[gmx-users] Umbrella Sampling - no PMF plateau

[Steven Neumann]

Dear Gmx Users,

I run US simulations between 2 nanotubes with attached proteins with
distance as a reaction coordinate. This is a coarse-grain simulation, both
tubes are placed across pbc so infinite in length. I have tabulated
potentials for both bonded and non-bonded interactions. I observed that
even at large distances between them in US windows when they do not
interact (even across pbc) the PMF is still increasing... no plateau is
observed...
I tried it in many systems increasing the box size but still the same
happening. The perido image is 15 nm away and the cutoff in potentials is 2
nm...

Would you please advise? May that correspond to position restraint dynamics
of tube atoms of one of them (from which I pulled the other one)? Or maybe
tabulated potentials used?

Steven