Initial sequences generated using BindCraft. 7 diferent sequences selected. Then, binders have been optimized using Rosetta, mutating residues in the interface and second shell like this: Rosetta Protocol:

1. Load complex PDB (binder chain, target chain).
2. Identify interface residues on binder (first shell).
3. Identify second-shell residues on binder.
4. Interface design loop:
   • Design only interface residues (binder chain only; target frozen).
   • Score interface with InterfaceAnalyzerMover (ddG, etc.).
   • Accept designs that improve ddG.
5. Second-shell design loop:
   • Keep interface residue identities fixed (frozen).
   • Design only second-shell residues (binder chain only; target frozen).
   • Same scoring/acceptance as above.