On Tue, Jun 28, 2005, Thomas E. Cheatham, III wrote:
> I am going to rebuild the prmtops w/ LEaP and parm94 to verify that I
can
> see what Ross does (as I did parm94 with PLEP not parm99).
I think(?) that all of the stuff Ross picked up on is unique to
parm99.dat:
that is the only place where there are the specific sugar-based torsions
(added by Junmei) that potentially conflict with wild-card torsions. As
I understand it, there is no corresponding conflict in parm94.dat.
I have created an updated version of parm99.dat that avoids all
specific-vs-generic torsions, and I will passing this around soon for
developers to look at (assuming tests work for me).
[I've also created some new versions of parm94.dat, etc. that fix an
unrelated problem: that LEap will occasionally use the wrong order for the
three peripheral atoms in an improper torsion. This has no real effect on
the
force field, but can show up in detailed comparisons to other programs.]
I don't think people should have their hair on fire too quickly:
independent
implementations of the Amber force field (e.g. most recently in GROMACS,
see
http://folding.stanford.edu/ffamber/) get results essentially
identical to
Amber. There _is_ some god-awful code regarding wild-card vs.
non-wild-card
torsions in LEaP, but we can work through this. Please be sure you are
comparing apples to apples (e.g. parm94 to parm94, or parm99 to parm99).
I'll report more about what I learn as it happens.
....dac
p.s. {tom: since you are obviously "alive", can we talk about the Sponer
book chapters, and about scheduling the fall Amber-developers' meeting?}
Received on Wed Apr 05 2006 - 23:49:55 PDT
