Hi everyone:
I have created a new library (with Pawel's help) for both monovalent
and divalent atomic ions, called "atomic_ions.lib" (created by running
"atomic_ions.cmd" through tleap.) This new library combines things
that were in ions08.lib and ions13.lib.
For ions where the PDB has chosen atom and residue names, we now follow
the pdb conventions; where there is no PDB-chosen name, I kept the
names we have been using in the past. I have not changed any of the
atom *types*, so there are no changes in parm.dat or frcmod files. The
leaprc files that used to load ions08.lib now load atomic_ions.lib. [The
"ions08.lib" file is still present, for backwards compatibility, but I
hope people will stop using it, or will post reasons why the old way is needed
for certain puposes. Since ions13.lib is quite new, I didn't see a need
to keep it; again, please pipe up if you think this is a bad choice.]
The goal is to have as many things as possible in the Amber libraries use PDB
atom and residue nomenclature; I think this is a big win for data interchange,
and a big help to most users. We are trying to agree with the PDB where
possible, although carbohydrates and lipids are big exceptions. (One thing we
need to track and document: where do the residue names we use conflict with
other uses in the PDB components library?)
As a heads-up, Pawel is working with the phenix crew to create pre-built
amber libraries and frcmod files for a large fraction of the 17,626
entries in the components.cif file from the PDB. Once this is completed,
it will be much easier to create plausible MM models for most of the
entries in the PDB, and Amber's atom and residue names will match those
used by the PDB.
Comments/suggestions are welcome. The atomic_ions change is quite small, and
it is my hope/intent to include this in the upcoming release. The library
of pre-built component libraries doesn't even exist yet, and will won't be
distributed until after a lot more testing.
...dac
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Fri Feb 21 2014 - 12:00:03 PST
