[gmx-users] Hamiltonian Replica Exchange

[Szilárd Páll]

PS: [Perhaps stating the obvious] Using an overly aggressive
Hamiltonian scaling  will only result in bad mixing at high
temperatures and therefore low efficiency, hence wasted compute time,
but it should not "hurt" your results. It is still quite useful to
ensure that you're not sampling an entirely different system at small
lambda factors.
--
Szilárd


On Wed, Jun 25, 2014 at 6:21 PM, Szilárd Páll <pall.szilard at gmail.com> wrote:
> Hi,
>
> Next time, you should perhaps use "plz, ..." (and other eye-catching
> formatting marks); a nice mix of font colors and typefaces could also
> help to highlight your questions. :D
>
> On Tue, Jun 24, 2014 at 10:32 PM, Thomas Evangelidis <tevang3 at gmail.com> wrote:
>> Greetings,
>>
>> I want to use the HREX implementation of GROMACS to study the dynamics of a
>> heterodimeric protein. The structure is a two helix bundle (two helical
>> monomers that are wrapped around each other) with disordered ends. I am
>> mainly interested in the dynamics of the disordered ends because I know
>> from NMR that the rest remains structured.
>
>> My question is, can I scale
>> selectively the Hamiltonian of the disordered ends whilst leaving the
>> Hamiltonian of the rest of the protein untouched in order to preserve the
>> dimeric structure?
>
> I'm no expert in protein simulations, so take this with a grain of salt. ;)
>
> Unless your helix bundle is rather unstable, I think your proposal is
> fine. However, why not try it first? Generate the scaled topologies
> and run the one with the highest effective temperature separately to
> assess the stability of the bundles.
>
>> Otherwise I 'll have to impose distance and secondary structure restraints
>> which will slow down the computations and render the dynamics of the
>> structured part unphysical. Is it possible to increase the force constant
>> of the harmonic restraints as lambda decreases to attenuate the stiffness
>> of the helices?
>
> Restraints are indeed sometimes required (and the above suggested
> experiment should tell whether that's the case, I think), but the
> lucky thing in the less than ideal topology hacking-based PLUMED
> approach is that you will actually by definition have N different
> inputs which can have differences not only in the scaled interactions,
> but e.g. in restraint strength.
>
>> The other alternative will be to use much fewer replicas (up to lambda ~0.8
>> to be on the safe side) thus with slower sampling.
>
> If you want to be on the safe side and you observe is a sudden change
> in the stability of the helix bundle from a certain lambda, you can
> just stick to scaling factors lower than this. Assessing what maximum
> lambda is reasonable is something you should probably anyway do.
>
> Cheers,
> Sz.
>
>> thanks,
>> Thomas
>>
>> --
>>
>> ======================================================================
>>
>> Thomas Evangelidis
>>
>> PhD student
>> University of Athens
>> Faculty of Pharmacy
>> Department of Pharmaceutical Chemistry
>> Panepistimioupoli-Zografou
>> 157 71 Athens
>> GREECE
>>
>> email: tevang at pharm.uoa.gr
>>
>>           tevang3 at gmail.com
>>
>>
>> website: https://sites.google.com/site/thomasevangelidishomepage/
>> --
>> 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.