[gmx-users] crosslinking polypeptide chains

[Mark Abraham]

Hi,

Please start new topics in new emails, rather than confusing the web
archives with replies to digests :-)

On Tue, Feb 16, 2016 at 5:00 PM Nash, Anthony <a.nash at ucl.ac.uk> wrote:

> Hi all,
>
> When executing pdb2gmx I am getting a fatal error due to dangling bonds. I
> know that it will be down to how I¹ve organised the .pdb file, I¹m just
> lacking in the experience with TERs, -chainsep and -merge to solve this. I
> would appreciate hints/tips/outright-solutions.
>
> My protein is very simple. It is a proof of concept for forcefield
> parameters I derived myself from QM data (in house software I've
> developed). There are two chains, both are the triplet NGLY-MOD-CGLY. The
> MOD residue is actually a lysine covalently bound to a carbon ring which
> is then covalently bound to the neighbouring chain¹s lysine. It is being
> treaded as one complete residue, MOD. E.g.,
>
> NGLY-MOD-CGLY
>
>       |
> NGLY-MOD-CGLY
>
>
> Just to reiterate, there is only one instance of the residue MOD, the
> instance has two defined backbones (I.e., backbone as part of each chain).


It seems like arranging for your .pdb file to have chain separators that
make gmx pdb2gmx -chainsep whatever -merge interactive would work smoothly,
and then have specbond mechanism do the crosslink. What can't work is a
single MOD residue, because the early stages of pdb2gmx only know how to
recognize (and handle termini for) an unbranched chain. You might need a
(complete) LYS+ring residue in one chain, and a normal LYS in the other.

Mark