[gmx-users] Re: Simulation in the high temperature conditions

[Tsjerk Wassenaar]

Hey James,

It's reassuring to find you did think of it. The approach may have
merits, or may be a waste of CPU. But the proof of the pudding is in
the eating of it. The connection with NMR data makes sense. The
restraints may be time-averaged, involving different conformational
states, hampering the usual annealing-based structure determination.
But room-temperature simulations may not allow the transitions
required. Hence, applying the same restraints at an elevated
temperature, probably using a higher force constant, could yield the
desired ensemble. To get a room-temperature ensemble by using some
restraints will probably not work, but it's an interesting list of
things that shouldn't have worked according to some, yet worked out
for those who tried :p Nonetheless, applying local restraints to
maintain integrity and use an elevated temperature to enhance
conformational sampling, might actually also be a valuable approach.
Good luck :)

Tsjerk

On Wed, Apr 11, 2012 at 8:27 PM, James Starlight <jmsstarlight at gmail.com> wrote:
> Tsjerk,
>
>
> As Justin already has noticed there were lot of examples of implementation
> of the high temperature aproach to increase conformation sampling. This
> could be usefull because of short time-scale of typical MD trajectory and
> hight energy barriers of the adjacent energy-minimums wich could correspond
> to the functional-relevant conformations.  So the main reason of such
> stimulations in to dicrease such barriers to promote transitions to the
> functional-relevant states conformations Finally I can sample the output of
> such simulation and compare this resulted conformations with the
> experimental data. If I obtain good similarity this may indicate about
> efficiency of such aproach
>
> On other hand such simulation could result in some non-native conformations
> due to the non-native condtions. So the artificial restrains like as the
> distance restrains within native conformation ( obttained by NMR for
> instance) could be usefull-aproach to keep the system in the native-like
> conformations while enhansing sampling of the native-like structures via
> rising of temperatures. So combinations of  such tricks could give
> physical-relevant results partly couldn't it ? However I suppose that the
> main disadvantage of such aproach is the non-physical intermediate states
> produced by non-native transitions pathways wich could arrise from such
> simulation.
>
>
>
> James
>
> 11 апреля 2012 г. 18:09 пользователь Tsjerk Wassenaar <tsjerkw at gmail.com>
> написал:
>
>> Hey James,
>>
>> Have you thought about the physical relevance of your results?
>>
>> Cheers,
>>
>> Tsjerk
>>
>> On Wed, Apr 11, 2012 at 12:01 PM, James Starlight
>> <jmsstarlight at gmail.com> wrote:
>> > Tsjerk,
>> >
>> > Thank you for suggestions!
>> >
>> >
>> > Indeed the hight temperature ( I'm using 700K) which I use for enhansing
>> > sampling rate  resulted in destabilisation of the secondary structure.
>> >
>> > To prevent this I've used two slightly different aproaches based on the
>> > restraint. But in both cases I've used slightly soft restrains with
>> > fs=10-200.
>> >
>> > The first aproach is the ussage of the posres applied on each backbone
>> > atom
>> > with fc=10-50. I've tested case with fc=50 and found that such restrains
>> > were very hight. I've not noticed any conformation sampling of my
>> > protein (
>> > rmsd of backbone < 0.3 nm) during 10ns of such simulation. So I've
>> > decided
>> > to test a case with fc=10 ( under calculation)
>> >
>> > The second approach is the application of the harmonic distance
>> > restrainse
>> > wich I've applied on each backbone atom pair in the CUTOFF radius of 1.0
>> > nm.
>> > The Rc value was chosen because of my protein is the 7 buddle of alpha
>> > helices so this aproach could be usefull but exactly value for R have
>> > been
>> > chosen empirically. I've selected deviation value wich are equal to 1\2
>> > of
>> > cutoff radius = 1.5 nm. I have not realise whaat I could obtain yet from
>> > this because I'm not quite sure about coccect values of such disres.
>> >
>> >
>> > James
>> >
>> > 11 апреля 2012 г. 13:35 пользователь Tsjerk Wassenaar
>> > <tsjerkw at gmail.com>
>> > написал:
>> >>
>> >> Hey :)
>> >>
>> >> I'd say a protein should be loosing structure at 700K. Can't even say
>> >> the force field is wrong there, even though it hasn't been
>> >> parameterized for such temperatures for certain. You'd have to check
>> >> with experiments.
>> >> Trying to do high-temperature simulations is fine. But trying to do
>> >> high temperature simulations to get results that match a room/body
>> >> temperature ensemble is completely bogus. If you aim to use this
>> >> approach for enhanced sampling, you're in for some serious
>> >> reparameterization. Part of the deal may indeed be adding
>> >> long(er)-range distance restraints. However, the force constants to
>> >> use will increase with temperature, and with very high temperatures
>> >> the force constants may end up so high that they give rise to fast
>> >> oscillations, which will require using a smaller time step.
>> >>
>> >> Just my 2 cents...
>> >>
>> >> Tsjerk
>> >>
>> >> --
>> >> Tsjerk A. Wassenaar, Ph.D.
>> >>
>> >> post-doctoral researcher
>> >> Molecular Dynamics Group
>> >> * Groningen Institute for Biomolecular Research and Biotechnology
>> >> * Zernike Institute for Advanced Materials
>> >> University of Groningen
>> >> The Netherlands
>> >> --
>> >> gmx-users mailing list    gmx-users at gromacs.org
>> >> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> >> Please search the archive at
>> >> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> >> Please don't post (un)subscribe requests to the list. Use the
>> >> www interface or send it to gmx-users-request at gromacs.org.
>> >> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>> >
>> >
>> >
>> > --
>> > gmx-users mailing list    gmx-users at gromacs.org
>> > http://lists.gromacs.org/mailman/listinfo/gmx-users
>> > Please search the archive at
>> > http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> > Please don't post (un)subscribe requests to the list. Use the
>> > www interface or send it to gmx-users-request at gromacs.org.
>> > Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>>
>>
>> --
>> Tsjerk A. Wassenaar, Ph.D.
>>
>> post-doctoral researcher
>> Molecular Dynamics Group
>> * Groningen Institute for Biomolecular Research and Biotechnology
>> * Zernike Institute for Advanced Materials
>> University of Groningen
>> The Netherlands
>> --
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
>
>
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists



-- 
Tsjerk A. Wassenaar, Ph.D.

post-doctoral researcher
Molecular Dynamics Group
* Groningen Institute for Biomolecular Research and Biotechnology
* Zernike Institute for Advanced Materials
University of Groningen
The Netherlands