Investigating Potential Biomarkers of Ankylosing Spondylitis: A Study on Mitochondrial and Senescence Pathways Using Machine Learning.

OBJECTIVE: Ankylosing spondylitis (AS) is a chronic immune-mediated inflammatory disorder characterized by inflammation and pathological bone formation. Growing evidence suggests that mitochondrial dysfunction and cellular senescence are key drivers of disease progression. This study aimed to identify novel biomarkers linking these processes to AS. METHODS: Transcriptomic datasets of AS patients and controls were analyzed to identify differentially expressed genes related to mitochondrial function and cellular senescence. Bioinformatics pipelines and multiple machine learning algorithms were used to screen candidate biomarkers, which were further validated in an independent dataset and in a collagen antibody-induced arthritis (CAIA) mouse model. Clinical diagnostic value was assessed using receiver operating characteristic analysis. RESULTS: We identified 25 mitochondrial- and 8 senescence-related genes differentially expressed in AS. Consensus machine learning analysis highlighted COX17 and MATK as robust candidates with significant diagnostic performance. Immune infiltration analysis suggested strong correlations between these genes and altered immune cell subsets. In vivo validation confirmed upregulation of COX17 and downregulation of MATK in the AS mouse model, accompanied by enhanced osteogenic activity. CONCLUSION: COX17 and MATK are promising biomarkers linking mitochondrial dysfunction and cellular senescence to AS. Their diagnostic potential highlights new avenues for improving early disease detection and personalized therapeutic strategies.