Altered somatic hypermutation patterns in COVID-19 patients classifies disease severity

Area of Science:

• Immunology
• Infectious Diseases
• Computational Biology

Background:

• The human immune response to SARS-CoV-2 infection involves lymphocytes and their antigen receptors.
• Identifying and characterizing these receptors is crucial for understanding and combating COVID-19.
• Current methods require improvement for accurate stratification of disease severity.

Purpose of the Study:

• To apply a machine learning approach to B cell receptor (BCR) repertoire sequencing data.
• To differentiate between infected and non-infected individuals, and between mild and severe COVID-19 cases.
• To identify features within BCR repertoires that correlate with infection status and severity.

Main Methods:

• Utilized BCR repertoire sequencing data from individuals with varying SARS-CoV-2 infection levels (severe, mild, and uninfected controls).
• Applied a machine learning model to analyze the sequencing data.
• Focused on somatic hypermutation patterns as key classification features.

Main Results:

• The machine learning approach successfully stratified individuals based on infection status (infected vs. non-infected).
• The model also accurately stratified disease severity (mild vs. severe COVID-19).
• Somatic hypermutation patterns were identified as the primary features driving this classification, indicating alterations in this process in COVID-19 patients.

Conclusions:

• The study demonstrates a novel machine learning application for analyzing BCR repertoires in COVID-19.
• Identified somatic hypermutation patterns as significant biomarkers for infection and severity.
• These findings provide a proof of concept for developing quantitative diagnostic and therapeutic antibody strategies for COVID-19 and future epidemiological challenges.