Human Genome Polymorphisms and Computational Intelligence Approach Revealed a Complex Genomic Signature for COVID-19...

Area of Science:

• Genetics
• Infectious Diseases
• Machine Learning

Background:

• Severe COVID-19 outcomes are influenced by a complex interplay of genetic and environmental factors.
• Identifying individuals at high risk for severe disease is crucial for effective public health strategies and personalized medicine.

Purpose of the Study:

• To develop and validate a machine learning model for predicting severe COVID-19 prognosis based on innate immunity gene polymorphisms.
• To identify specific single nucleotide polymorphisms (SNPs) associated with severe COVID-19 risk.

Main Methods:

• Genotyping of 296 innate immunity loci in 96 Brazilian severe COVID-19 patients and controls.
• Utilizing recursive feature elimination with support vector machine (SVM-RFE) for feature selection.
• Employing a support vector machine with a linear kernel (SVM-LK) for patient classification and prognosis.

Main Results:

• The SVM-RFE model identified 12 key SNPs in genes including PD-L1, IL10RA, JAK2, IFIT1, and IL10.
• The SVM-LK model achieved 85% accuracy, 80% sensitivity, and 90% specificity in classifying severe COVID-19 patients.
• Univariate analysis highlighted specific alleles in PD-L1, IFIT1, JAK2, and IFIH1 as risk or protective factors.

Conclusions:

• Genetic context plays a significant role in the development of severe COVID-19.
• The developed classification method can identify individuals at high risk for severe COVID-19 outcomes, even before infection.
• This predictive capability represents a novel approach to COVID-19 prognosis and risk stratification.