On Tue, Feb 17, 2009, Arnaud wrote:
>
> I perfomed three time the same PMF calculation on the same complex (a drug
> and receptor) but using amber8, amber9 and amber10. In each case, PMF was
> evaluated from the WHAM post-processing. For all my calculations parameters
> were strictely the same (same condition, same time simulation, same windows,
> same constraint force, ...)
>
> For all simulations, dissociation pathways were similar but PMF values
> obtained with amber8 were twice those obtained with amber9 or amber10.
I don't understand the result you found. A PMF is a function of some
pathway coordinate, and its zero of energy is arbitrary. What do you
mean when you say that the PMF values in one case were "twice" those in
another case?
You should be able to take the *same* prmtop and restart files, and run
a short simulation in each of the three codes...that should pin down any
differences that may exist. Since energy and restraint part of Amber
has not changed in many years, I would not expect to see such changes,
but it is best to check. Especially since something(?!?) must have led
you to perform the same calculation three times.
If individual energies and restraints are the same, look at the actual
distributions of points along the reaction, not just what WHAM puts out as
the final PMF. Of course, you should also check is whether your PMF's are
indeed converged.
...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Feb 18 2009 - 01:24:22 PST
