Bioinformatics Identification of Immunomodulatory Genes Related to Neonatal Sepsis and Their Incorporation into a Diagnostic Model

Objective Although neonatal sepsis (NS) is a main driver of neonatal morbidity and mortality, reliable molecular biomarkers for early detection are lacking. This study identified immunomodulation-related differentially expressed genes (IMRDEGs) linked to NS through integrated bioinformatics analysis. Methods Two GEO datasets (GSE25504 and GSE69686) containing 90 NS and 122 control samples were combined via the R packages "GEOquery" and "sva". Differentially expressed genes (DEGs) were detected via "limma", and functional enrichment was determined via the GO and KEGG databases. Enriched pathways were further identified via gene set enrichment analysis (GSEA) and gene set variation analysis (GSVA). Then, we developed a diagnostic model via logistic regression (LR), SVM-RFE, and LASSO regression. Results In total, 360 DEGs were identified, including 69 IMRDEGs. Enrichment analyses highlighted significant associations with inflammatory and immune regulation pathways. Seven hub genes (HGs; ARG2 , IL18R1 , IL1RN , MERTK , RETN , STAT3 , and TSPO ) were incorporated into the diagnostic model, which displayed high accuracy (AUC > 0.9) in ROC curve analysis. Immune infiltration analysis elucidated close interconnections between the HGs and specific immune cell (IC) subsets. Conclusion These outcomes illustrate that the detected HGs represent biomarkers for early NS diagnosis and provide insights into potential therapeutic targets. Upcoming studies should concentrate on the functional validation and clinical translation of these biomarkers.