Dear Daniel, Dear David,
thanks for your replies. Yes, I see the same behavior with the Monte
Carlo barostat.
I can probably work around this by using no or very weak restraints
during the equilibration phase, as you suggested. If there are
unexpected problems with this, I'll write you again.
Thanks for your help!
Best regards,
Franz Waibl
On 6/24/19 1:31 PM, Daniel Roe wrote:
> Hi,
>
> Out of curiosity, does the same thing happen with the Monte Carlo
> barostat (which does not use virial)?
>
> Alternatively, you could just equilibrate your system in NPT ensemble
> without restraints, then do your dynamics in NVT ensemble with
> restraints.
>
> -Dan
>
> On Mon, Jun 24, 2019 at 7:07 AM Franz Waibl <Franz.Waibl.uibk.ac.at> wrote:
>> Dear Amber people,
>>
>>
>> I am working on a system containing of 2 chains in explicit TIP3P
>> solvent, and I want to restrain the solute for subsequent GIST
>> calculations. However, the density of the NpT equilibrated system is
>> significantly different depending on the restraint_wt setting.
>>
>> I attach a plot of the box volume vs. time with different restraint_wt
>> settings to illustrate the problem. I am aware that restraints of 300 or
>> 1000 are way too strong, but the effect is already quite pronounced with
>> 10 kcal/mol/A².
>>
>> It seems to me that this is the same problem that was described in
>> http://archive.ambermd.org/201103/0111.html a few years ago. In brief,
>> the problem seems to occur because the center of mass of the 2 molecules
>> is scaled with the box size during pressure equilibration. This leads to
>> clashes between them. Of course, those clashes are larger when stronger
>> restraints are used. It seems that the restraint forces are not included
>> in the virial, but the clashes are, so that the calculated pressure of
>> the system is too high, and the barostat tries to counteract this by
>> increasing the box size again.
>>
>> Is there any elegant way to avoid this problem? There seems to have been
>> no real solution in the original thread. A quick fix would be to do
>> multiple equilibration steps where each restart file is the reference
>> for the next step. But first of all I would still be worried about the
>> ensemble, and also there could be larger deviations from the starting
>> structure with each step. What is the recommended way of dealing with this?
>>
>>
>> Best regards,
>>
>> Franz Waibl
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Jun 24 2019 - 07:30:03 PDT
