Identification of Genes as Potential Biomarkers for Sepsis-related ARDS using Weighted Gene Co-expression Network Analysis

Background: Acute respiratory distress syndrome (ARDS) caused by sepsis presents a high mortality rate; therefore, identification of susceptibility genes of sepsis to ARDS at an early stage is particularly critical. Methods: The GSE66890 dataset was downloaded and analyzed by WGCNA to obtain modules. Then, GO and KEGG analyses of the module genes were performed. Then, the PPI network and LASSO model were constructed to identify the key genes. Finally, expression levels of the screened genes were validated in clinical subjects. Results: We obtained 17 genes merged modules via WGCNA, and the dark module and tan module were the most positively and negatively correlated with sepsis-induced ARDS, respectively. Based on gene intersections of the module genes, 11 hub genes were identified in the dark module, and 5 hub genes were identified in the tan module. Finally, the six key genes were identified by constructing the LASSO model. We further detected the screened genes expression in clinical samples, and as the bioinformatics analysis revealed, the expressions of NANOG, RAC1, TWIST1, and SNW1 were significantly upregulated in the ARDS group compared to the sepsis group, and IMP3 and TUBB4B were significantly downregulated. Conclusion: We identified six genes as the potential biomarkers in sepsis-related ARDS. Our findings may enhance the knowledge of the molecular mechanisms behind the development of sepsisinduced ARDS.