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, and COL1A2) 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.

中文翻译：

---

基于多芯片集成分析和机器学习的骨关节炎诊断生物标志物识别

骨关节炎（OA）的发病机制仍不清楚。因此，重要的是要鉴定OA的相关诊断标记基因。我们对多个微阵列数据队列进行了综合分析，以确定OA发育的潜在转录组标记。此外，为了鉴定OA诊断标记，我们基于涉及这些差异表达基因（DEG）的蛋白质-蛋白质相互作用网络建立了基因调控网络。利用支持向量机（SVM）模式识别，建立了OA预测和预防的诊断模型。京都基因与基因组百科全书通路分析表明，190个DEG主要集中在肿瘤坏死因子信号通路，白介素17信号通路，促分裂原激活的蛋白激酶信号通路，核因子κ-轻链增强剂激活的B细胞信号通路和破骨细胞分化。八个枢纽基因（POSTN，MMP2，CTSG，ELANE，COL3A1，MPO，COL1A1和COL1A2）被认为是OA的潜在生物标志物，曲线下面积（AUC）> 0.95，显示出很高的准确性。在多个外部验证队列中，基于这八个基因的OA SVM模型的敏感性和特异性达到100％。我们的研究为临床医生诊断OA提供了理论依据。