Abstract
BACKGROUND: A chronic progressive degenerative joint disease, such as osteoarthritis (OA) is positively related to age. The medical economy is facing a major burden, because of the high disability rate seen in patients with OA. Therefore, to prevent and treat OA, exploring the diagnostic biomarkers of OA will be of great significance. METHODS: Differentially expressed genes (DEGs) were obtained from the Gene Expression Omnibus database using the RobustRankAggreg R package, and a protein-protein interaction network was constructed. The module was obtained from Cytoscape, and the four algorithms of degree, MNC, closeness, and MCC in CytoHubba were used to identify the hub genes. A diagnostic model was constructed using Support Vector Machines (SVM), and the ability of the model to predict was evaluated by other cohorts. RESULTS: From normal and OA samples, 136 DEGs were identified, out of which 45 were downregulated in the normal group and 91 were upregulated in the OA group. These genes were associated with the extracellular matrix-receptor interactions, the PI3K-Akt signaling pathway, and the protein digestion and absorption pathway, as per a functional enrichment analysis. Finally, we identified the 7 hub genes (COL6A3, COL1A2, COL1A1, MMP2, COL3A1, POST, and FN1). These genes have important roles and are widely involved in the immune response, apoptosis, inflammation, and bone development. These 7 genes were used to construct a diagnostic model by SVM, and it performed well in different cohorts. Additionally, we verified the methylation expression of these hub genes. CONCLUSIONS: The 7-genes signature can be used for the diagnosis of OA and can provide new ideas in the clinical decision-making for patients with OA.