On improving experimental binding affinity predictions with synthetic data
Published in BioRxiv, 2026
We extend SAIR with ~80K absolute free energy perturbation calculations and show that carefully curated synthetic structures improve binding affinity prediction models.
Recommended citation: Ryczko, Kevin; Zin, Phyo Phyo; Crivelli-Decker, Jordan; Le, Ly; Jha, Punit K.; Shields, Benjamin J.; Lemos, Pablo; Bandi, Sasaank; van Damme, Maarten; Sood, Amogh; Huntington, Lee; Pitman, Mary; Ganahl, Martin; Bortolato, Andrea. "On improving experimental binding affinity predictions with synthetic data", BioRxiv, 2026. https://doi.org/10.64898/2026.03.02.708607
