Re: Free energy did not converge by using ABF

From: Vermaas, Joshua (Joshua.Vermaas_at_nrel.gov)
Date: Thu Dec 07 2017 - 19:23:53 CST

Hi Yaxin,

I'm super confused. A water "box" suggests that you have periodic
boundary conditions. How can your alkane molecule move out of water this
way? Wouldn't a value of 45 imply that you are really at -5 and
re-entering the waterbox from the bottom? This should be really obvious
if you visualize a water surface, as I'd expect a hole in the water
surface on the "bottom" that matches where the alkane is on the top.
Normally I see this done using FEP or TI rather than a pulling
technique, since then PBC makes perfect sense and you don't need to pull
your molecule into a vacuum.

-Josh

On 12/07/2017 04:28 PM, Yaxin An wrote:
> Hi All:
>
> I am running ABF simulations for the free energy of hydration of
> alkanes. But I got different values for the free energies every time
> when I repeat the simulations, like 1.5 and 3.0 Kcal/mol.
>
> In the simulation, there was one water box (40 Angstrom by40 by 40).
> The reaction coordinate is the Z distance between COM of the water box
> and the one alkane molecule. The reaction coordinate spans from 0 to 45.
>
> I split the reaction coordinate into 9 windows: 0-5, 5-10, ....,40-45.
> And I ran the simulation in each window for 50 ns. Each time the free
> energy converged. But when I repeated it, the free energy converged to
> a different value.
>
> Hope someone can give me some suggestion.
>
> Thanks
> Yaxin
>
>
>

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:21:49 CST