VMD-L Mailing List

From: Josh Vermaas (joshua.vermaas_at_gmail.com)
Date: Thu Apr 09 2020 - 17:04:14 CDT

Hi Hendrik,

I missed a "guesscoords" command initially (put it right after regenerating
angles and dihedrals). Since the names for the hydrogens in the NMR
structure and the CHARMM force field are not identical, some elements of
the structure look a little freaky, since the default is for psfgen to put
everything at the origin. guesscoords guesses based on connectivity where
the missing atoms should go.

Now for the fun part: you've got a patch that works, but isn't ready for
simulation quite yet. To demonstrate this, try the following in the
tkconsole.

set sel [atomselect top "all"]
vecsum [$sel get charge]

You will see that you have a non-integer charge. That is the other part
that goes in the patch, changing atomtypes and charges so that the physics
that you are modeling is correct. What I'd do in your shoes is to change
the charges and atomtypes around the new linkage so that they match a usual
peptide bond (see the DISU patch for an example of how CHARMM expects the
format of changed atom types and charges). For simulation, note that you'll
*still* need to figure out what parameters you are missing (I just let NAMD
tell me one at a time). Alot of them will be pretty easy to do by analogy,
since CHARMM has standard-ish parameters for alot of interactions, but that
will also be a decision you'll need to make.

-Josh

On Thu, Apr 9, 2020 at 3:40 PM Hendrik Schröder <
hendrik.v.schroeder_at_gmail.com> wrote:

> Hi Josh
>
>
>
> Sorry to bother you again, but your recommendations have been very helpful
> thus far! I try to create a psf for a peptide (PDB: 1pp5) with linkage
> between N-terminus and a side chain (isopeptide bond).
>
> I followed your idea of creating an own patch and tried to copy a patch
> for cyclic peptides (LIG1) with taking the removal of atom OE2 into
> account. Psfgen works just fine with the topology file but during
> inspection of the pdb/psf, the connectivity of some atoms is messed up. I
> feel that something is wrong with the IC values in the patch. Enclosed, you
> find patch and files.
>
> Any help is very much appreciated! :)
>
>
>
> Best wishes,
>
> Hendrik
>
>
>
> My patch for isopeptide bonds:
>
>
>
> PRES IPB 0.00 ! patch for isopeptide bond between N-terminus
> which is a Gly and Glu
>
> ! use in a patch statement
>
> ! generation using first NONE
>
> ! follow with AUTOgenerate ANGLes DIHEdrals command
>
> GROUP
>
> ! OE1
>
> ! //
>
> ! CG--CD--N--
>
> ! /
>
> ! HN
>
> DELETE ATOM 1OE2
>
> BOND 1CD 2N
>
> IMPR 2N 1CD 2CA 2HN 1CD 1CG 2N 1OE1
>
> IC 1N 1CG 1CD 2N 0.0000 0.0000 180.0000 0.0000 0.0000
>
> IC 2N 1CG *1CD 1OE1 0.0000 0.0000 180.0000 0.0000 0.0000
>
> IC 1CG 1CD 2N 2CA 0.0000 0.0000 180.0000 0.0000 0.0000
>
> IC 1CD 2N 2CA 2C 0.0000 0.0000 180.0000 0.0000 0.0000
>
> IC 1CD 2CA *2N 2HN 0.0000 0.0000 180.0000 0.0000 0.0000
>
>>