Molecular Classification of Patients With COVID-19 Based on Transcriptional Profiling

Area of Science:

• Immunology
• Genomics
• Computational Biology

Background:

• COVID-19 remains a global health crisis with diverse patient immune responses complicating treatment.
• Molecular classification is crucial for developing precision therapies for COVID-19.

Purpose of the Study:

• To classify COVID-19 patients into distinct molecular endotypes.
• To identify reliable biomarkers for endotype-specific patient stratification.

Main Methods:

• Unsupervised consensus clustering of blood transcriptomic data from 351 COVID-19 patients.
• Validation in an independent cohort of 56 patients.
• Application of XGBoost, LASSO, and random forest algorithms for biomarker identification.

Main Results:

• Identification of three distinct COVID-19 endotypes with unique biological and clinical profiles.
• Endotype 1: favorable outcomes, DNA replication, IL7. Endotype 2: hypoxia, angiotensin pathways. Endotype 3: TLR4 activation, impaired NK cytotoxicity, poor outcomes.
• Discovery of predictive biomarker pairs (e.g., STAT4:S100A11) for accurate endotype classification via RT-qPCR.

Conclusions:

• This molecular classification highlights COVID-19 heterogeneity.
• Proposes biomarker-guided strategies for improved patient stratification and management.