DDG Predictor is the deep-learning model from "Deep Learning-Guided Optimization of Human Antibody Against SARS-CoV-2 Variants with Broad Neutralization" (Helixon Protein). Given a wild-type complex PDB and a mutant complex PDB (same length, mutated structures typically built by Rosetta), it predicts the change in protein-protein binding free energy (ddG) for the mutation as a single scalar.