[gmx-users] topolbuild 1.2, gromacs topology builder extended to include gromacs united atoms force fields
Bruce D. Ray
brucedray at yahoo.com
Tue Oct 14 14:58:57 CEST 2008
I have submitted a further revision of topolbuild as file topolbuild1_2.tgz
Please see the end of this message for an important post-submission
Reads a Tripos .mol2 file with charges to generate something approximating
gromacs .gro, .top, and .itp files from it based on selected force field parameters.
Now includes some support for united atom force fields. Includes tables to
support amber, gaff, glycam, and gmx type force fields. Requires that the
.mol2 file have syntactically correct Tripos atom types and absolutely will
not work with other input atom types. Includes capability of pruning dihedral
angles to a possibly more reasonable set.
1. Corrected improper dihedrals determination.
2. Added gromacs united atoms force fields, along with some required
atom type definition files and parameters files. Atom type definition
files and parameter files are included for gmx43a1, gmx43a2, gmx43b1,
gmx45a3, gmx53a5, and gmx53a6. Only the table for gmx53a6 has
received any checking. All of these tables are incomplete and to be
considered experimental. As with any automatic system for
parameterization, careful examination and editing of the results is
required to generate a well behaved final topology.
3. Added glycam force fields option and ATOMTYPE_GLYCAM.DEF.
The glycam option is experimental and may not give a correct result.
4. Added option to prune excess dihedral angles.
5. Added conversion of some ring planarity dihedral angles to be output
in the topology as improper dihedral angles.
6. Added an internal adjustment to correct for the problem of programs
other than Sybyl generating larger charges than those from the standard
Sybyl residue dictionary and charge generation methods.
7. Altered topology outputs to allow for various types of bond, angle, and
dihedral angle parameters.
8. Made changes to antechamber data tables to correct for problems related
to gromacs case insensitivity and for the absence of some atom type
combinations in the force field parameters.
Special Processing Used for Gromacs Force Fields
Gromacs force fields have a limited set of atom types that are used in multiple
chemical environments. This works quite well when one can use a residue types
file to generate a topology. This presents problems when the goal is to determine
the topology of an arbitrary molecule for which residue type information does not
exist. Therefore, conversion to gromacs force fields requires double determination
of atom types. A first determination of atom types is performed with a rich set
of atom types derived from a combination of the gaff and amber atom type
definitions. This assignment is used to determine force constants, bond lengths,
angles, dihedral angles, and improper angles. A second determination of atom
types is performed to assign the true gromacs atom types to match the atom
types in the appropriate ffG####nb.itp from the gromacs distribution. A major
problem of this double conversion is that for something found in a residue type
file, topolbuild does not always select the exact same parameters as are
found in the residue type file’s parameterization. However, tests so far give
the same parameterization better than 95% of the time.
Post-submission code correction: I discovered that code in the file improper.c
did not generate united atom model improper dihedral angles for tetrahedral
carbons with a single attached proton. To correct this, please replace files in
topolbuild1_2/src with the attached files.
Once again, I hope people find this useful.
Bruce D. Ray, Ph.D.
Associate Scientist, and Operations Director
402 N. Blackford St.
Indianapolis, IN 46202-3273
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 11161 bytes
Desc: not available
More information about the gromacs.org_gmx-users