From: Josh Vermaas (
Date: Tue Apr 07 2020 - 15:29:25 CDT

Hi Hendrik,

This isn't a beginner question unfortunately, since these structures don't
come up all the frequently. You'll need to defined a patch to make the bond
you want, and to change the atomtypes around within that patch so that the
peptide bond between residues 1 and 8 is modeled consistently with the rest
of the force field. You can *probably* do this by analogy, but that is
going to be a longer discussion. You are likely going to need to make the
psf yourself, and I'd use psfgen. The script would look something like this:

topology top_all36_prot.rtf
segment P {
pdb 1pp5.pdb
coordpdb 1pp5.pdb P
patch XXX P:1 P:8
regenerate angles dihedrals
writepsf output.psf
writepdb output.pdb

The trick is to define the patch, since I can't find one that currently
exists within the CHARMM force field that does what you want.


On Tue, Apr 7, 2020 at 12:48 PM Hendrik Schröder <> wrote:

> Hi VMD and NAMD users
> I try to build a psf file for a peptide with linkage between its
> N-terminus and a side chain (=isopeptide bond, =cyclic-branched peptide).
> Psfgen does not recognize this extra bond and creates a linear version of a
> peptide (see for example PDB: 1PP5). After reading forum threads on
> cyclic peptides and checking manual/tutorials, it seems like a very common
> problem without solution yet.
> I would be very happy if someone could point out a strategy (preferably
> for beginners) to convert a pdb of a peptide with side chain linkage into a
> psf-file. Maybe someone has a suitable topology file already and could send
> it to me.
> Thank you so much and with my best wishes,
> Hendrik