Re: Asking for Free energy calculation advice

From: Aron Broom (broomsday_at_gmail.com)
Date: Fri Dec 06 2013 - 11:48:03 CST

If the force is being applied, but its not high enough to help it explore
the reaction coordinate, you can perhaps wait longer, as the force will
keep adapting. The other option may be to change the "fullSamples" value
to something larger, such that you have a more accurate approximation of
the force before it gets applied.

In particular, the concern with the latter case is that your simulation has
overestimated the force needed to reach the wall and is now holding it
there. You say that it "soon found its way to the wall", but that implies
it was spontaneous, whereas, it seems more likely that your ABF force is
what drove it there, and if the "fullSamples" value was too low, it may
have become set at too high of a force, and now it will be very hard to
change that, as it needs to sample the solvated coordinate in order to
update the force.

If you want to completely avoid the above issue, you can use the same
colvars, but use metadynamics instead. Of course it comes with it's own
issues.

~Aron

On Thu, Dec 5, 2013 at 3:18 PM, Kasra Fattah <kasra.fattah_at_gmail.com> wrote:

> Hi All,
>
> I'm doing a free energy calculation of a nano-particle with the reaction
> coordinate being the center of the mass distance of the particle from a
> wall with a water as solvent surrounding the particle and the wall. I
> started the simulation with the particle far above the wall as it's
> hydrophobic it soon find its way to land and stays on the wall. I'm doing
> an ABF with colvars, after the particle reaches the wall it stays there for
> a very long time, actually the simulation is still running and 8ns has
> already passed but after the initial 2ns that took for the particle to
> reach the wall it's spending rest of it on the wall.
> Although the ABF is doing its job to negating the force on the particle
> (as I checked it in the trajectory file of the colvars) but it how can I
> make the particle to move around and sample the reaction coordinate? it's
> like only the few bins near the wall are only sampled as the particle is
> stuck there (maybe because it's waiting for the diffusion to kick it out of
> that position?!) ...I would really appreciate it if you can comment on this
> kind of situation and how I can circumvent this situation.
>
> Cheers,
> Kasra.
>

-- 
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 - 23:24:04 CST