Identification of COVID-19-Specific Immune Markers Using a Machine Learning Method.

Area of Science:

• Immunology
• Genomics
• Computational Biology

Background:

• Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) significantly impacts the human immune system, involving distinct stages of immune defense and inflammation.
• COVID-19's immune response involves various cells like CD4+ T cells, CD8+ T cells, monocytes, dendritic cells, B cells, and natural killer cells, differentiating it from other respiratory diseases.

Purpose of the Study:

• To identify specific immune cell markers for differentiating COVID-19 from common inflammatory responses, other severe respiratory diseases, and healthy states.
• To establish quantitative rules for distinguishing disease status based on immune cell gene expression profiles.

Main Methods:

• Single-cell gene expression profiling of six immune cell types.
• Application of Boruta and mRMR feature selection methods to identify significant cell markers.
• Utilized the IFS method to determine optimal feature subsets and classifiers for two classification algorithms.

Main Results:

• Identified specific features like IFI44L in B cells, S100A8 in monocytes, and NCR2 in natural killer cells associated with the innate immune response in COVID-19.
• Highlighted ZFP36L2 in CD4+ T cells as a regulator of the COVID-19 inflammatory process.
• Established quantitative rules for disease status classification based on identified markers.

Conclusions:

• The identified cell markers and quantitative rules provide a basis for distinguishing COVID-19 from other conditions.
• This research offers theoretical support for further investigation into COVID-19 pathogenesis and the development of targeted intervention strategies.