I'm trying to construct a homo-dimer of a protein+ligand (so a total of two peptide chains and two ligand chains) using docking_protocol with the -docking::partners flag.
I've parameterized the ligand using molfile_to_params.py and docking_protocol runs without generating any fatal errors (although there are numerous warnings about discarding atoms, missing heavyatoms, can't find atoms, etc.).
However, when I check the output structures, I find that both peptide chains are intact but both ligands have completely come apart.
My suspicion is that when Rosetta is switching from centroid back to all-atom that somehow the ligand bonds are not being reassembled correctly.
I've attached the two ligand parameter files as well as the input file (with .txt extensions added solely to allow uploading) so that the problem might first be reproduced, then I welcome any advice the community may have.