Re: ABF Displacement

From: Jérôme Hénin (
Date: Mon Mar 29 2010 - 12:10:04 CDT

Hi Patrick,

PMFs are meaningful only up to an additive constant, which means that
you are free to add or subtract a constant so as to redefine the
zero-free-energy point. For distance coordinates, a common choice is
to let the PMF tend towards zero at large distances, meaning "no
interaction". Another convenient option is to set the free energy to
be zero at the measured global minimum, so all free energies are
measured with respect to this "ground state".

Note that the current code (colvars module, in NAMD 2.7b1 and later)
applies the latter convention by default.


On 27 March 2010 20:42, Patrick Yee <> wrote:
> Hi NAMD users,
> My question is about the data file generated using the ABF method. I
> tried running two simulations, one as the NaCl tutorial and one as a
> methane-methane simulation in water, to find the PMF's of both
> systems. My results come out with great accuracy, but each simulation
> has a high displacement from expected values. For example, the PMF for
> the Na-Cl ion pair was calculated to be about 12 kcal/mol high for
> each value, but after translating my data to a known value, results
> looked fine. Is the displacement on expected results considered normal
> behavior for using the ABF method or is there a problem with them?
> Please let me know if you need any other parameters than those
> included below.
> After equilibrating the systems, I ran an NVE simulation with ABF parameters:
> abf coordinate            distance
> abf abf1                     1
> abf abf2                     2
> abf ximin                   0.05
> abf ximax                  14.95
> abf dxi                       0.1
> abf fullsamples           4000
> abf forceconst            10.0
> writexifreq                  10000
> Energy in the ABF simulations was fairly constant and rigidbonds on
> water molecules was not used.
> Thank you very much!
> Patrick

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:53:57 CST