| Objective To use machine learning algorithms to screen differentially apoptotic related genes (DARGs) in knee osteoarthritis, in order to further identify relevant biomarkers and to elucidate the role of apoptosis in the pathogenesis of knee osteoarthritis. Methods Firstly, the gene expression matrices of datasets GSE55235 and GSE169077 were downloaded and integrated from the GEO database. The differential analysis and WGCNA analysis on the integrated gene expression matrices were performed using R packages, respectively. Then, the two analysis results were intersected with apoptosis related genes obtained from the Genecards database to obtain DARGs, which were subjected to GO, KEGG, and GSEA enrichment analysis. Finally, the MCODE plugin in Cytoscape was used to screen the modules with the closest protein-protein interactions among DARGs. The Lasso algorithm was used to screen the hub gene biomarkers, and ROC curves were drawn in the osteoarthritis dataset GSE178557 for validation. The gene miRNA transcription factor drug regulatory network and immune infiltration analysis were performed on the hub gene biomarkers with good validation results. Results After screening, a total of 189 DARGs were obtained. After validation with a validation set, 7 reliable hub gene biomarkers were ultimately obtained. The miRNA regulatory network and transcription factor regulatory network of these genes were predicted. The potential drug and immune cell infiltration analysis results for hub gene biomarkers were obtained. Conclusion Apoptosis plays an important role in the pathogenesis of knee osteoarthritis. Seven hub gene biomarkers including TYROBP, COL14A1, CD74, LAMA4, COL5A1, MMP13, and IGFBP5 provide valuable clues for diagnosis and in-depth understanding of the mechanism of knee osteoarthritis. |