Prediction of Enzyme Species by Graph Neural Network

Choosing an effective classification and recognition method in a large protein database plays a crucial role in the classification of enzymes. In previous studies on enzyme classification, only node characteristic information of amino acid were generally considered in the process of model training. The characteristics of amino acid nodes and topological structure in enzyme protein structure are proposed in this paper. The model was trained by graph neural network. By comparing with K nearest neighbor, support vector machine, random forest and multi-layer perceptron, it is shown that the graph neural network method has great advantages. The accuracy obtained by graph neural network is obviously higher than others.