I’m not entirely sure, but I believe that they used the standard protein-protein docking protocol, but instead of two full-sized proteins they used one full-sized protein and one single residue “protein”. I haven’t tried it myself, but from what I understand, that should work*.
If you need help getting started with protein-protein docking, please see the Rosetta tutorials (https://www.rosettacommons.org/demos/latest/Home#tutorials) or ask additional questions here, if something isn’t clear.
*) You may have issues with the docking protocol forming interactions with the backbone of the single residue partner. But you should be able to manually filter out those docked conformations as a post-processing step.
.
