On Mon, Mar 30, 2015 at 1:02 AM, Shubhadip Das <d.shubhadip.yahoo.com>
wrote:
> > Dear user,
> >
> > I am doing umbrella sampling in AMBER12 and to construct PMF i have used
> > WHAM. The umbrella sampling was conducted for 3.0 A^0 to 14.5 A^0. which
> > was divided into 24 bins with window lengths of 0.5 A^0. Using
> > K=25kcal/mol/A^0^2. I run each an every simulation upto 10ns and i have
> > checked that windows have overlap properly. And next to run WHAM i have
> > used the following command
> > /home/shubhadip/WHAM/wham/wham/wham 2.5 15.0 24 0.01 298 0 meta.dat
> results.dat
> > but still when i was changing the hist_min and hist_max by 0.1 A^0 the
> > shape of the PMF was changed abruptly. Am i doing something wrong?
>
​What you describe is not necessarily unusual behavior for methods relying
on discrete binning with a finite bin size. But "shape of the PMF changed
abruptly" is a very vague description. I have no idea what that actually
means (did it go from concave-down to concave-up? Did it go from smooth to
fractal-like? Did it just change barrier heights more than you
expected?). Pictures will help illustrate what you are observing better
than any description you can give will do. How much data did you feed WHAM?
Another thing you should check is *unweighted* binning. Histogram the raw
data. If you adjust the histogram dimensions by 0.1 A on either side (but
still keep the same number of bins), do you also see an "abrupt shape
change" in the distribution? If so, then it seems that your data is more
likely to be the culprit than WHAM.
One final note: histogramming often requires a lot of data -- at least a
*lot* more data than you have bins. If you don't have a lot of data, then
changing the bin locations and widths, even slightly, can give rise to
unexpectedly large changes in the apparent shape of the distributions.
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Mar 30 2015 - 06:01:50 PDT
