From: Paul O'Leary (paul.oleary_at_monash.edu)
Date: Thu Aug 20 2020 - 04:19:38 CDT

Hi all,
We are interested in carrying out simulations using peptides that contain
non-natural amino acids. As an example of a non-natural amino acid, we are
looking at ornithine, which has been included as an additional residue in a
VMD tutorial and in Swiss Side Chain Charmm22 topology and parameter files (
https://www.swisssidechain.ch/MD/charmm.php). We followed the Swiss Model
instructions and included their provided parameters into the original
par_all22_prot.prm file.
When we run autopsf with the swissmodel toplogy file (
top_all22_prot_swisssidechain.inp) and the modified prm file, the pdb
plugin recognises the peptide correctly (ie correct# of atoms, bonds and
residues). However, it seems at the topotools step (I have tried both Linux
topotools 1.8 and Windows 1.7) it splits the peptide into segments
containing the natural amino acid sequence and the non-natural sequence. In
further steps it appears to create three segments comprising of n_termini
sequence, non natural amino acid and c-termini sequence. When we go back to
the segments identified in the Autopsf dialog box and set a single segment
to encompass the entire peptide it produces a psf with the non-natural
amino acid as the new C-terminal residue, whilst spacially located
reasonably.

Any help will be greatly appreciated including direction to further
resources.

We plan to also investigate setting the non-natural amino acid to HETATM
and develop a patch to see if this circumvents these problems.

Best regards,
Paul