Hi,
In addition to Dave's reply, the 'OPT' keyword that you include in your
template is overridden by the 'Force' keyword that is added by sander.
This keyword asks Gaussian to compute only the energy and the forces
(=minus the gradient). That's all. That's why you see a huge difference
in timings and outputs between keeping or manually removing the 'Force'
keyword.
Don't forget that, you are asking a QM/MM calculation, that means you
need to compute the energy of the MM part (done by sander), the QM part
(done by Gaussian), and the interaction between the QM and the MM part
(done partially by sander and partially by Gaussian). You cannot have
the two programs to change the system coordinates independently!
Therefore, only one program should handle the optimization (and,
therefore, take care of modifying the coordinates of every atoms).
In other (now politically incorrect?) words, let's say that in Amber,
sander is the "master" program and Gaussian, and by extension all EXTERN
programs, are "slave" programs. Only the master program has the right to
modify the coordinates, here either for minimization or molecular dynamics.
Gerald.
On 06/13/2014 04:58 PM, Kshatresh Dutta Dubey wrote:
> Dear Prof Monard,
> Thanks for your detailed explanation, it is really helpful. However, I
> have still some queries. QM/MM run using Gaussian is producing some file,
> like old.gau_job.log, old.gau_job.inp and others. When I am checking
> *old.gau_job.log
> *file there is no proper convergence informations as usually we find in any
> Gaussian optimized output file. Indeed it successfully finishes proper
> termination in less than 1 hour while number of atoms are more than 100.
> When I ran same old.gau_job.inp file by remove 'Force' keyword appended by
> sander, it runs as usual QM optimizations and prints all convergence
> informations and takes relatively very long time in proper convergence (
> same input without 'Force' keyword on same machine). Therefore, I guess
> that even assigning min=1 keyword, Gaussian is not optimizing the system
> and only computing its energy (as single point calculations).
> I am attaching old.gau_job.inp as well as QM-MM.in. I request you to
> tell me where am I doing a mistake. I will be thankful for your help.
>
> With Regards
> Kshatresh
>
>
>
> On Fri, Jun 13, 2014 at 4:48 PM, Gerald Monard <
> Gerald.Monard.univ-lorraine.fr> wrote:
>
>> Hi,
>>
>> On 06/13/2014 01:26 PM, Kshatresh Dutta Dubey wrote:
>>> Dear Jason,
>>>
>>> I am looking for QM optimization, I know that using imin=1 will
>> give me
>>> minimized structure but I think that it will molecular mechanically
>>> minimize the system, but what for QM optimizations?
>>
>> No. imin=1 will minimize the _whole_ system: QM _+_ MM, except if you
>> use some kind of restraints (ibelly, which is not recommended for
>> minimization).
>> Minimization means "energy minimization": you compute the energy of the
>> _whole_ system (QM + MM and their interactions), the gradient, and from
>> the latter, the minimization algorithm decide how to modify the
>> coordinates of the whole system to obtain a new set of coordinates that
>> should lead to a lower energy.
>> That is to say: to perform minimization, the algorithm must have
>> information on the whole system, not part of it.
>>
>> If you refer to some kind of minimization algorithm as develop in the
>> group of Weitao Yang some years ago or in the group of Keiji Morokuma
>> (iterative minimization, J. Chem. Phys. 112, 3483--3492 (2000) or J.
>> Comp. Chem., 24, 760-769 (2003)), this is not implemented in Amber.
>>
>>
>>> If I have to use
>>> sander for minimization then what is the use of implementation of
>> Gaussian
>>> interface? I am curious to know that - is it possible to perform QM
>>> optimization of QM region using external QM program and sander for MM
>>> region. If yes then what should I do? Using following input (complete
>>> input, I used) does not optimize the QM region, it does single point
>> energy
>>> calculations only, while sander does optimize MM region.
>>
>> In the input below, both regions are optimized at the same time (as
>> there is only 1 energy computed: a QM/MM energy).
>>
>> Gerald.
