Integrated analysis of multi-omics data reveals T cell exhaustion in sepsis.

Background: Sepsis is a heterogeneous disease, therefore the single-gene-based biomarker is not sufficient to fully understand the disease. Higher-level biomarkers need to be explored to identify important pathways related to sepsis and evaluate their clinical significance. Methods: Gene Set Enrichment Analysis (GSEA) was used to analyze the sepsis transcriptome to obtain the pathway-level expression. Limma was used to identify differentially expressed pathways. Tumor IMmune Estimation Resource (TIMER) was applied to estimate immune cell abundance. The Spearman correlation coefficient was used to find the relationships between pathways and immune cell abundance. Methylation and single-cell transcriptome data were also employed to identify important pathway genes. Log-rank test was performed to test the prognostic significance of pathways for patient survival probability. DSigDB was used to mine candidate drugs based on pathways. PyMol was used for 3-D structure visualization. LigPlot was used to plot the 2-D pose view for receptor-ligand interaction. Results: Eighty-four KEGG pathways were differentially expressed in sepsis patients compared to healthy controls. Of those, 10 pathways were associated with 28-day survival. Some pathways were significantly correlated with immune cell abundance and five pathways could be used to distinguish between systemic inflammatory response syndrome (SIRS), bacterial sepsis, and viral sepsis with Area Under the Curve (AUC) above 0.80. Seven related drugs were screened using survival-related pathways. Conclusion: Sepsis-related pathways can be utilized for disease subtyping, diagnosis, prognosis, and drug screening.