Dear Miro,
again, I am not the expert on the inner workings of the Amber MD engine,
but I believe that the barostat scales the coordinates per molecule.
I.e., all atoms of the same molecule get shifted by the same amount.
Therefore, the molecular geometry stays intact when scaling the
coordinates. However, if two restraint molecules touch, and their
reference coordinates are shifted by different amounts, this leads to
clashes between the two molecules.
Best regards,
Franz
On 11/22/22 15:09, Miroslav Suruzhon via AMBER wrote:
> Dear Franz,
>
> Thank you for the reply. So basically this means that in your case the reference restraint coordinates are not scaled accordingly with the barostat? Wouldn’t this result in potentially high energies and numerical instability in any case? Why is this affected by the number of restraints?
>
> Or is it the case that one of the restraints is scaled correctly but the other one isn’t?
>
> Many thanks.
>
> Kind regards,
> Miro
>
> Miroslav Suruzhon MSc PhD
> Molecular Dynamics Research Scientist
> +44 (0) 7719 025700
> Exscientia
>
> The Schrödinger Building, Heatley Road, Oxford Science Park, Oxford, OX4 4GE
>
>> On 17 Nov 2022, at 13:10, Franz Waibl via AMBER <amber.ambermd.org> wrote:
>>
>> Dear Miro,
>>
>>
>> I am one of the persons who previously asked about this in the mailing
>> list, since I did NpT simulations with restraints in previous Amber
>> versions - to prepare the system for the GIST method.
>>
>> In previous versions, the NpT simulations run fine if only one molecule
>> has positional restraints. If two molecules (e.g., two protein chains)
>> have restraints, the barostat tries to move them relative to each other.
>> Since this produces a high energy, the volume cannot equilibrate
>> properly. This can be avoided by modifying the topology file such that
>> all restraint molecules are treated as one "molecule". There don't need
>> to be any bonds between them, just modify the SOLVENT_POINTERS and
>> ATOMS_PER_MOLECULE.
>>
>> I am not completely up-to-date on the changes in Amber22, but I would
>> assume that the situation is still the same. If not, I hope that someone
>> can correct me.
>>
>>
>> This is my old question in the mailing list:
>>
>> http://archive.ambermd.org/201906/0295.html <http://archive.ambermd.org/201906/0295.html>
>>
>>
>> Best regards,
>>
>> Franz Waibl
>>
>>
>> Am 16.11.2022 um 13:43 schrieb Miroslav Suruzhon via AMBER:
>>> Dear AMBER users,
>>>
>>> I have seen some past questions/issues about positional restraints not being compatible with constant-pressure simulations because the restraint positions are not scaled by the barostat. Is this still the case in AMBER 2022?
>>>
>>> Many thanks.
>>>
>>> Miro
>>>
>>> Miroslav Suruzhon MSc PhD
>>> Molecular Dynamics Research Scientist
>>> +44 (0) 7719 025700
>>> Exscientia
>>>
>>> The Schrödinger Building, Heatley Road, Oxford Science Park, Oxford, OX4 4GE
>>>
>>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber <http://lists.ambermd.org/mailman/listinfo/amber>
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Nov 23 2022 - 03:30:03 PST
