Pretty much all of the protein protein docking in current versions of Rosetta uses RosettaScripts. I’ve attached an example script. The flexible backbone movement in this protocol is all after-the-fact, though. A docked conformation is found, and the backbone is minimized to be “optimal” for that rigid body orientation. You then count on the multiple output structures from Rosetta to allow backbone sampling in different rigid body orientations.
If you want to do more extensive sampling, for example active loop remodeling, that’s a little harder to do with RosettaScripts (we don’t yet have a canned loop remodeling protocol exposed on the RosettaScripts level). But you can do smaller active backbone sampling in RosettaScripts (e.g. the “backrub”, “small” and “shear” moves), and you can move output through multiple programs. (E.g. run your rough docking in RosettaScripts, use the output to do active loop remodeling in another Rosetta program, and then bring the output of that run back into RosettaScripts to finish it off.)
The exact details will depend on what exactly you’d like to do. RosettaScripts is rather flexible, assuming it has all the parts which you want to use.
.
