From: Jim Phillips (
Date: Fri Jun 24 2005 - 15:51:37 CDT


I can't diagnose this without the actual vmd commands you're running. I'm
pretty sure that the psfgen coordinate-guessing code doesn't delete atoms.

You need to use psfgen commands to create a full psf file in VMD. Using
the VMD file writers produces a limited psf file, as you have seen.


On Mon, 13 Jun 2005 wrote:

> i have a protein crystal which I constructed from individual pdb protein.
> The file i'm analyzing contains water molecules from the original pdb
> file, which i would like to keep. however, when i run the structure
> through psfgen, it will guess some coordinates of hydrogens and delete any
> water oxygen it can't measure a coordinate for. only a quarter of the
> water molecules or less remain after the hydrogens are added. i can add
> hydrogens to it with other programs, but then can not produce a proper psf
> file.
> i added hydrogens with chem 3d. i then took the pdb file, loaded it into
> vmd, selected all, and wrote a psf file in tckon. however the psf file
> fails to show which atoms are bonded to each other.
> any suggestions on how to make a psf file directly from a pdb file, or
> another way to add hydrogen atoms without losing water molecules would be
> appreciated.
> Ken jordan