Used the newly released Protein-Hunter Boltz-2 prediction pipeline with the configuration below. I used PeSTo to identify putative protein binding sites on the provided structure, then fed those constraints into Protein-Hunter. I modified the code to skip the Alphafold3 refolding and validation, choosing to use the original Boltz-2 binder-target predictions for Rosetta scoring, followed by re-prediction of the binders that passed the default filters to calculate the ipSAE scores.