Identification of DNA damage repair-related genes in sepsis using bioinformatics and machine learning: An observational study.

Abstract

Sepsis is a life-threatening disease with a high mortality rate, for which the pathogenetic mechanism still unclear. DNA damage repair (DDR) is essential for maintaining genome integrity. This study aimed to explore the role of DDR-related genes in the development of sepsis and further investigated their molecular subtypes to enrich potential diagnostic biomarkers. Two Gene Expression Omnibus datasets (GSE65682 and GSE95233) were implemented to investigate the underlying role of DDR-related genes in sepsis. Three machine learning algorithms were utilized to identify the optimal feature genes. The diagnostic value of the selected genes was evaluated using the receiver operating characteristic curves. A nomogram was built to assess the diagnostic ability of the selected genes via "rms" package. Consensus clustering was subsequently performed to identify the molecular subtypes for sepsis. Furthermore, CIBERSORT was used to evaluate the immune cell infiltration of samples. Three different expressed DDR-related genes (GADD45A, HMGB2, and RPS27L) were identified as sepsis biomarkers. Receiver operating characteristic curves revealed that all 3 genes showed good diagnostic value. The nomogram including these 3 genes also exhibited good diagnostic efficiency. A notable difference in the immune microenvironment landscape was discovered between sepsis patients and healthy controls. Furthermore, all 3 genes were significantly associated with various immune cells. Our findings identify potential new diagnostic markers for sepsis that shed light on novel pathogenetic mechanism of sepsis and, therefore, may offer opportunities for potential intervention and treatment strategies.