Structural Hole Spanner in HumanNet Identifies Disease Gene and Drug targets

Traditional methods of analyzing genes mainly focus on a single gene. However, they are time consuming, costly, and neglect the interaction between genes. To overcome the shortcomings of traditional methods, many computational techniques have been proposed for gene analysis. Computational methods posit that the interaction between genes forms a network. By analyzing the topological characteristics of the network, a type of node, “hubs,”is identified. Only a small fraction of nodes is called “hubs,”but they interact with many partners in the network. Several methods are used to find “hubs”in a network. However, they are not comprehensive. Nevertheless, searching for more information about genes in a network remains challenging. In this paper, we integrate HumanNet and a framework to find a structural hole (SH) spanner. We classify genes into three classes (SH, normal, and non-SH). We also classify genes into three classes according to degree and betweenness. Enrichment analysis reveals that the SH genes are enriched for an essential one. Furthermore, enrichment analyses of disease mutations, viruses, and drugs consistently show that the SH genes are preferred targets and are important in cellular state transition. Functional enrichment shows that the biological function that the SH genes over represent is related to cell function. We demonstrate that by finding the structural hole spanner in HumanNet, we can find the SH genes that are essential and indispensable in cellular life. In addition, we identify the SH genes that are key players in cellular state transition.