[gmx-users] Constraints & Restraints

[Ran Friedman]

Erik Marklund wrote:
> Mark Abraham skrev:
>> Darrell Koskinen wrote:
>>> Hi Tsjerk,
>>> So then, if I understand correctly, setting "constraints =
>>> all-bonds" is not as realistic as setting "constraints = none",
>>> since the latter will allow for flexible (e.g. harmonic) behavoir
>>> which is more realistic than fixing the bond to a certain distance,
>>> correct?
>>
>> Actually not. It would be a better model of a harmonic potential
>> (duh), but it has been shown that the use of constraints can lead to
>> a (more?) acceptable model of a real system, and they allow a larger
>> integration time to boot. Check out the papers for the constraint
>> algorithms (refs in GROMACS manual).
>>
>> Mark
>>
> Agreed. And this is especially true for hydrogen atoms as I understand
> it, since their behaviour as quantum particles deviate more from a
> classical treatment than is the case for heavier nuclei. This is
> mentioned in the gromacs manual.
>
> /Erik
This depends on the system you study. In some cases it is necessary not
to constrain the hydrogen atoms to get a better agreement with the
experiment. Also,  if one needs to deal with vibrational spectra
involving hydrogens, they must be mobile.

Ran.
>>>> Date: Wed, 11 Nov 2009 21:45:44 +0100
>>>> From: Tsjerk Wassenaar <tsjerkw at gmail.com>
>>>> Subject: Re: [gmx-users] Constraints & Restraints
>>>> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>>>> Message-ID:
>>>>     <8ff898150911111245u63c6aa65sa839f246343521f4 at mail.gmail.com>
>>>> Content-Type: text/plain; charset=ISO-8859-1
>>>>
>>>> Hi Darrell,
>>>>
>>>> Constraints and restraints also apply to relative positions. A bond
>>>> constraint fixes the bond to a certain distance. constraints =
>>>> all-bonds means that all bonds are to be converted to constraints,
>>>> rather than have them flexible, e.g. harmonic. Harmonic bonds are
>>>> actually more like restraints, penalizing deviations from the
>>>> equilibrium values. These equilibrium values and the 'penalty
>>>> function' are described in the force field.
>>>>
>>>> Hope it helps,
>>>>
>>>> Tsjerk
>>>>
>>>> On Wed, Nov 11, 2009 at 9:32 PM,  <darrellk at ece.ubc.ca> wrote:
>>>>  
>>>>> Hi,
>>>>> I just have a quick question on contraints and restraints. My
>>>>> understanding is that "constraints" fix the position of an atom in
>>>>> space and "restraints" restrain the deviation of the atom's position
>>>>> from its equilibrium point. Is that correct? If so, then I am a
>>>>> little
>>>>> confused by the purpose of "constraints = all-bonds" or "constraints
>>>>> = none" in an mdp file, since by selection of a force field, which
>>>>> has
>>>>> bond/angle/dihedral stretching/bending/torsion constants, we are
>>>>> specifying the constraints applied to the simulation. So then what is
>>>>> the purpose of "constraints = all-bonds" and "constraints = none"?
>>>>>
>>>>> Thanks.
>>>>>
>>>>> Darrell
>>>>>     
>
>


-- 
------------------------------------------------------
Ran Friedman
Postdoctoral Fellow
Computational Structural Biology Group (A. Caflisch)
Department of Biochemistry
University of Zurich
Winterthurerstrasse 190
CH-8057 Zurich, Switzerland
Tel. +41-44-6355559
Email: r.friedman at bioc.unizh.ch
Skype: ran.friedman
------------------------------------------------------