A novel method for drug-target interaction prediction based on graph transformers model

Abstract Background:drug-target interactions prediction(DTIs) becomes more and moreimportant for accelerating drug research and drug repositioning. drug-targetinteraction network is a typical model for DTIs prediction. As many differenttypes of relationships exist between drug and target, drug-target interactionnetwork can be used for modeling drug-target interaction relationship. Recentworks on drug-target interaction network are mostly concentrate on drug node ortarget node and neglecting the relationships between drug-target. Results:We propose a novel prediction method for modeling the relationshipbetween drug and target independently. Firstly, we use different level relationshipsof drugs and targets to construct feature of drug-target interaction. Then, we useline graph to model drug-target interaction. After that, we introduce graphtransformer network to predict drug-target interaction. Conclusions:We introduce line graph to model the relationship between drug andtarget. After transformed drug-target interaction from links into nodes, we usegraph transformer network to fulfill drug-target interaction prediction task. Keywords: drug-target interaction; graph attention network; line graph