Single-cell atlas of COVID-19 based on multiple sample integration.

Area of Science:

• Immunology
• Genomics
• Computational Biology

Background:

• Understanding the immune landscape in severe COVID-19 is crucial for clinical practice.
• Predicting patient outcomes remains a significant challenge.

Purpose of the Study:

• To construct a comprehensive single-cell atlas of COVID-19.
• To develop a predictive model for severe COVID-19 patient outcomes.

Main Methods:

• Single-cell RNA sequencing was performed on COVID-19, influenza, and healthy control samples.
• Analysis included cell abundance, transcriptomic profiles, and pathway activities.
• A logistic regression model was built to predict severe COVID-19 prognosis.

Main Results:

• The study analyzed 434,703 cells, identifying significant alterations in T/NK and myeloid cells.
• Differential gene expression in inflammatory and metabolic pathways correlated with disease severity.
• The predictive model, using IL1R2, PI3, IGHG3, and CTTN, achieved 80-81% sensitivity and ~70% AUC.

Conclusions:

• The single-cell atlas provides insights into COVID-19's cellular dynamics and mechanisms.
• The predictive model shows potential for aiding clinical decisions in severe COVID-19 cases.