Fig. 1From: Node-adaptive graph Transformer with structural encoding for accurate and robust lncRNA-disease association predictionThe NAGTLDA workflow. Step1: Construct the integrated similarity network, extract the local features of the heterogeneous network and the integrated similarity network adopting NAFS, and encode the structural information of the integrated similarity network applying SDNE. Step2: Learn global information of heterogeneous network nodes by Transformer architecture. Step3: Adaptively fusing local information of nodes, global information and structural coding of the network by Transformer architecture. Step4: Predict associations using bilinear encoderBack to article page