Curvature-Guided Geometric Representation for Protein-Ligand Binding Affinity Prediction

In three linesRicciBind, a geometric representation framework, predicts protein-ligand binding affinity by combining Ricci curvature to capture local organization and optimal transport for cross-domain alignment. Demonstrates superior performance on PLA benchmarks and virtual screening tasks.

Summary generated by Claude — human-verified