I have a protein sequence that consists of two domains: A and B. The two domains presumably interact and there are some residue pairs that can be used as constraints. However, I don't know whether these domains interact within a single protein (A-B) or a dimer (A-B', B-A'). Given the fact that both A and B have many good templates in PDB, I thought that Rosetta comparative modeling protocol could be used to investigate these two scenarios. I would appreciate suggestions on how to set up such an analysis.