<p>Autophagy plays a significant role in the pathogenesis of osteoarthritis (OA). The objective of this study was to identify diagnostic signature genes associated with OA from the autophagy-related genes (ARGs). This study first obtained gene expression profiles (GSE114007 and GSE57218) of OA cartilage from the GEO database and subsequently identified differentially expressed genes (DEGs) using the limma method. To further explore the biological functions of these DEGs in OA, Gene Ontology (GO) and KEGG enrichment analyses were conducted. Additionally, to screen for autophagy-related differentially expressed genes (ADEGs), the study visualized these genes using Venn diagrams. Next, the LASSO algorithm was applied to identify potential OA marker genes within the ADEGs. Finally, to validate the expression and effects of these genes, RT-qPCR analysis was performed in an in vitro OA model, and Western blotting was used to analyze the effects of gene overexpression in OA cells. The results demonstrated that PIM2, PTPN22, and RRAGD were identified as signature genes, exhibiting low expression levels with high diagnostic efficacy in OA samples from both experimental and test group datasets. Overexpression of PIM2, PTPN22, and RRAGD in OA cells showed suppression of inflammation-associated proteins (MMP13 and ADAMTS) and enhancement of cartilage matrix-associated proteins (Aggrecan and Collagen II). In conclusion, this study highlights PIM2, PTPN22, and RRAGD as novel and crucial diagnostic markers for OA through the utilisation of machine learning techniques, and may also offer novel insights for potential targeted therapeutic interventions in the future.</p> Graphical abstract <p></p>