In Annals of translational medicine
Background : At present, the progression mechanism of knee osteoarthritis (KOA) has not been fully elucidated, and there is a clinical need for late KOA-specific diagnostic markers to provide reference for preventive treatment. This study aimed to analyze the sequencing results of early- and late-stage KOA synovial tissue based on the key genes of late-stage KOA in combination with a machine learning algorithm.
Methods : The whole transcriptome sequencing results of synovial tissue from KOA patients (GSE176223 and GSE32317) were downloaded from the gene expression omnibus (GEO) database. Thirty-nine early KOA synovial tissue samples and 31 late KOA synovial tissue samples were included in this study. The diagnostic criteria and baseline data balance of early and late KOA were referred to the data source literature, and the two groups of data had good baseline data balance. R software (V3.5.1) and R packages were used for screening and enrichment analysis of differentially expressed genes (DEGs). The key genes were screened by weighted correlation network analysis (WGCNA) and least absolute shrinkage and selection operator (LASSO) regression analysis. A receiver operating characteristic curve (ROC) curve was used to evaluate the diagnostic efficacy of key genes for advanced KOA.
Results : A total of 211 DEGs related to knee arthritis were screened out. Compared with synovial tissue of early knee arthritis, 111 genes were upregulated and 100 genes were downregulated in the synovial tissue of late knee arthritis. Sixty-six key genes were screened out through WGCNA and 34 key genes were screened out in the LASSO analysis. The genes obtained by the two algorithms combined with three overlapping genes, namely interleukin- 6 (IL-6), C-X-C chemokine ligand 12 (CXCL12), and macrophage migration inhibitor factor (MIF). The areas under the ROC curves of CXCL12, IL-6, and MIF were 0.96, 0.944, and 0.961, respectively (P<0.001).
Conclusions : IL-6, CXCL12, and MIF are the key pathogenic genes of KOA, which have good diagnostic efficacy for advanced KOA.
Yang Yongju, Zhang Yuqian, Min Dongyu, Yu Heshan, Guan Xuefeng
Knee osteoarthritis (KOA), diagnostic efficiency, differentially expressed genes (DEGs), machine learning