Enhancing Ligand Binding Affinity Prediction in Drug Discovery through Pairwise Comparison Network

Researchers have developed a new AI model called the pairwise binding comparison network (PBCNet) that can accurately predict the relative binding affinity of ligands in lead optimization for drug discovery. PBCNet outperformed other high-throughput methods, except for FEP+, and showed robustness and stability in its performance. The model was also able to accelerate lead optimization projects by up to 473% and reduce resource investment by 30%. PBCNet's efficiency and accuracy make it a valuable tool for guiding lead optimization in drug discovery.