[gmx-users] PMF changing drastically with the number of bins

[Sajeewa Pemasinghe]

Dear all,

I am calculating the PMF as the COM distance of two molecules are changing
using umbrella sampling ( pull = umbrella ). Each umbrella sampling
simulation runs for 50ns.  When I calculate the PMF with default (200 bins)
settings using gmx wham I get the minimum in free energy of -1.19 kCal
mol-1 at 24.9 Angstrom. When I change the number of bins to 100 the new
value for minimum is -0.558 kCal mol-1 at 24.8 Angstrom and for 50 bins it
is -0.195 kCal mol-1 at 24.7 Angstrom. Can the free energy value be this
much dependent on the number of bins? I suppose the normalization of total
probability to 1 happens inside the algorithm. Could anyone give me a
suggestion to this much of a variation?

Best regards

Sajeewa