Curated Antibody Binding Affinity Datasets (AbCDR-Binding)

This dataset contains six curated antibody-antigen binding affinity datasets, as used in the paper Preferential CDR masking in paired antibody language models improves binding affinity prediction". Contents: Six binding affinity datasets covering diverse antigen systems: D44 dataset (N=2,048): Single-point mutants of anti-lysozyme antibody D44.1 G6 dataset (N=4,275): Single-point mutants of anti-lysozyme antibody HyHEL-10 Trastuzumab dataset (N=422): Single-point mutants of therapeutic anti-HER2 antibody anti-Fluorescein dataset (N=11,052): Single and combinatorial mutants binding to fluorescein anti-H1 Hemagglutinin dataset (N=1,038): Single and combinatorial mutants targeting influenza H1 anti-HR2 SARS-CoV-2 dataset (N=71,830): Combinatorial mutants targeting SARS-CoV-2 spike protein Data sources: D44, G6: Koenig et al. (2017), Warszawski et al. (2019) Trastuzumab: Shanehsazzadeh et al. (2023) anti-Fluorescein: Adams et al. (2016) anti-H1 Hemagglutinin: Petersen et al. (2024) anti-HR2 SARS-CoV-2: Lim et al. (2023) Citation If you use this dataset, please cite both the original OAS publication and our published article: Talaei, M., Walker, K. C., Hao, B. et al. (2026). Preferential CDR masking in paired antibody language models improves binding affinity prediction. Communications AI & Computing, 1, 7. https://doi.org/10.1038/s44488-026-00010-2