[gmx-users] Umbrella sampling force autocorrelation

[gozde ergin]

Dear Gromacs user,

I did umbrella sampling by pulling a water molecule from gas phase to
organic coated bulk water.
By clicking the link below you can see my PMF profile.
In order to estimate the transmission coefficient, I put the pulled
molecule on the top of the PMF barrier and did simulation for 600 ps.
I extracted the instantaneous force of z direction on this molecule by
using the command of :

*g_traj -f traj.trr -s md.tpr -of force.xvg -n index.ndx -nox -noy*

And to estimate the auto-correlation for  instantaneous force:

*g_analyze -f force.xvg -ac auto_corr.xvg -temp 300 -nonormalize*

However I wonder why the fluctuation on auto-correlation is big and never
converges to the zero.
Please find the profile.xvg, force.xvg and auto_corr.xvg on
http://imgur.com/a/IVP5f


I would greatly appreciate of any respond.

Best regards