VMD-L Mailing List

From: Pawel Kedzierski (pawel.kedzierski_at_pwr.edu.pl)
Date: Fri Feb 02 2018 - 05:45:47 CST

Hi Bryan,
W dniu 02.02.2018 o 11:09, Bryan Roessler pisze:
> Hello,
>
> I have a short engineered protein fragment that is correctly ordered
> and numbered with resids and nums (701, 702, 703, 704, 705) in the
> original pdb, however when I run autopsf on the original pdb the
> resulting *_autopsf_formatted.pdb places the residues in the wrong
> order (702, 703, 704, 701, 705). This results in peptide bonds being
> generated between non-consecutive residues. Thus the carboxyl in
> residue 701 ends up bound to the amino in residue 705 even though they
> are at opposite ends of the peptide. I have tried renumbering the
> residues from 1 to no avail. The only oddity in my peptide is that
> residues 701 and 705 are non-standard and use custom topologies, with
> the correct DECL and BOND +N for residue 701 and -C for residue 705.
Autopsf divides your system into segments, with every protein chain or
nucleic acid strand ending up in separate segments. I guess that your
nonstandard residues make up another segment and this results in reordering.
I an not sure if this is done by residue names or by checking if an atom
belongs to ATOM or HETATM records.

You may try a few things:

  * if you use the Automatic PSF builder GUI, at Step 3 you can delete
    the wrong segment with 701 and 705 and edit the first one to include
    your entire peptide.
  * alternatively, try to change the first 6 characters of the ATOM and
    HETATM lines in your PDB file to be all ATOM or all HETATM.
  * if neither of the above works, you could try to fool autopsf by
    renaming residues, but this would need to be done both in PDB and
    topology files. There should be an easier way, though.

HTH,

Pawel

>
> Thanks,
> Bryan
>
> *Bryan Roessler | Researcher
> *
> UAB | The University of Alabama at Birmingham
> *_uab.edu/cmdb <http://uab.edu/cmdb>_*
> Knowledge that will change your world
>