Abstract
PURPOSE: Early diagnosis of sepsis-induced acute respiratory distress syndrome (ARDS) is critical for effective treatment. We aimed to identify early stage biomarkers. MATERIALS AND METHODS: Differentially expressed genes were identified in whole blood samples from patients with sepsis or ARDS based on the Gene Expression Omnibus (GEO) datasets GSE32707, GSE54514 and GSE10361. Functional enrichment analysis explored the biological characteristics of differentially expressed genes. Genes with high functional connectivity based on a protein-protein interaction network were marked as hub genes, which were validated using the GEO dataset GSE76293, and a gene set variation analysis index (GSVA) was assigned. Diagnostic and predictive ability of the hub genes were assessed by receiver operating characteristic (ROC) curve analysis. DNA methylation levels of hub genes were quantified using the GEO dataset GSE67530. RESULTS: Forty-one differentially expressed genes were shared between sepsis-specific and ARDS-specific datasets. MAP2K2 and IRF7 functional activity was highly connected in sepsis-induced ARDS. Hub genes included RETN, MVP, DEFA4, CTSG, AZU1, FMNL1, RBBP7, POLD4, RIN3, IRF7. ROC curve analysis of the hub gene GSVA index showed good diagnostic ability in sepsis or ARDS. Among genes related to sepsis-induced ARDS, 17 were differentially methylated. Principal component analysis and heatmaps indicated that gene methylation patterns differed significantly between ARDS patients and controls. CONCLUSION: We identified a genetic profile specific to early-stage sepsis-induced ARDS. The abnormal expression of these genes may be caused by hypomethylation, which may serve as a biomarker for early diagnosis of ARDS.