Re: MetaDynamics in 2.9b2 Question

From: Giacomo Fiorin (giacomo.fiorin_at_gmail.com)
Date: Fri Nov 09 2012 - 11:38:10 CST

Hello Robert, yes, the energy of metadynamics is counted in the MISC
section. You don't see this explicitly for metadynamics because it is a
feature shared by all biases (metadynamics, harmonic, abf), through
inheritance from the base class.

Thank you for taking the initiative to develop a very useful set of scripts!

Giacomo

On Fri, Nov 9, 2012 at 12:17 PM, Robert Elder <rmelder_at_gmail.com> wrote:

> As a follow-up to the discussion below, I have a question: is the energy
> from the metadynamics history-dependent bias reported by NAMD under the
> MISC header?
>
> More specifically, I would like to perform replica-exchange metadynamics,
> and I need to ensure that the full potential energy is used in the
> Metropolis criterion. I already have the RE-metadynamics scheme
> *functional* but need to ensure that it is *correct*. I've looked through
> the NAMD source code and it seems that all the Colvar bias energies
> (including metadynamics) are reported under MISC, but I am very new to the
> code so any advice would be helpful.
>
> Thanks to Aron Broom and Dr. Fiorin for the useful exchanges that have
> allowed me to make it this far :)
>
> --
> Robert Elder
> PhD Student
> Department of Chemical and Biological Engineering
> University of Colorado-Boulder
>
> On Wed, Apr 4, 2012 at 6:32 PM, Giacomo Fiorin <giacomo.fiorin_at_gmail.com>wrote:
>
>> Hello Aron, yes to (1), but since this is new code, so keep checking that
>> the contents of the PMF files across all replicas are the same (or close
>> to).
>>
>> About (2), I'm not even aware there is a formalism to handle more than
>> one level of replica-parallelism at once.
>>
>> Anyway, temperature-exchange should be possible, for as long as you
>> include the metadynamics energy in the Metropolis criterion. Likewise,
>> bias-exchange is also possible by swapping the state files.
>>
>>
>> On Wed, Apr 4, 2012 at 9:54 PM, Aron Broom <broomsday_at_gmail.com> wrote:
>>
>>> I had two questions concerning metadynamics in NAMD 2.9b2, the first
>>> quite specific to 2.9b2, and the other perhaps is more general and could
>>> relate to 2.8 also.
>>>
>>> 1) In 2.8 multiple replicas (multiple walkers) metadynamics was not
>>> possible while useGrids was enabled, but I see the particular note
>>> concerning this has been removed from the 2.9b2 user guide. It does,
>>> however, say that a *.hills file is used for each replica as a kind of
>>> communication buffer, which perhaps suggest that the hills still need to be
>>> recorded analytically? Anyway, I'm wondering if it is possible now to do
>>> multiple walkers/replicas for metadynamics while also having the benefit of
>>> the grids.
>>>
>>> 2) Is it possible (has anyone done this in NAMD?) to use the temperature
>>> replica exchange Tcl script to perform Replica-Exchange metadynamics? I
>>> don't see why it would be different in form from any other
>>> temperature-based replica exchange method, but as I'm no expert in these
>>> things I could easily be missing something. Moreover, the section on
>>> temperature replica exchange claims that the scripts are in spirit easily
>>> adapted for exchange based on things other than temperature. Is it
>>> therefore possible to perform bias-exchange metadynamics? Could one even
>>> combine both of those in order to do nested replica exchange for different
>>> parameters (temperature and bias for instance)?
>>>
>>> Thanks,
>>>
>>> ~Aron
>>>
>>> --
>>> Aron Broom M.Sc
>>> PhD Student
>>> Department of Chemistry
>>> University of Waterloo
>>>
>>>
>>
>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:22:44 CST