>>
>>
>>>
>>> *QM-MM calculations for FMN domain *
>>> *&cntrl*
>>> * imin=1, maxcyc=10,*
>>> * ntb=0,*
>>> * cut=20.,*
>>> * ifqnt=1*
>>> * /*
>>> * &qmmm*
>>> * qmmask ='.819-830,2244-2295',*
>>> * qmcharge=0,*
>>> * spin =1,*
>>> * qm_theory='EXTERN',*
>>> * qmcut=20.0*
>>> * /*
>>> * &gau*
>>> * mem = '2GB'*
>>> *! method = 'B3LYP',*
>>> *! basis = '6-31G',*
>>> * use_template =1,*
>>> * num_threads = 2*
>>> * /*
>>>
>>> Thanks and regards
>>> Kshatresh
>>>
>>>
>>>
>>> On Fri, Jun 13, 2014 at 2:01 PM, Soumendranath Bhakat <
>>> bhakatsoumendranath.gmail.com> wrote:
>>>
>>>> Use these input for minimization if you are using QM/MM. Please use
>> sander
>>>> and sander.MPI for running
>>>>
>>>> Full_Mini.in
>>>>
>>>> Initial minimization of QMMM: solvent molecules and added ions
>>>> &cntrl
>>>> imin = 1,
>>>> maxcyc = 2500,
>>>> ncyc = 750,
>>>> ntb = 1,
>>>> ntr = 1,
>>>> cut = 12.0,
>>>> ifqnt = 1,
>>>> /
>>>> &qmmm
>>>> qmmask= '557,250,238.75,86,89,75',
>>>> qmcharge=0,
>>>> qmtheory=7, *(please suggest check once)*
>>>> qmshake=1,
>>>> qm_ewald=1,
>>>> qm_pme=1,
>>>> /
>>>>
>>>> Hold the Protein fixed
>>>> 10.0
>>>> RES 1 557
>>>>
>>>> All the best.
>>>>
>>>>
>>>> On Fri, Jun 13, 2014 at 12:54 PM, Jason Swails <jason.swails.gmail.com>
>>>> wrote:
>>>>
>>>>>
>>>>>
>>>>>> On Jun 13, 2014, at 5:50 AM, Kshatresh Dutta Dubey <
>>>> kshatresh.gmail.com>
>>>>> wrote:
>>>>>>
>>>>>> Dear Prof Monard,
>>>>>>
>>>>>> Many thanks for your quick response, still I am unable to find any
>>>>>> keyword that handles the optimizing process in sander. I went through
>>>> the
>>>>>> manual bt I was unable to find such keyword. I will deeply appreciate
>>>> you
>>>>>> if you help me to handle this process so that I can get proper QM
>>>>>> optimization.
>>>>>> Thanking you again.
>>>>>
>>>>> What you're looking for is a basic minimization. Look at the imin and
>>>>> ntmin flags.
>>>>>
>>>>> HTH,
>>>>> Jason
>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Fri, Jun 13, 2014 at 12:31 PM, Gerald Monard <
>>>>>> Gerald.Monard.univ-lorraine.fr> wrote:
>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> You cannot use the Gaussian optimizer since it is sander that masters
>>>>>>> what happens to the coordinates (from a coordinate set, sander asks
>>>>>>> gaussian to compute an energy, it does not let gaussian modify them,
>>>>>>> hence the "force" keyword).
>>>>>>> You'd better use one of the optimizers available in sander (&cntrl
>>>>>>> namelist) instead.
>>>>>>>
>>>>>>> Gerald.
>>>>>>>
>>>>>>>
>>>>>>>> On 06/13/2014 11:16 AM, Kshatresh Dutta Dubey wrote:
>>>>>>>> Dear Users
>>>>>>>>
>>>>>>>> I am using Amber 14 for QM/MM calculation with Gaussian as
>>>> external
>>>>>>>> interface. I have to optimize a system containing some amino acids
>>>> and
>>>>> a
>>>>>>>> ligand in QM regions. I found that Amber automatically adds some
>>>>> keywords
>>>>>>>> in route section of Gaussian input that force Gaussian to do single
>>>>> point
>>>>>>>> calculations. For example, I used :
>>>>>>>> *&gau*
>>>>>>>> * mem = '2GB'*
>>>>>>>> *! method = 'B3LYP',*
>>>>>>>> *! basis = '6-31G',*
>>>>>>>> * use_template =1,*
>>>>>>>> * num_threads = 2*
>>>>>>>> * /*
>>>>>>>> for &gau namelist and :
>>>>>>>>
>>>>>>>> *#P OPT B3LYP/6-31G SCF=(Conver=8)*
>>>>>>>>
>>>>>>>> for template file namelist. The calculations runs fine and it shows
>>>>>>> normal
>>>>>>>> termination, but I found that it does only single point calculation
>>>> not
>>>>>>>> optimizing the structure while I have mentioned optimization in
>> route
>>>>>>>> sections. I notice that Amber adds some additional keyword such as
>>>>>>> 'Force'
>>>>>>>> in route section which restricts Gaussian to calculate forces only.
