ipsae

You can rank protein binder designs by computing ipSAE scores on AF2, AF3, or Boltz predictions to distinguish true binders from decoys. The Skill requires aligned-error information, chain assignments, and structure data to generate chain-pair and residue-level scores for filtering top candidates.

ipSAE scoring is used to rank protein-protein interaction binder designs by evaluating predicted structures. It computes chain-pair and residue-level scores from aligned-error data to help researchers prioritize true binders over decoys, reducing experimental screening effort for synthesis or testing.

Yes, you can batch process multiple protein designs and compare ipSAE scores against ipTM or iPAE metrics. The Skill supports evaluating many AF2, AF3, or Boltz predictions simultaneously, outputting score tables you can export for downstream QC and selection to improve ranking confidence.

To compute ipSAE scores for binder ranking, you need aligned-error information, chain assignments, and structure data from your AF2, AF3, or Boltz predictions. These inputs allow the Skill to calculate chain-pair and residue-level scores that distinguish true binders from decoys across your design pipeline.

Yes, ipSAE works with AlphaFold2, AlphaFold3, and Boltz structure predictions for scoring binder designs. It processes the aligned-error information and chain assignments from these tools to generate ranking scores applicable to your protein-protein interaction design pipeline.

ipSAE provides chain-pair and residue-level scores that can be compared against ipTM or iPAE to improve ranking confidence for protein binder candidates. Using ipSAE alongside these metrics helps better distinguish true binders from decoys, enabling more effective prioritization for experimental screening.