Is it possible to define improper dihedral angles for sidechain-backbone crosslinks


I’m trying to simulate a system with two protein chains. The C-terminus of the first protein is covalently linked to a Cysteine sidechain on the second protein, forming a thioester bond.

I have created the required forcefield files and parameters and can create the system topology with pdb2gmx. However I’m usure how to define the improper dihedral angle of the thioester.

Because the connection is sidechain->mainchain i can not simply write: [C CA +SG O].

Gromacs will accept this definition, and pdb2gmx will even create a topology, but thee topology will not include an entry for the improper.

One solution would be to modify the topology file and add the missing improper. However, if possible I would prefer a more general solution.

If anyone know a solution or has had a similar issue in the past, I’d love to hear how you solved it.

GROMACS version: 22.5
GROMACS modification: No