Thanks Fabricio, this is a useful reference, captures the points that we discussed.
All the best,
Andy
—
Dr. Andreas W. Goetz
Assistant Project Scientist
San Diego Supercomputer Center
Tel: +1-858-822-4771
Email: agoetz.sdsc.edu
Web: www.awgoetz.de
> On Aug 5, 2016, at 6:30 PM, Fabrício Bracht <fabracht1.gmail.com> wrote:
>
> That is a good point. Found this while searching on the subject. Maybe
> you've seen it already.
> http://pubs.acs.org/doi/abs/10.1021/acs.jctc.6b00269?journalCode=jctcce
>
> 2016-08-05 18:17 GMT-03:00 Dr. Andreas W. Goetz <agoetz.sdsc.edu>:
>
>> The solvent cap should be large enough to solvate the entire protein and
>> keep it solvated during the simulation, such that only water will be at the
>> vacuum boundary. The question is whether you prefer to simulate a protein
>> in a highly concentrated solution or in a tiny water droplet. I am not
>> generally advising against using PBC with a large qmcut value for NVT
>> simulations, which is common practice. But you should be aware that there
>> are discontinuities in the potential energy surface. The error due to the
>> use of approximate force fields and QM Hamiltonians is most likely larger.
>>
>> All the best,
>> Andy
>>
>> —
>> Dr. Andreas W. Goetz
>> Assistant Project Scientist
>> San Diego Supercomputer Center
>> Tel: +1-858-822-4771
>> Email: agoetz.sdsc.edu
>> Web: www.awgoetz.de
>>
>>> On Aug 5, 2016, at 11:06 AM, Fabrício Bracht <fabracht1.gmail.com>
>> wrote:
>>>
>>> I know. Sorry, I should have sent you the updated example.
>>> In the case of a solvent cap, since one cannot use implicit solvation
>> with
>>> QM/MMext simulations, does this mean that the rest of the protein will
>>> effectively be in a vacuum? And if so would it be wise to apply some
>>> restraining force on the atoms. There are some very mobile groups in my
>>> structure. Would this be better than using cutoff with a periodic
>>> simulation in NVT (since the cutoff in NVE leads to discotinuities)?
>>> Thank you
>>> Fabrício
>>>
>>> 2016-08-05 10:50 GMT-03:00 David A Case <david.case.rutgers.edu>:
>>>
>>>> On Thu, Aug 04, 2016, Fabrício Bracht wrote:
>>>>
>>>>> Thank you David. But, for the qm/mm md simulation, by specifying ntb=0
>>>>> would mean I would not be using periodic conditions, right?
>>>>
>>>> Correct: but in your example input, you did not specify ntb=0, and you
>>>> clearly
>>>> had a periodic simulation. (Defualt is ntb=1).
>>>>
>>>> ...dac
>>>>
>>>>
>>>> _______________________________________________
>>>> 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
>>
> _______________________________________________
> 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 Sat Aug 06 2016 - 13:30:02 PDT
