Hello,
Firstly, it is not compulsory to include a native file to run docking. You could choose not to use a native file; it does not affect the docking algorithm in any way.
A native file is used when we want to compare the RMSD of the models to a known structure, eg. if you are trying to benchmark a protocol or if you have a homologous complex. If you have a homologous complex, I would individually align A and B to their respective homologs in the complex, say A’ and B’ and then save the PDB with that orientation of A-B. You could then use this PDB as your native file.
If using a native file, make sure:
-
The order of the chains is the same as in the input, and
-
The same number of residues are present in the input and native.
Best,
Shourya