On Wed, May 11, 2011 at 2:32 PM, Massimiliano Porrini <M.Porrini.ed.ac.uk>wrote:
> Dear Dan,
>
> Thanks a lot for your reply.
>
> But still I do not get the answer to my question and
> probably I should read your articles to get it (I must say I have already
> carefully read the Sugita and Okamoto work).
>
> I explain what I meant with an example.
>
> As you wrote we relax every replica within NPT ensemble (let's assume
> P = 1 atm).
> Let's consider the first two replica for simplicity at the beginning
> of the run, with their own
> temperature (T) and volume (V) values (obtained from the aforesaid
> relaxation):
>
> - Rep1: T1 and V1
> - Rep2: T2 and V2
>
> and let's assume T2 > T1, thus we have V2 > V1.
>
I would suggest *against* doing this. You should relax and heat up to only
the lowest temperature, followed by some NPT to relax the density. Do this
only for a single structure. From this heated structure, heat up a bunch of
replicas to your target temperature with NVT. This way, you ensure that all
volumes are exactly the same, and you get rid of this problem altogether.
Here's another good reason this is a better approach -- let's take the
(fairly common) case in which you have replicas at temperatures > 373 K. In
this case, if you run NPT to relax the density, you'll start seeing the
water boil! This is certainly not something that you want. Therefore, if
you follow the above recipe you will never have a case where the volumes of
different replicas are different, and the simple scaling should always work
to drop your system instantly into "equilibrium" with its new temperature.
If the temperatures are swapped between these two replicas and we are
> running
> in a NVT ensemble, the systems will end up having a new temperature
> "not compatible"
> with the actual volume, which should change accordingly.
> Specifically, after the exchange, V2 should contract whereas V1 should
> expand, but as we have
> frozen the initial volumes (by using NVT), these processes are not allowed.
>
> I hope I wrote more clearly what I meant and I apologise if the
> answers are in your papers.
>
Ultimately the only trajectory you're interested in is the one at "room"
temperature (perhaps a little above or a little below). Keep in mind that
the force fields were not parametrized for high temperatures, so any
"results" that you get will be meaningless, anyway. The low-temperature
replica, however, will have a valid ensemble, since the detailed balance
equations ensure a proper Markov chain of states.
HTH,
Jason
> All the best,
> Max
>
>
> 2011/5/11 Daniel Sindhikara <sindhikara.gmail.com>:
> > Since AMBER uses a velocity rescaling factor upon exchange, the replicas
> are
> > in equilibrium at the moment of exchange for NVT. To my knowledge, AMBER
> > does not do NPT exchanges. (Not sure if it is allowed or not). Most
> people
> > run relaxation simulations at NPT prior to REMD then freeze the volume
> for
> > NVT REMD simulation.
> >
> >
> >
> > For information about rescaling see the original paper by Okamoto:
> > * Y. Sugita, Y. Okamoto, Chem. Phys. Lett. 314 (1999) 141. *
> > For information regarding frequency of exchanges (within NVT) see:
> >
> > *D. Sindhikara, Y.L. Meng, A.E. Roitberg, J. Chem. Phys. 128 (2008)
> 24103.*
> > * *
> >
> > *D.J. Sindhikara, D.J. Emerson, A.E. Roitberg, *J. Chem. Theory Comput.*.
> 6
> > (2010) 2804.*
> >
> > Hope this helps!
> > --Dan
> >
> >
> > On Wed, May 11, 2011 at 8:19 PM, Massimiliano Porrini <
> M.Porrini.ed.ac.uk>wrote:
> >
> >> Dear Amber users and developers,
> >>
> >> I am going to do some REMD calculations on a small peptide in a
> >> water molecules box and a question came up to me.
> >>
> >> When the first (and then the following ones) attempt to swap
> temperatures
> >> between two adjacent replica is successful, does not it creat
> instability
> >> in the
> >> two replica?
> >>
> >> I mean each of the replica involved has a new temperature and therefore
> is
> >> no
> >> longer at the equilibrium, what about the value of density (and
> >> pressure) at that
> >> temperature?
> >>
> >> If one is using an NPT ensemble, the replica are likely to `relax' and
> >> reach the equilibrium again.
> >> However, even in this case, the temperature change might be too abrupt
> >> and to not give enough
> >> time to the replicas volume to contract or expand.
> >>
> >> Am I right or I am missing something here?
> >>
> >> I would really appreciate if someone can give clarification about this.
> >>
> >> All the best,
> >> MP
> >>
> >>
> >> --
> >> Dr Massimiliano Porrini
> >> P. E. Barran Research Group
> >> Institute for Condensed Matter and Complex Systems
> >> School of Physics & Astronomy
> >> The University of Edinburgh
> >> James Clerk Maxwell Building
> >> The King's Buildings
> >> Mayfield Road
> >> Edinburgh EH9 3JZ
> >>
> >> Tel +44-(0)131-650-5229
> >>
> >> E-mails : M.Porrini.ed.ac.uk
> >> mozz76.gmail.com
> >> maxp.iesl.forth.gr
> >>
> >> _______________________________________________
> >> AMBER mailing list
> >> AMBER.ambermd.org
> >> http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >
> >
> >
> > --
> > Dr. Daniel J. Sindhikara
> > Institute for Molecular Science
> > E-mail: sindhikara.gmail.com
> > Website: http://sites.google.com/site/dansindhikara/
> > --
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
> >
>
>
>
> --
> Dr Massimiliano Porrini
> P. E. Barran Research Group
> Institute for Condensed Matter and Complex Systems
> School of Physics & Astronomy
> The University of Edinburgh
> James Clerk Maxwell Building
> The King's Buildings
> Mayfield Road
> Edinburgh EH9 3JZ
>
> Tel +44-(0)131-650-5229
>
> E-mails : M.Porrini.ed.ac.uk
> mozz76.gmail.com
> maxp.iesl.forth.gr
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed May 11 2011 - 06:00:04 PDT
