Integration of bioinformatics and machine learning reveals oxidative stress-associated gene signatures and molecular subtypes in myositis

Oxidative stress (OS) contributes significantly to idiopathic inflammatory myopathy (IIM), but its molecular mechanisms remain unclear. This study aimed to investigate OS-related genes (OSRGs) in IIM, identify molecular subtypes, and explore potential biomarkers for diagnosis and treatment.

Differentially expressed genes (DEGs) between IIM and healthy controls were intersected with OSRGs to obtain DE-OSRGs. Consensus clustering based on DE-OSRGs defined OS subtypes. Machine learning identified key genes; a nomogram was constructed and validated using calibration curves, decision curve analysis, and external datasets. Candidate drugs were screened via DrugRep and assessed by molecular docking. Key gene expression was experimentally validated.

Forty-four DE-OSRGs were identified, enriched in oxidative metabolism and immune-inflammatory regulation. Cluster analysis classified IIM patients into two subgroups; C2 exhibited greater immune activation and cell infiltration, indicating a more severe inflammatory microenvironment. Two hub genes, CDKN1A and CAMK2G, were identified. A prediction model incorporating these genes demonstrated clinical potential. Drug screening identified multiple candidate compounds with strong binding affinity. Preliminary experimental validation confirmed the bioinformatics findings.

This study underscores the role of OS in IIM pathogenesis and identifies CDKN1A and CAMK2G as potential diagnostic markers and therapeutic targets. Key Points • Based on oxidative stress-related genes, we delineated two distinct OS-related subtypes in IIM, which exhibited significant biological and immunological differences. • Oxidative stress serves as a critical non-immune pathological driver in IIM progression.