VMD-L Mailing List

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Sat Jul 27 2013 - 13:04:11 CDT

Hello:
I was trying to merge a dimeric protein containing a ligand and a
water molecule with an organic ligand. The resulting psf (charmm27)
shows correct partial charges and atom types (new ones) for the
ligand, while the pdb coordinates for the ligand are zero.

The first part of the combines psf reads

 REMARKS segment TWA { first NTER; last CTER; auto angles dihedrals }
 REMARKS segment TWB { first NTER; last CTER; auto angles dihedrals }
 REMARKS segment O2A { first NONE; last NONE; auto none }
 REMARKS segment O2B { first NONE; last NONE; auto none }
 REMARKS segment WA { first NONE; last NONE; auto none }
 REMARKS segment WB { first NONE; last NONE; auto none }
 REMARKS segment PQA { first ; last ; auto none }
 REMARKS defaultpatch NTER TWA:0
 REMARKS patch CTER TWA:254
 REMARKS defaultpatch NTER TWB:0
 REMARKS patch CTER TWB:254

O2A/O2B and WA/WB are the ligand and water originally present in the
protein. I do not understand why the ligand PQA, that is intended to
merge, is read {first; last; auto none}. In its psf, the ligand shows:

 REMARKS segment PQA{ first NONE; last NONE; auto none }

I suppose I am missing something, however unable to remedy.

Incidentally, I am at trying the combine route because, very
unusually, I found problems in generating psf/pdb for the whole
complex (vmd -dispdev ...). That is, psfgen adds atoms to the ligands
(which bears carboxylate functional groups) assuming that it deals of
an amino acid, while it is not ;

ATOM 8189 C PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA C
ATOM 8190 OT1 PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA O
ATOM 8191 OT2 PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA O
ATOM 8192 N PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA N
ATOM 8193 HT1 PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA H
ATOM 8194 HT2 PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA H
ATOM 8195 HT3 PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA H
ATOM 8196 CA PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA C
ATOM 8197 HA PXX A 500 0.000 0.000 0.000 -1.00 0.00 PQA H

I tried variants, with atom names that do not exist in charmm27, for
example changing "N" that vmd takes for amnino and triprotonates it.
The ligand turns out to be structurally correct, at the correct place
in the protein, but the psf is faulty, with partial charges (taken
from charmm27 according to the vmd arbitrary choice of atom types)
added for atoms with coordinate zero

Generating psf/pdb for the ligand alone has no problems: correct
charges and global charge, i.e., the ff with the new atom types is OK.

thanks for advice

francesco pietra