Figure 2.

The performances of the representative area-based models implemented in the web serverAREA-AFFINITYand some of the other methods. A. Sixty area-based models, PRODIGY and LISA for protein–protein binding affinity prediction. The data is based on refs (18 and 20). B. Thirty-seven area-based models and CSM-AB for antibody–protein antigen binding affinity prediction. The data is based on our earlier study.¹⁹