Abstract
BACKGROUND: Neonatal sepsis (NS) is a systemic inflammatory response to severe pathogenic infections, and is a major cause of high morbidity and mortality in newborns. Currently, there is a lack of efficient diagnostic technology to accurately and rapidly diagnose NS, and the precise pathogenesis of NS has yet to be fully elucidated. The present study aimed to identify the optimal biomarkers in the progression of NS. METHODS: The differentially expressed genes (DEGs) between NS and controls in the discovery datasets were screened. Gene set variation analysis (GSVA) was used to enrich the changes in biological functions and pathways in sepsis patients compared to healthy individuals. The differences in immune cell infiltration between these two groups were assessed using CIBERSORT. Furthermore, LASSO algorithm and ROC analysis were performed to identify and evaluate the gene signature. RESULTS: A total of 85 upregulated and 40 downregulated overlapping DEGs were screened in sepsis samples. The GSVA results indicated that DEGs largely contributed to upregulated inflammation and metabolism-related processes, and suppressed adaptive immune responses in NS. Markedly lower infiltration of most types of immune cell was observed in sepsis patients, except for some innate immune cells. Moreover, 57 genes with AUC >0.9 in both discovery sets were selected and applied to a LASSO model. Using this model, a seven-gene signature was acquired, which was validated in the discovery and independent validation sets. Five genes among the gene signatures with optimal diagnostic performance were obtained and further validated in clinical samples using RT-qPCR. Finally, three genes SLC2A3, OSCAR, and CD3G were identified as key biomarkers for NS. CONCLUSIONS: Our findings will provide novel insights into the pathogenesis of NS, and the potential biomarkers may have promising application values for its early detection and therapeutic intervention.