I am currently embarking on the same task. I think what batch2k meant was defining specific disulfides in a protein, aka AA 25 to AA 45. I am trying to use pose.conformation().bond_length() as a way to tell if the correct disulfides have been created. If the wrong one(s) were created during a perturbation step, the pose will revert back to its original form. And eventually, the correct disulfides should emerge. We shall see.