THE identification of four key genes and potential therapeutic compounds using machine learning could improve the diagnosis and treatment of sepsis. Sepsis remains a leading cause of mortality in intensive care units, with current biomarkers often lacking specificity and rapid response to treatment. Advances in machine learning offer new possibilities for early detection, personalised treatment, and prognosis evaluation. By integrating gene expression data and computational models, researchers aim to improve diagnostic accuracy and identify novel therapeutic targets.
In this study, gene expression datasets were analysed and combined following quality control and standardisation. A total of 405 differentially expressed genes were identified, including 334 upregulated and 71 downregulated genes. Further analysis using weighted gene co-expression network analysis refined this to 308 potential genes, which were subjected to 113 combined machine learning algorithms. This process identified 22 hub genes, and subsequent protein-protein interaction network analysis narrowed this down to four key genes: CD177, GNLY, ANKRD22, and IFIT1. Functional enrichment analysis highlighted immune response and bacterial infection as primary drivers of sepsis. The diagnostic model was validated using nomogram, decision curve analysis, and clinical impact curves, confirming its predictive reliability. Further molecular docking analysis identified three potential therapeutic compounds (Dieckol, Grosvenorine, and Tellimagrandin II) from a traditional Chinese medicine compound library.
These findings suggest that machine learning-based diagnostic models could enhance the early identification of sepsis, leading to more timely and targeted interventions. The discovery of key genes associated with sepsis also opens avenues for personalised treatment approaches. The identification of potential therapeutic compounds provides a foundation for future drug development, potentially improving survival rates in critically ill patients.
Jenna Lorge, EMJ
Reference
Zhang W et al. A diagnostic model for sepsis using an integrated machine learning framework approach and its therapeutic drug discovery. BMC Infect Dis. 2025;DOI:10.1186/s12879-025-10616-z.
