[gmx-users] EnMin & constraints: Hamletic doubt

Justin Lemkul jalemkul at vt.edu
Fri Jun 20 15:06:22 CEST 2014 


On 6/20/14, 8:58 AM, Nicola Staffolani wrote:
> Sorry, that's true, I am mixing things up! I am using "constraints = none"
> for EnMin (and here our findings apply → Δt = .5 fs (at least)) & then

dt is irrelevant during energy minimization; only emstep matters since the 
minimization algorithms are not dynamical.

> "constraints = h-bonds" for the MD (and then so here I can probably safely
> enough set Δt = 1 - 2 fs...
>

Beware that if you are using constraints in the MD, you should minimize with 
constraints, otherwise the minimized structure may have bond lengths that cause 
the constraint algorithm to fail.

-Justin

> Thank you and sorry for being not sufficiently clear!!
>
> Nicola
>
>
> On 20 June 2014 14:36, Justin Lemkul <jalemkul at vt.edu> wrote:
>
>>
>>
>> On 6/20/14, 8:34 AM, Nicola Staffolani wrote:
>>
>>> .5 fs even with h-bonds?
>>>
>>>
>> No, you can use 1-2 fs with h-bonds, but based on your citation of a
>> previous post that did not use constraints and then you said:
>>
>> "And I was thinking that we should remove the constraints if we want to
>> see the appropriate oscillations!"
>>
>> it sounded like you would not be using constraints.
>>
>> -Justin
>>
>>
>>
>>> On 20 June 2014 14:29, Justin Lemkul <jalemkul at vt.edu> wrote:
>>>
>>>
>>>>
>>>> On 6/20/14, 4:01 AM, Nicola Staffolani wrote:
>>>>
>>>>   On 19 June 2014 23:59, Justin Lemkul <jalemkul at vt.edu> wrote:
>>>>>
>>>>>
>>>>>
>>>>>> On 6/19/14, 10:05 AM, Nicola Staffolani wrote:
>>>>>>
>>>>>>    Dear Justin,
>>>>>>
>>>>>>>
>>>>>>> thank u 4 replying!
>>>>>>>
>>>>>>>
>>>>>>> On 19 June 2014 15:04, Justin Lemkul <jalemkul at vt.edu> wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>   On 6/19/14, 4:11 AM, Nicola Staffolani wrote:
>>>>>>>>
>>>>>>>>     Dear GMX community,
>>>>>>>>
>>>>>>>>
>>>>>>>>> ​I am running simulations of a solvated protein on a (frozen) Au
>>>>>>>>> substrate,
>>>>>>>>> but I think that my question is so basic that it falls outside the
>>>>>>>>> details
>>>>>>>>> of my case...
>>>>>>>>>
>>>>>>>>> So, the question is: when I prepare the system by minimizing the
>>>>>>>>> energy,
>>>>>>>>> should I set constraints = all_bonds, h_bonds or none?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>    What was your force field parametrized to use?  Most commonly,
>>>>>>>>> modern
>>>>>>>>>
>>>>>>>> force fields constrain bonds involving H, though all bonds are
>>>>>>>> frequently
>>>>>>>> constrained in practice because the differences are generally small
>>>>>>>> and
>>>>>>>> the
>>>>>>>> rigid representation of a bond is considered by many to be a more
>>>>>>>> realistic
>>>>>>>> representation of the ground state than a harmonic function.
>>>>>>>>
>>>>>>>> I do not understand your question... ​Do you mean which force field I
>>>>>>>> am
>>>>>>>>
>>>>>>>>    using? In this case the answer is gromos43a1​...
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>    Constraining all bonds is common with Gromos96 parameter sets.  I
>>>>>>> know
>>>>>>>
>>>>>> that 53a6 was parametrized with all bonds constrained; I am not sure if
>>>>>> the
>>>>>> full details of 43a1 were ever published.  If someone else knows,
>>>>>> please
>>>>>> chime in.  The oldest public reference I know of for Gromos96 is a
>>>>>> paper
>>>>>> about the software suite in general, but not the original
>>>>>> parametrization
>>>>>> protocol.
>>>>>>
>>>>>>
>>>>>>        IMHP, I should set it to all_bonds, otherwise I may risk that the
>>>>>>
>>>>>>>
>>>>>>>   system's
>>>>>>>>
>>>>>>>>    energy minimum corresponds to a configuration very different from
>>>>>>>> the
>>>>>>>>
>>>>>>>>> initial one... However, I saw  somebody is using constraints = none
>>>>>>>>> and
>>>>>>>>> I
>>>>>>>>> was wandering why...
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>     Without context, answering that is impossible.
>>>>>>>>>
>>>>>>>>>
