The high disability rate of osteoarthritis (OA), a joint disease with an insidious onset and widespread effects, places a heavy financial burden on patients, families, and society. Traditional diagnostic approaches, including radiology and physical examination, cannot achieve early-stage screening of OA and thus, miss early intervention for patients. Therefore, the need of biomarkers for the early diagnosis of OA is crucial.

390 differentially expressed genes (DEGs) were identified from the training set, and 1077 key module genes were found by constructing a weighted gene co-expression network, and 161 key genes were obtained as a result. Four diagnostic marker genes highly associated with OA were screened for key genes using machine learning algorithms, and the resulting nomogram model showed excellent predictive power and clinical value. After further background studies, immune infiltration and functional enrichment analysis, we found that FKBP5 may play an important role in the prognosis and immune infiltration of multiple cancers, and this hypothesis was verified by pan-cancer analysis.

We screened four diagnostic marker genes (FKBP5, EPYC, KLF9 and PDZRN4) that are highly associated with OA. And this led to a diagnostic model, which was assessed to have good predictive power and clinical value. FKBP5 may be a potential intervention target for human diseases such as osteoarthritis and tumors.