Identification Of Diagnostic Biomarkers Of Osteoarthritis Based On Multi-Chip Integrated Analysis And Machine Learning

The pathogenesis of osteoarthritis (OA) is still unclear. It is therefore important to identify relevant diagnostic marker genes for OA. We performed an integrated analysis with multiple microarray data cohorts to identify potential transcriptome markers of OA development. Further, to identify OA diagnostic markers, we established gene regulatory networks based on the protein-protein interaction network involved in these differentially expressed genes (DEGs). Using support vector machine (SVM) pattern recognition, a diagnostic model for OA prediction and prevention was established. Kyoto Encyclopedia of Genes and Genomes pathway analysis revealed that 190 DEGs were mainly enriched in pathways like the tumor necrosis factor signaling pathway, interleukin-17 signaling pathway, mitogen-activated protein kinase signaling pathway, nuclear factor kappa-light-chain-enhancer of activated B cells signaling pathway, and osteoclast differentiation. Eight hub genes (POSTN,MMP2,CTSG,ELANE,COL3A1,MPO,COL1A1, andCOL1A2) were considered potential diagnostic biomarkers for OA, the area under curve (AUC) was >0.95, which showed high accuracy. The sensitivity and specificity of the SVM model of OA based on these eight genes reached 100% in multiple external verification cohorts. Our research provides a theoretical basis for OA diagnosis for clinicians.