Structure-Guided Adaptive Propagation for Protein-Protein Interaction Site Prediction
Researchers have developed a new deep learning model called SGAP-PPIS for predicting protein-protein interaction sites. This model utilizes an adaptive propagation mechanism that adjusts information diffusion based on the geometric environment of each residue. By incorporating multi-scale geometric states from an equivariant graph neural network, SGAP-PPIS can better distinguish true interaction sites from similar non-interacting neighbors. Experiments show that SGAP-PPIS achieves competitive performance with existing state-of-the-art methods. AI
IMPACT This model could improve the accuracy of predicting protein interactions, aiding drug discovery and understanding cellular processes.