I have looked into it, and found the problem. AmberTools 12 never had this
problem, since the statistics were completely redone in a safer(/better)
way.
Please apply the attached patch from AMBERHOME using the commands:
cd $AMBERHOME
patch -p0 -N < std_dev.patch
cd AmberTools/src/mmpbsa_py && make install
Then you should be able to use -rewrite-output to get the correct standard
deviations. The bug was subtle enough such that the test case actually
didn't catch it for me (a coincidence based on the number of frames used in
the test).
I suggest upgrading to AmberTools 12, although it is still slightly
cumbersome to use AmberTools 12 with Amber 11 to perform decomposition
analysis.
Thanks for the report!
Jason
On Tue, May 29, 2012 at 6:42 AM, Jason Swails <jason.swails.gmail.com>wrote:
>
>
> On May 28, 2012, at 11:52 PM, "Cao Yang" <muxiachuixue.163.com> wrote:
>
> > Hello Bill,
> >
> > I got the same problem . When I use AmberTools1.5(with .bugfix27), the
> Std. Dev. is very high. Dose AmberTool12 help in this case?
>
> Yes. I still have to look into why AmberTools 1.5 has such poor standard
> deviation estimates for the binding entropy. (indeed, there is a lot I need
> to look into). However, AmberTools 12 handles statistics in a more reliable
> way, and I've already verified that this problem is fixed in that version.
>
> Thanks!
> Jason
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Candidate
> 352-392-4032
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue May 29 2012 - 10:30:04 PDT
