Re: Accelerated MD in NAMD

From: James Starlight (jmsstarlight_at_gmail.com)
Date: Fri Nov 08 2013 - 00:53:41 CST

So I'd like to make some suggestions about boost. Firstly I'd like consider
only dihedral boost for simplicity. The hight <Udihe> meas that <on
average> molecule tends to explore "tense" dihedral values (angles= 120,
240, 360 ) which increase its <Udihe>. So addition of the extra boost will
further tend to increase frequency being in that tense angle range.
Assuming that this boost will move system from ony potential well across
the barier to the another well (Accompanied by the rotations about
dihedrals-> conformational change) how we could conclude about
physical-meaningful of such transitions ? (In other worlds could such
escaping from minimum exactly bee seen in the real systems (as I've
understood it call as the satisfaction of the Boltzmann distribution in
terms of the statistics) in case of energy contribution from any external
factors but not artificial dihedral term which looks for me very
unnatural-> sorry I'm biochemist not a physicist).

One extra example could arise from the total boost where such boost apply
on the total U. For example my membrane protein has very negative
contribution from the Coulomb term (due to the big number of the salt
bridges in the protein hydrophobic interior, assuming low dielectric in
that fields). Addition of the boost will increase the <Ucoul> so IT might
be supposes that it could result in the destabilization of the salt bridges
(Accompanied to the increase in the Ucoul) and moving protein to another
potential well ( conformation). Does this statements correct? How I could
make suggestions (based on the above mentioned statistical re-welling)
about possibility of the addition of such artificial boosts in the system?

James

2013/11/7 Jeff Wereszczynski <jwereszc_at_iit.edu>

> The difference in force field could certainly be the issue.
>
> Jeff Wereszczynski
> Assistant Professor of Physics
> Illinois Institute of Technology
> http://www.iit.edu/~jwereszc
>
>
> On Thu, Nov 7, 2013 at 1:18 PM, James Starlight <jmsstarlight_at_gmail.com>wrote:
>
>> Jeff,
>>
>> thanks again for suggestion!
>>
>> might the difference in force fields be the source of such BIG difference
>> in the actual values of the potential? As I wrote previously for the same
>> system as in the Irina paper I have total of 10.000 Kcal/mol in charm ff
>> (2000 from protein and 8000 from lipids) versus 1500/1000 in the Irina's
>> case with amber ff. Or with my system is something wrong (Actually I've
>> analyzed this trajectory and didnt notice any strange behavior in terms of
>> RMSD, SS preservation etc) ?
>>
>> James
>>
>>
>>
>>
>> 2013/11/7 Jeff Wereszczynski <jwereszc_at_iit.edu>
>>
>>> Something else to keep in mind is that most of the back of the envelope
>>> estimates for initial aMD parameters that have been quoted in this thread
>>> are based upon the amber force field. The charmm force field (which was
>>> used in this study) has different dihedral energies, so that could
>>> contribute to different boosting parameters.
>>>
>>> Also, I don't see any reason from a theoretical standpoint why you
>>> couldn't use aMD with an external electric field. I don't know of anyone
>>> that's done that, but it should be straightforward. Your potential will be
>>> position dependent, but I would think the overall electrostatic energy of
>>> the system would dominate your potential function anyway, so there
>>> shouldn't be too much worry about needing position-dependent boost terms
>>> (which is good, because that would be difficult to do from a computational
>>> standpoint). The only question is if the code is set up to properly at
>>> your energy from the electric field into the boost. My guess is yes, but
>>> thats just a guess. I'd say give it a shot and compare the aMD outputs for
>>> applied boosts to what you would expect from the aMD equations and make
>>> sure they match.
>>>
>>> Cheers,
>>>
>>>
>>> Jeff Wereszczynski
>>> Assistant Professor of Physics
>>> Illinois Institute of Technology
>>> http://www.iit.edu/~jwereszc
>>>
>>>
>>> On Thu, Nov 7, 2013 at 9:59 AM, Thomas Evangelidis <tevang3_at_gmail.com>wrote:
>>>
>>>>
>>>>
>>>>
>>>>> 3. You might also find this paper interesting:
>>>>> http://www.pnas.org/content/110/27/10982.full
>>>>> There is some discussion in the SI about choosing aMD parameters for a
>>>>> dual-boost setup for a different GPCR that may be helpful.
>>>>>
>>>>> Very good one! The first one I see to implement dual-boost aMD on the
>>>> whole protein-membrane system. The authors introduce a different empirical
>>>> rule to choose dihedral Energy cutoff and alpha value, but use the old rule
>>>> for the total Potential Energy threshold and the respective alpha:
>>>>
>>>> Edihed = Vdihed_avg + lamda*Vdihed_avg
>>>> αdihed= lamda*Vdihed_avg/5
>>>> Etotal = Vtotal_avg + 0.2*Natoms
>>>> α total = 0.2*Natoms
>>>>
>>>>
>>>> Jeff, since you are very familiar with the physics behind aMD and its
>>>> effects on system dynamics, I wonder if you could comment on my recent post
>>>> about combination of aMD with an external electric field:
>>>>
>>>>
>>>> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2013-2014/1718.html
>>>>
>>>> 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/
>>>>
>>>>
>>>>
>>>
>>
>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:23:58 CST