ProLMS-GNN (Prediction of Ligand Molecule-binding Sites using Graph Neural Network) is a web server for predicting small-molecule binding using a graph neural network model. Starting from given PDB ID, UniProt ID, or structure data, ProLMS-GNN predicts ligand-binding residues and the binding ability for the ligand using simple structural features. Currently, prediction models for pyridoxal 5'-phosphate (PLP), flavin-adenine dinucleotide (FAD), and nicotinamide adenine dinucleotide (NAD) binding have been prepared in addition to a prediction model for general ligand binding residues. The model for NAD-binding prediction can also predict the binding for NAD-like molecules such as NADP and NADPH.
Predict binding to by model
Filter: binding probability >