Identification of telomere-related diagnostic markers in osteoarthritis based on bioinformatics analysis and machine learning.

Abstract

Osteoarthritis (OA) is one of the most prevalent joint disorders, with aging considered a primary, irreversible factor contributing to its progression. Telomere-related cellular senescence may be a crucial factor influencing the OA process, yet biomarkers for OA based on telomere-related genes have not been clearly identified. The datasets GSE51588, GSE12021, and GSE55457 were retrieved from the Gene Expression Omnibus database. Initially, R software was utilized to identify differentially expressed genes between OA and normal samples. Subsequently, differentially expressed telomere-related genes (DETMRGs) were obtained, and their functional enrichment was analyzed. Feature genes for OA diagnosis were selected from DETMRGs using a combination of least absolute shrinkage and selection operator, support vector machine-recursive feature elimination, and Random Forest algorithms. The diagnostic value of these feature genes was then validated through receiver operating characteristic (ROC) curves and decision curve analysis. Additionally, CIBERSORT and xCell were employed to assess the infiltration of immune cells in OA tissues. Finally, potential drugs targeting candidate genes were predicted. Three telomere-related genes, PGD, SLC7A5, and TKT, have been identified as biomarkers for OA diagnosis and were confirmed through ROC diagnostic tests. The immune infiltration of mast cells, neutrophils, common lymphoid precursors, and eosinophils associated with PGD, SLC7A5, and TKT was reduced. Recognizing telomere-related genes PGD, SLC7A5, and TKT as potential diagnostic biomarkers for OA is significant, as it offers valuable insights into the role of telomere-related genes in OA. This discovery also provides valuable information for the diagnosis and treatment of OA.