Machine learning and experiments revealed a novel pyroptosis-based classification linked to diagnosis and immune landscape in spinal cord injury

Abstract

Background: Rising evidence indicates the development of pyroptosis in the initiation and pathogenesis of spinal cord injury (SCI). However, the associated effects of pyroptosis-related genes (PRGs) in SCI are unclear. Methods: We obtained the gene expression profiles of SCI and normal samples in the GEO. Database: The R package limma screened for differentially expressed (DE) PRGs and performed functional enrichment analysis. Mechanical learning and PPI analysis helped filter essential PRGs to diagnose SCI. Peripheral blood was collected for validation from ten SCI patients and eight healthy individuals. The association of essential PRGs with immune infiltration was evaluated, and pyroptosis subtypes were recognized in SCI patients by unsupervised cluster analysis. Besides, a SCI model was built for in vivo validation of essential PRGs. Result: We identified 25 DE-PRGs between SCI and normal controls. Functional enrichment analysis revealed the principal involvement of DE-PRGs in pyroptosis, inflammasome complex, interleukin-1 beta production, etc. Subsequently, three essential PRGs were identified and validated, showing excellent diagnostic efficacy and significant correlation with immune cell infiltration. Additionally, we developed diagnostic nomograms to predict the occurrence of SCI. Two pyroptosis subtypes exhibited distinct biological functions and immune landscapes among SCI patients. Finally, the expression of these essential PRGswas verified in vivo. Conclusion: The current study described the vital effects of pyroptosis-related genes in SCI, providing a novel direction for effective assessment and management of SCI.

Keywords

• SCI
• Pyroptosis
• Machine learning
• Diagnosis
• Immunity