Objective:

To characterize immune-cell composition, glycolysis-related pathways, and intercellular interactions in sepsis, and to identify potential biomarkers using specific machine-learning techniques.

Key Findings:

• Sepsis showed myeloid dominance and elevated glycolysis indices, especially in monocytes and plasma cells, suggesting a potential target for therapeutic intervention.
• Five candidate biomarkers (GLRX, MDH1, MDH2, TGFBI, COPB2) were identified with strong discriminatory efficacy, warranting further investigation.
• TGFBI was enriched in monocytes and was central to a communication network involving B cells, plasma cells, and neutrophils, indicating its role in immune response.
• qRT-PCR confirmed significant differences for TGFBI in the CLP model, reinforcing its potential as a biomarker.

Interpretation:

The findings suggest that elevated glycolysis-related pathways are associated with a monocyte-focused communication framework involving TGFBI, highlighting its potential as a biomarker in sepsis.

Limitations:

• The study relies on publicly available datasets, which may introduce variability that could affect the robustness of the findings.
• The datasets used differ in disease stage and sample composition, limiting direct comparisons and generalizability of the results.

Conclusion:

The study emphasizes the role of glycolysis in sepsis and identifies TGFBI as a promising biomarker for further validation, highlighting the need for additional research.