From: Rajan Vatassery (rajan_at_umn.edu)
Date: Mon Feb 06 2012 - 11:02:14 CST

Flavio, Chris and Axel,
        Thank you for your input. I ended up making a PSF containing the "!
NATOM" list and the "!NBOND" list, and using topotools (v 1.1) I
confirmed the number of bonds (52 -- from the partial PSF I loaded along
with the pdb), angles (0), and dihedrals (0) using:

vmd > topo numbonds
52
vmd > topo numangles
0
vmd > topo numdihedrals
0

Then for the actual angle and dihedral creation, I used

vmd > topo guessangles
vmd > topo numangles
99
vmd > topo guessdihedrals
vmd > topo numdihedrals
139

Then write the psf using

vmd > animate write psf test2.psf
psfplugin) Writing angles/dihedrals/impropers...
Info) Opened coordinate file test2.psf for writing.
Info) Finished with coordinate file test2.psf.
1

This process is disconcertingly easy, so I went through a majority of
the angles and dihedrals by hand and confirmed that they are correct and
no angle or dihedral is duplicated or missing.

        I consider this solved on the basis of the topotools package. Many
thanks to Axel and anyone else developing that toolbox.

rajan

On Sun, 2012-02-05 at 14:13 -0600, Rajan Vatassery wrote:
> Hi All,
> I am looking for a way to avoid creating a topology file whenever I
> need to create a new PSF. For example, my current need is for a PSF for
> oleic acid. I can easily create a pdb file, but the creation of a PSF is
> substantially more time-intensive. Currently I would need to make a
> topology file, which although is not terribly difficult, is not used
> after the creation of the PSF.
> My hope is that there is some way that I can use the bonds given in the
> pdb file to produce the essential details that PSFGEN needs to then go
> on to writing the angles and dihedrals. As it is now, if I try to create
> anything in terms of a segment for writing (segment U {pdb Oleate.pdb}),
> I get the usual:
>
> unknown residue type
> ERROR: failed on end of segment
> MOLECULE DESTROYED BY FATAL ERROR! Use resetpsf to start over.
>
> because I haven't given a topology file to PSFGEN.
> My impression is that during the normal PSF creation from a topology
> file, PSFGEN calculates the angles and dihedrals from only the list of
> bonds that are given in the topology file. So my question stated in
> another way is, can I trick PSFGEN into reading the bond list from the
> PDB file? I would even rather write the first part of the PSF (atoms and
> bonds) if I could get the angles and dihedrals automatically generated.
>
> Thanks in advance,
>
> rajan
>