骨关节炎（OA）是一种慢性疾病，其发病率和致残率较高，其分子机制尚不清楚。本研究旨在通过生物信息学分析，确定与滑膜炎和软骨细胞凋亡相关的OA标记物。我们从Gene Expression Omnibus数据库中选择了五个基因表达谱。我们将GEO与GeneCards数据库结合起来，并进行了基因本体论和Kyoto基因和基因组百科全书的分析。然后，我们使用最小绝对收缩和选择算子（LASSO）算法来确定特征基因，并建立了预测风险评分。我们使用统一流形近似和投影（UMAP）方法来识别OA患者的亚型，而CytoHubba算法和GOSemSim R包用于筛选出核心基因。接下来，使用单样本基因集富集分析和CIBERSORTx进行免疫评估。选取了56个与OA相关的差异型基因，并通过LASSO算法确定了10个特征基因。通过UMAP将OA样本分为簇1和簇2的亚型，聚类结果表明这些簇之间的特征基因有显著差异。MYOC、CYP4B1、P2RY14、ADIPOQ、PLIN1、MFAP5和LYVE1在簇2中高表达，ANKHLRC15、CEMIP、GPR88、CSN1S1、TAC1和SPP1则在簇1中高表达。蛋白质相互作用网络分析显示MMP9、COL1A和IGF1为高节点，而差异基因影响了IL-17途径和肿瘤坏死因子途径。GOSemSim R包显示ADIPOQ、COL1A和SPP1与31个核心基因的功能密切相关。此外，还确定了mmp9和Fos与多个转录因子存在相互作用关系，ssGSEA和CIBERSORTx算法揭示了两个OA亚型之间免疫浸润的显著差异。最后，进行了一项qPCR实验，探索大鼠软骨和滑膜组织中的重要基因；qPCR结果显示COL1A和IL-17A在OA大鼠的滑膜组织和软骨组织中均高表达，与预测结果一致。未来，可能会找到用于同时缓解OA两种表型的常见治疗靶点。 版权所有 © 2023 阳、余、刘和曹。

Osteoarthritis (OA) is a chronic disease with high morbidity and disability rates whose molecular mechanism remains unclear. This study sought to identify OA markers associated with synovitis and cartilage apoptosis by bioinformatics analysis.A total of five gene-expression profiles were selected from the Gene Expression Omnibus database. We combined the GEO with the GeneCards database and performed Gene Ontology and Kyoto Encyclopedia of Genes and Genome analyses; then, the least absolute shrinkage and selection operator (LASSO) algorithm was used to identify the characteristic genes, and a predictive risk score was established. We used the uniform manifold approximation and projection (UMAP) method to identify subtypes of OA patients, while the CytoHubba algorithm and GOSemSim R package were used to screen out hub genes. Next, an immunological assessment was performed using single-sample gene set enrichment analysis and CIBERSORTx.A total of 56OA-related differential genes were selected, and 10 characteristic genes were identified by the LASSO algorithm. OA samples were classified into cluster 1 and cluster 2 subtypes byUMAP, and the clustering results showed that the characteristic genes were significantly different between these groups. MYOC, CYP4B1, P2RY14, ADIPOQ, PLIN1, MFAP5, and LYVE1 were highly expressed in cluster 2, and ANKHLRC15, CEMIP, GPR88, CSN1S1, TAC1, and SPP1 were highly expressed in cluster 1. Protein-protein interaction network analysis showed that MMP9, COL1A, and IGF1 were high nodes, and the differential genes affected the IL-17 pathway and tumor necrosis factor pathway. The GOSemSim R package showed that ADIPOQ, COL1A, and SPP1 are closely related to the function of 31 hub genes. In addition, it was determined that mmp9 and Fos interact with multiple transcription factors, and the ssGSEA and CIBERSORTx algorithms revealed significant differences in immune infiltration between the two OA subtypes. Finally, a qPCR experiment was performed to explore the important genes in rat cartilage and synovium tissues; the qPCR results showed that COL1A and IL-17A were both highly expressed in synovitis tissues and cartilage tissues of OA rats, which is consistent with the predicted results.In the future, common therapeutic targets might be found forsimultaneous remissions of both phenotypes of OA.Copyright © 2023 Yang, Yu, Liu and Cao.