A New Disease Candidate Gene Prioritization Method Using Graph Convolutional Networks

Abstract

Identifying disease genes from a large number of candidate genes by laboratory methods is very costly and time consuming, so it is necessary to prioritize disease candidate genes before laboratory work. Recently, many gene prioritization methods have been proposed using various datasets such as gene ontology and protein-protein interaction, which are often based on text mining, machine learning, and random walk methods. Due to the good performance and increasing use of deep graph networks in the representation of graph problems, in this study, a method based on graph convolutional networks has been developed to represent the graph on the protein-protein interaction. The results show that the proposed method is effective and the performance of the proposed method better than other methods in some cases.