>>>>>>>> ​I see... How can I better describe the context ? Which info do u
>>>>>>>> need? ​
>>>>>>>>
>>>>>>>>    ​I have seen setting the constraints to none with the same very
>>>>>>>>
>>>>>>> system as
>>>>>>> mines (solvated azurin on Au​
>>>>>>> ​ layers) where the ff used was again gromos43a1... ​
>>>>>>>
>>>>>>>
>>>>>>>    The context to which I refer is: what were the goals of the study?
>>>>>>>   If
>>>>>>>
>>>>>> one
>>>>>> wants to examine, for instance, vibrational spectra, it would be
>>>>>> inappropriate to apply constraints (and you'd be forced to use a very
>>>>>> small
>>>>>> time step).  For normal MD investigations, constraints are used to
>>>>>> increase
>>>>>> the time step for better throughput.  My comment about context is also
>>>>>> a
>>>>>> general one.  People often mention previous posts without providing
>>>>>> links
>>>>>> or looking at the whole thread.  There may be very specific reasons
>>>>>> for a
>>>>>> particular approach, or it may turn out that the posted settings were
>>>>>> wrong
>>>>>> ;)
>>>>>>
>>>>>>
>>>>>>    ​Thanks for explaining! Indeed we are in a early phase of our study
>>>>>> and
>>>>>>
>>>>> are
>>>>> interested in the vibrational (Raman) spectra (!!) of Cu-ligands; we
>>>>> want
>>>>> to see whether we are able to reproduce the experimental spectra in
>>>>> order
>>>>> to validate our model and then go further... And I was thinking that we
>>>>> should remove the constraints if we want to see the appropriate
>>>>> oscillations!
>>>>>
>>>>> It is also very interesting your comment about the relation between time
>>>>> step duration and aim of the MD simulation: I had been suggested to use
>>>>> Δt
>>>>> = 0.002 ps, but then I had issues with the frozen Au layers and changed
>>>>> it
>>>>> to 0.001 ps. Afterwards, in my last runs, I changed some other
>>>>> parameters
>>>>> and decided to give a try again to 0.002 ps, but it seems more
>>>>> appropriate
>>>>> - thanks to your comment - to switch back to 0.001 ps...
>>>>>
>>>>>
>>>>>   Without constraints, you may need an even smaller time step, 0.5 fs at
>>>> most, for proper energy conservation.
>>>>
>>>> -Justin
>>>>
>>>> --
>>>> ==================================================
>>>>
>>>> Justin A. Lemkul, Ph.D.
>>>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>>>
>>>> Department of Pharmaceutical Sciences
>>>> School of Pharmacy
>>>> Health Sciences Facility II, Room 601
>>>> University of Maryland, Baltimore
>>>> 20 Penn St.
>>>> Baltimore, MD 21201
>>>>
>>>> jalemkul at outerbanks.umaryland.edu | (410) 706-7441
>>>> http://mackerell.umaryland.edu/~jalemkul
>>>>
>>>> ==================================================
>>>> --
>>>> Gromacs Users mailing list
>>>>
>>>> * Please search the archive at http://www.gromacs.org/
>>>> Support/Mailing_Lists/GMX-Users_List before posting!
>>>>
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>> * For (un)subscribe requests visit
>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>> send a mail to gmx-users-request at gromacs.org.
>>>>
>>>>
>>>
>>>
>>>
>> --
>> ==================================================
>>
>> Justin A. Lemkul, Ph.D.
>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>
>> Department of Pharmaceutical Sciences
>> School of Pharmacy
>> Health Sciences Facility II, Room 601
>> University of Maryland, Baltimore
>> 20 Penn St.
>> Baltimore, MD 21201
>>
>> jalemkul at outerbanks.umaryland.edu | (410) 706-7441
>> http://mackerell.umaryland.edu/~jalemkul
>>
>> ==================================================
>> --
>> Gromacs Users mailing list
>>
>> * Please search the archive at http://www.gromacs.org/
>> Support/Mailing_Lists/GMX-Users_List before posting!
>>
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>> * For (un)subscribe requests visit
>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>> send a mail to gmx-users-request at gromacs.org.
>>
>
>
>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================

More information about the gromacs.org_gmx-users mailing list