Prediction method and prediction system for G-protein-coupled receptor-ligand interaction relationship

The invention discloses a multi-information integration-based prediction method for G-protein-coupled receptor (GPCR)-ligand interaction. The method comprises the steps of constructing a GPCR-ligand interaction network structure graph of a non orphan receptor; embedding a receptor-receptor relationship between an orphan receptor and the non orphan receptor into a receptor-ligand relationship network; and based on a distance relationship between nodes in the network, constructing a feature relationship of a receptor-ligand interaction pair, and in combination with an ensemble learning method, realizing effective prediction of a potential receptor-ligand relationship. The prediction method suitable for prediction of a GPCR-ligand interaction relationship and prediction of a potential ligandof an orphan GPCR is established and realized for bio-features of the GPCR and the ligand of the GPCR; and an experiment shows that the method is good in prediction effect and has a relatively wide application prospect. The i