[gmx-users] Re: g_wham giving flat free energy

[Stefan Hoorman]

>
> Michael McGovern wrote:
> > Hi everyone.  I'm having some trouble using g_wham to analyze some
> > results.  I have some simulations from gromacs 3.3 using the pull
> > code.  I'm using the latest version of g_wham to analyze the results,
> > which the help file says is ok, using the -ip option.
> >
> > I've gotten almost everything to work, and it generates histograms
> > that seem perfectly reasonable, but sometimes the free energy it
> > outputs is completely flat, with each bin having a free energy value
> > of 0.
> >
> > I've played around with the -b and -e options using only part of my
> > data, and sometimes it gives non-flat free energies, other times not.
> >  I see no logic to when it works and when it doesn't.  My data does
> > have good overlap of the histograms, and it sometimes converges even
> > when I use data that have no overlaps, so that doesn't seem to be the
> > problem.
> >
> > I'm completely confused.  Does anyone know what could cause this?
> >
> >
>
> Hi Micheal,
>
> how large is your data? Could you send me a tar.gz of your pdo files? I
> would like to make sure that there is no bug in g_wham.
>
> Jochen
>
>
I had this problem also when I tried to calculate dimerization free energies
using umbrella sampling and g_wham. The problem seemed to occur every time I
included .tpr files that contained configurations in which there were no
more interaction between the previously dimerized set. What I mean is, when
I monitored the LJ and Coul interactions between my molecules, I found out
that there would be no more interaction beyond, lets say, 2.1 nm (distance
between COM of my molecules). When I included the 2.0, 2.1, 2.2 nm etc
windows in my g_wham analysis, I got a flat free energy graphic.
Just wanted to add that to the post.
Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20100428/f8778484/attachment.html>