2.4. Comprehensive Analysis of PPI Networks and Modules

To analyze the interactions between the proteins encoded by DEGs in the DG and ACC, the STRING (https://string-db.org/) database was used to predict the interactions between the proteins encoded by DEGs [35]. The active interaction sources included Textmining, Experiments, Databases, Coexpression, Neighborhood, Gene Fusion, and Cooccurrence, with a confidence score > 0.4 being defined as significant. Then, the PPI network was constructed by Cytoscape software (version 3.5.1) [36]. The topological properties of the PPI network, including the node degree and betweenness centrality, were determined. The significance of a gene in the network was evaluated by measuring its “betweenness centrality” and “degree,” and the gene nodes with the top scores for both betweenness centrality and degree were identified as hub genes. Furthermore, module analysis was performed by using the PEWCC plugin (version 1.0) to explore the clustering modules, which included the hub genes in the vast PPI network (join parameter > 0.5, overlap threshold > 0.8) [37]. Finally, the ClueGO plugin (version 2.5.6) was used to analyze the biological functions of the clustering modules, and P < 0.05 was considered statistically significant [32, 38].