Fig. 2From: Long-distance dependency combined multi-hop graph neural networks for protein–protein interactions predictionThe illustration of the LDMGNN framework. Here \(G_{1}\) indicates the original PPI network, and \(G_{2}\) indicates the THPPI network. This symbol \(\oplus\) indicates addition by element-wise. \(e_i^{T}\) and \(e_j^{T}\) represent embedding vectors for proteins \(p_i\) and \(p_j\), respectively. And the MUL block represents the dot product operation. The FC block is designed to classify the classes of protein interactions. And as shown in the figure, there are seven different typesBack to article page