Are you looking for automation for prediction of a large number of structures, or do you have just a few structures but are wary about the amount of work involved?
If it’s the latter, I’d recommend you reconsider. It’s not actually all that involved, it’s just a fair number of moving parts. If you haven’t looked at the tutorial at Rosetta/demos/tutorials/rosetta_cm, I’d recommend doing so.
The reason I encourage you to be a little bit more “hands-on” in your comparative modeling is that the success of a comparative modeling run depends critically on the quality of your alignments. If you have poor alignments, you get poor results. The automated protocols on Robetta do a decent job, but they don’t necessarily have the trained eye you have for your particular system of interest. You’ll likely get better results if you run through the protocol manually, spending the time to examine the output results, adjusting the sequence alignment, adding/trimming the template structures carefully, and double checking the threading.
I’d only recommend you try to automate the process if you’re running a pipeline where you’re predicting the structures for a large number of unrelated structures. Unfortunatley, the software behind Robetta isn’t externally availible (nor am I aware of another off-the-shelf RosettaCM-based prediction pipeline), though if you understand the RosettaCM process, it should be realatively straightforward to automate most of the steps. (Except the critically important manual inspection phases.)
.
