atoms)
> >>> and electrostatic interaction of QM nuclei and electron cloud with
> >>> unscaled point charges of the MM region?
> >>>
> >>> Vdw interaction between QM and MM region is in included into 'VDWAALS'
> >>> and '1-4 VDW' ?
> >>> Possible bond, angle and dihedral terms between QM and MM regions are
> >>> included into the classical terms (BOND, ANGLE, DIHED)?
> >>> Solvation energy of QM region is in included into EGB ?
> >>>
> >>> Did I get it right?
> >>>
> >>> terveisin, Markus
> >>
> >> --
> >> Dr. Andreas W. Goetz
> >> San Diego Supercomputer Center
> >> Tel : +1-858-822-4771
> >> Email: agoetz.sdsc.edu
> >> Web : www.awgoetz.de
> >>
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> AMBER mailing list
> >> AMBER.ambermd.org
> >> http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >
> >
> >
> > --
> > Dmitry Nilov,
> > Lomonosov Moscow State University
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
>
> --
> Dr. Andreas W. Goetz
> San Diego Supercomputer Center
> Tel : +1-858-822-4771
> Email: agoetz.sdsc.edu
> Web : www.awgoetz.de
>
>
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Graduate Student
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Dec 17 2010 - 18:30:05 PST
