Integrated machine learning and clinical validation identify COL21A1 as a potential biomarker for sarcopenia

Sarcopenia is a progressive skeletal muscle disorder characterized by loss of mass, strength, and function. However, reliable molecular biomarkers for early detection remain limited. This study aimed to identify and validate potential biomarkers for sarcopenia through a combination of machine learning analysis and experimental validation. Differential expression analysis between sarcopenia and control samples was performed using the GSE1428 dataset. Three machine learning algorithms were applied to screen candidate hub genes, and the diagnostic performance was evaluated by receiver operating characteristic curve analysis. Candidate gene expression was further validated in independent clinical samples using quantitative real‑time PCR (qRT‑PCR), Western blotting, and immunohistochemistry. Functional analyses were conducted to explore the underlying mechanisms. A total of 39 differentially expressed genes were identified. The intersection of machine learning algorithms identified four hub genes: C1QA , COL21A1 , SLC38A1 , and HOXB2 . All four genes showed good diagnostic accuracy, with COL21A1 achieving an area under the curve value of 0.967. Clinical validation by qRT‑PCR showed that among the four hub genes, only COL21A1 was significantly upregulated at the mRNA level in sarcopenia samples compared with controls. Western blot confirmed that COL21A1 protein expression was significantly increased in sarcopenia. Immunohistochemistry showed enhanced cytoplasmic and extracellular matrix staining in sarcopenic samples. Functional analyses indicated significant associations between COL21A1 expression and immune‑related pathways. This study identified and validated COL21A1 as a potential candidate biomarker for sarcopenia. These findings offer new molecular insights for sarcopenia and may suggest a candidate biomarker for future therapeutic investigation.