[gmx-users] Re: converting pdb to gmx successfully

[David Mathog]

Well I'm getting closer to "automatic" pdb->gmx but it's still
more of a slog than it should be.

1.  Run raw.pdb file through "reduce -BUILD" -> input.pdb
    (This puts on all required protons except, alas, for the ones
    at the N termini).
2.  Run raw.pdb through "whatif".  Copy the 3 lines for the 3 N
    terminal hydrogens into the first Amino acid residue (per chain)
    in input.pdb.
3.  Use modified Amber ff (O -> O2 for all THY residues) and add
    that to the first position in FF.dat.
4.  Run  pdb2gmx  (my own patched up version that automagically
    changes residue and atom names for Amber, to be posted for your
    use soon).

The pdb2gmx runs successfully.  The amount of manual cut and pasting
has been reduced from a ton to just three lines. 

Unfortunately pdb2gmx won't by itself put on all the hydrogens
Amber wants so specifying -ignh or feeding in raw.pdb directly doesn't work.

The carboxyl end is wrong of course, since even if there's an OXT 
in the PDB file nothing useful shows up in the output gro file.  In
most cases it shouldn't matter much.

The nucleic acid 5' end must be a OH, since the 5' models in the Amber
ff set are all like that.  

The ASP/ASPH/GLU/GLUH interactive stuff works and properly ends up
as the Amber ASP/ASH/GLU/GLH.

Ideally it would be nice if pdb2gmx could generate the N and C termini
all by itself.  It looks like "all" that's needed for that are -n.tdb
and -c.tdb files.  However I don't know yet how to write those.  Does
anybody have these already for the Amber FF files?

Thanks,

David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech