From: Peter Freddolino (
Date: Tue Jun 15 2010 - 20:33:02 CDT

Hi Kelly,
You've found a case that upsets the logic that autopsf uses to split
segments -- it is incorrectly identifying both CYCs as being part of the
same segment, and because you can only have one residue with a given
segment name /residue ID pair, and they have the same resid, one of them
is being dropped. The short term solution is to manually edit the
chains, deleting the automatically generated O2 (which contains both CYC
residues) and instead adding one chain for each CYC residue. This can be
done easily in the GUI when you reach the "Segments identified" step.

I have committed a fix to cvs that will prevent this problem in future
versions. However, in your case, you're going to need to manually edit
the segment definitions anyway because autopsf flags the MEN residues as
being non-protein, and thus breaks protein chain B into two separate
chains separated by a single isolated residue, which is not what you
want. This is a good example of where autopsf guesses right for common
cases, but anything unusual like nonstandard residues can cause
problems. The autopsf GUI should again be usable to fix this problem.

Thanks for bringing this to our attention.


On 06/15/2010 11:35 AM, Kelly Leung wrote:
> Hello,
> I recently ran AutoPSF for phycocyanin (2VJT on the RCSB protein data
> bank), complete with the toppar files for the residue MEN (n-methyl
> asparagine and CYC (phycocyanobilin).
> The output pdb file is missing one of the CYCs (The file directly
> downloaded from RCSB has two CYCs, the one resulting from AutoPSF only
> has one) and I can't determine why. Could it be a matter of chirality?
> --
> Kelly Wai-Wah Leung
> Energy, Environmental, & Chemical Engineering
> Washington University in St. Louis '12
> Campus Box 3787, 6515 Wydown Blvd, St. Louis, MO 63105.
> 847.840.3412 |
> <>