AMMP only understands its own language. You read an existing coordinate file
by compiling it into AMMP language. The program preammp does this.
Preammp only works one residue at a time. YOU WILL NEED TO SUPPLY THE
INTERMONOMER LINKAGE IN A POLYMER. Scripts such as peplink.amp (for proteins) and
ndick.amp (for standard DNA/RNA linkage) are supplied. These were generated by
generating the template for the dimer and extracting the appropriate terms. They
are also a good place to see how to manipulate atoms and potentials in an
automatic script.
Preammp requires several files:
Atoms file. This file is the definition of the various kinds of atoms which are in the
coordinate file. Preammp uses this to calculate bond and angle parameters and
set up the nonbonded potentials. A typical entry is:
c3 .757 109.47 0 0 3.851 .105 1.912 5.343 10.126 12.000 -.459 2.119 2.
atomname, halfbond, angle, abc parameter1,abc paramter2, vdw minimum location,
vdw minimum energy, pseudo charge for force constants, Jaa, Qaa (charge
terms), mass, default charge, torsion parameters
C3 is the type for an sp3 carbon. Generally atoms are named by the atom
symbol followed by a flag to describe the chemistry. C2 is an sp2 carbon. Ca is an
aromatic carbon.
Templates Preammp has to know which atoms are bonded to which. The minimal template
is simply the number of atoms, followed by the atoms, then the number of bonds
followed by the bonds. Additional terms for special angles, chiral groups, and
torsion parameters are allowed. Normally these fields are zero.
PDB file. Input coordinates in (more or less) PDB format. The new pdb format
flags for segid, atom type and 'charge' are ignored. Preammp is not fussy, as
long as the fields are separated by spaces. Unknown atoms can be given an
x,y,z value of 0.,0.,0. and then AMMP will generate them. This is useful for
adding single instances of unusual geometry such as a terminating carboxyl group.
Preammp will complain, but pass the atom through. This is also useful for
generating a polymer. If one atom in each monomer is specified with dummy
coordinates, then preammp will be able to generate the appropriate monomer terms. You
will then need to supply the linkage geometry. In thecase of a terminating
carboxyl, the file oxtlink.amp has that information. When converting from
pseudo-pdb files generated by other molecular mechanics programs, it is generally a
good idea to remove the hydrogens and rebuild them. While standard names for heavy atoms in proteins and nucleic acids are
generally followed, accord on hydrogen names is far from universal. Another
problem you may run into is the multiple naming conventions on sugars (C* .vs. C' and
O4 vs. O1 in the ring).