[Identification of oxidative stress-related biomarkers in chronic rhinosinusitis with nasal polyps using WGCNA combined with machine learning algorithms].

Objective: To identify diagnostic markers related to oxidative stress in chronic rhinosinusitis with nasal polyps (CRSwNP) by analyzing transcriptome sequencing data, and to investigate their roles in CRSwNP. Methods: Utilizing four CRSwNP sequencing datasets, differentially expressed genes (DEGs) analysis, weighted gene co-expression network analysis (WGCNA), and three machine learning methods for Hub gene selection were performed in this study. Subsequent validation was carried out using external datasets, as well as real-time quantitative polymerase chain reaction (Real-time qPCR), and immunofluorescence staining of clinical samples. Moreover, the diagnostic efficacy of the genes was assessed by receiver operating characteristic (ROC) curve, followed by functional and pathway enrichment analysis, immune-related analysis, and cell population localization. Additionally, a competing endogenous RNA (CeRNA) network was constructed to predict potential drug targets. Statistical analysis and plotting were conducted using SPSS 26.0 and Graphpad Prism9 software. Results: Through data analysis and clinical validation, CP, SERPINF1 and GSTO2 were identified among 4 138 DEGs as oxidative stress markers related to CRSwNP. Specifically, the expression of CP and SERPINF1 increased in CRSwNP, whereas that of GSTO2 decreased, with statistically significant differences (P<0.05). Additionally, an area under the curve (AUC)>0.7 indicated their effectiveness as diagnostic indicators. Importantly, functional analysis indicated that these genes were mainly related to lipid metabolism, cell adhesion migration, and immunity. Single-cell data analysis revealed that SERPINF1 was mainly distributed in epithelial cells, stromal cells, and fibroblasts, while CP was primarily located in epithelial cells, and GSTO2 was minimally present in the epithelial cells and fibroblasts of nasal polyps. Consequently, a CeRNA regulatory network was constructed for the genes CP and GSTO2. This construction allowed for the prediction of potential drugs that could target CP. Conclusion: This study successfully identifies CP, SERPINF1 and GSTO2 as diagnostic and therapeutic markers related to oxidative stress in CRSwNP.