Quick question: It is impossible with the normal Relax application to relax both the peptide chain and ligand(s), right?
Every time I have done a relax with a ligand, the ligand has not “wobbled” at all. So I currently relaxing the apostructure, add the ligand and use ligand_dock recursively for the ligands. (I know I ought to be using a tidy Rosetta script as opposed to an overly byzantine bash script that relies on a depracated application)