On Wed, May 23, 2012, Jason Swails wrote:
> On Wed, May 23, 2012 at 11:24 AM, marawan hussain wrote:
>
> > I feel like there are some issues with the MMPBSA.py in this version of
> > AMBER (12)..From the bunch of posts i have seen i'm a bit afraid to use the
> > results i got form the script..Some errors even may be silent and can't be
> > easily figured out..I appreciate if you can release a new version of the
> > script that we can safely use...
> >
>
> There have been a number of issues that have cropped up, but these have all
> been either poor error handling or attempts to get AmberTools12 working
> with Amber11 for decomposition (there was a recent thread talking about
> this). To clarify, the errors reported here have been instances in which
> MMPBSA.py was not used correctly, but the error message was not helpful.
>
> Anything tested for in the test suite (AmberTools/test/mmpbsa_py/) should
> be as reliable as the previous version was. As deficiencies in the error
> messages are reported, we will fix them.
To add to Jason's comments, I'd like to quote from the opening of Chapter
10 in the AmberTools12 Reference Manual:
"None of these should be considered as a “black-box”, and users should be
familiar with Amber before attempting these sorts of calculations. These
scripts automate a series of calculations, and cannot trap all the types of
errors that might occur. You should be sure that you know how to carry out an
MM-PBSA calculation “by hand” (i.e., without using the scripts); if you don’t
understand in detail what is going on, you will have no good reason to trust
the results."
To add some specific points:
1. The pbsa code in Amber now has *lots* of options, and may simply be giving
some people enough rope to hang themselves. If you are using pbsa,
be sure that you have run some of the calculations by hand, and have carefully
considered what options you want. I'd suggest starting with the simplest
models (using radii from the prmtop file, just a surface area non-polar term);
you can then carefully evaluate more advanced features like "optimized" radii
and more elaborate non-polar energy terms.
2. Using the mm-pbsa model where mono-atomic ions like Mg2+ are an important
feature of the affinity is a dangerous task, and I am not sure that anyone
knows how to do this in a reliable fashion.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed May 23 2012 - 04:30:03 PDT
