Multiomics Profiling Identifies Blood-Based Diagnostic Markers for Sepsis

Area of Science:

• Biomolecular Signatures
• Infectious Disease Pathogenesis
• Machine Learning in Medicine

Background:

• Sepsis presents a significant clinical challenge due to rapid immune dysregulation and multiorgan failure.
• Lack of stable biomarkers hinders early sepsis diagnosis and risk assessment.
• Transcriptomic profiling and machine learning offer potential for biomarker discovery.

Purpose of the Study:

• To identify and validate a novel gene signature for sepsis diagnosis and severity stratification.
• To elucidate the cellular localization and molecular characteristics of identified sepsis biomarkers.
• To confirm the translational relevance of the diagnostic panel in clinical settings.

Main Methods:

• Large-scale transcriptomic profiling integrated with machine learning algorithms.
• Single-cell RNA sequencing to determine cellular expression patterns.
• Murine cecal ligation and puncture (CLP) models and in vitro LPS-stimulated cellular models for validation.
• Independent clinical serum sample validation.

Main Results:

• A three-gene signature (TLR5, HMGB2, C19orf59) was identified for sepsis diagnosis.
• Gene upregulation was localized to myeloid immune cells (monocytes, neutrophils).
• TLR5 and HMGB2 effectively stratified high-risk sepsis, while C19orf59 showed consistent efficacy across all severity levels.
• In vivo and in vitro models confirmed persistent signature upregulation.
• HMGB2 exhibited a biphasic kinetic profile indicative of danger-associated molecular patterns (DAMPs).
• Clinical validation confirmed the translational relevance of the identified biomarkers.

Conclusions:

• TLR5, HMGB2, and C19orf59 constitute a reliable diagnostic and severity-stratification panel for sepsis.
• These genes provide novel molecular insights into the septic pathological cascade.
• The findings support the clinical utility of this signature for improved sepsis management.