>> I
>>>>>>> guess
>>>>>>>> it is fine for QM-MM/MD since we need forces for dynamics, but here
>> I
>>>>> am
>>>>>>>> interested in QM/MM optimization only. The route section modified by
>>>>>>> Amber
>>>>>>>> is :
>>>>>>>>
>>>>>>>> *#P OPT B3LYP/6-31G SCF=(Conver=8) NoSymm Force Guess=Read Charge
>>>>> Prop=*
>>>>>>>> * (Field,Read)*
>>>>>>>>
>>>>>>>> How can I disable 'Force' keyword so that I can perform proper
>>>>>>>> optimization. I will be thankful for your help.
>>>>>>>>
>>>>>>>> thanks
>>>>>>>
>>>>>>> --
>>>>>>>
>>>>>>>
>>>>>
>>>>
>> ____________________________________________________________________________
>>>>>>>
>>>>>>> Prof. Gerald MONARD
>>>>>>> SRSMC, Université de Lorraine, CNRS
>>>>>>> Boulevard des Aiguillettes B.P. 70239
>>>>>>> F-54506 Vandoeuvre-les-Nancy, FRANCE
>>>>>>>
>>>>>>> e-mail : Gerald.Monard.univ-lorraine.fr
>>>>>>> tel. : +33 (0)383.684.381
>>>>>>> fax : +33 (0)383.684.371
>>>>>>> web : http://www.monard.info
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>
>>>>
>> ____________________________________________________________________________
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> AMBER mailing list
>>>>>>> AMBER.ambermd.org
>>>>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> With best regards
>>>>>>
>>>>>
>>>>
>> ************************************************************************************************
>>>>>> Dr. Kshatresh Dutta Dubey
>>>>>> Post Doctoral Researcher,
>>>>>> c/o Prof Sason Shaik,
>>>>>> Hebrew University of Jerusalem, Israel
>>>>>> Jerusalem, Israel
>>>>>> _______________________________________________
>>>>>> AMBER mailing list
>>>>>> AMBER.ambermd.org
>>>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>>
>>>>> --
>>>>> Jason M. Swails
>>>>> BioMaPS,
>>>>> Rutgers University
>>>>> Postdoctoral Researcher
>>>>> _______________________________________________
>>>>> AMBER mailing list
>>>>> AMBER.ambermd.org
>>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Thanks & Regards;
>>>> Soumendranath Bhakat
>>>> _______________________________________________
>>>> AMBER mailing list
>>>> AMBER.ambermd.org
>>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>>
>>>
>>>
>>>
>>
>> --
>>
>> ____________________________________________________________________________
>>
>> Prof. Gerald MONARD
>> SRSMC, Université de Lorraine, CNRS
>> Boulevard des Aiguillettes B.P. 70239
>> F-54506 Vandoeuvre-les-Nancy, FRANCE
>>
>> e-mail : Gerald.Monard.univ-lorraine.fr
>> tel. : +33 (0)383.684.381
>> fax : +33 (0)383.684.371
>> web : http://www.monard.info
>>
>>
>> ____________________________________________________________________________
>>
>>
>> _______________________________________________
>> 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
>
--
____________________________________________________________________________
Prof. Gerald MONARD
SRSMC, Université de Lorraine, CNRS
Boulevard des Aiguillettes B.P. 70239
F-54506 Vandoeuvre-les-Nancy, FRANCE
e-mail : Gerald.Monard.univ-lorraine.fr
tel. : +33 (0)383.684.381
fax : +33 (0)383.684.371
web : http://www.monard.info
____________________________________________________________________________
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Jun 13 2014 - 13:00:02 PDT
