Hi Maryam,
First, what protocol to use depends on what you intend to accomplish. What part of the ab-ag are you interested in? Is your starting structure a crystal structur or a homology model? How well do you know the structure of loops?
Snugdock is more about refining structures (CDRs, L-H, and LH-A interface) than predicting the interface of an ab-ag complex.
Yea, it looks like that documentation is not quite there, but definitely checkout the paper for more. I’ve used snugdock in the past. Its super simple to use, though I’ve never done a production run using it. Checkout the paper for what they used in terms of numbers of output decoys, but you should be able to use it like any other Rosetta protocol. Pass -s for your antibody and your structure needs to have L, H, and A(ntigen) chains, in that order (which PyIgClassify will do for you).
You will need to have your PDB renumbered. I just tested locally with the AHO_Scheme (renumbered via PyIgClassify), and looks to work with alternative numbering schemes. Looking at the code, it should work with other CDR definitions too. See https://www.rosettacommons.org/docs/latest/application_documentation/antibody/General-Antibody-Options-and-Tips for more.
My simple command line was snugdock.macosclangrelease -s local_renum_ab.pdb -numbering_scheme AHO_Scheme -ignore_unrecognized_res
.
