Hepatitis C Virus Saint Petersburg Variant Detection With Machine Learning Methods.

Area of Science:

• Virology
• Molecular Epidemiology
• Bioinformatics

Background:

• Hepatitis C virus (HCV) infection is a major global health issue.
• Direct-acting antivirals (DAAs) are effective but treatment failures occur, especially with specific HCV genotypes like the 2k/1b recombinant variant.
• Accurate HCV genotyping is critical for effective treatment, particularly when pan-genotypic DAAs are not accessible.

Purpose of the Study:

• To develop and validate machine learning models for predicting the Hepatitis C virus 2k/1b recombinant variant.
• To create an accessible tool for physicians and researchers to identify HCV genotypes, including the 2k/1b variant.

Main Methods:

• Machine learning models were trained using sequence data from nonstructural proteins of HCV 1b and 2k/1b variants.
• Model performance was evaluated for predicting the 2k/1b variant.
• An open-access tool was developed integrating these predictive models.

Main Results:

• Machine learning models demonstrated high accuracy in predicting the Hepatitis C virus 2k/1b variant.
• The developed tool provides a reliable method for HCV genotyping, including the challenging 2k/1b variant.
• The study highlights the utility of computational approaches in molecular epidemiology.

Conclusions:

• Machine learning offers a robust solution for accurate Hepatitis C virus 2k/1b variant detection.
• The developed open-access tool can assist clinicians in selecting appropriate treatments, especially in regions where pan-genotypic DAAs are limited.
• Molecular epidemiology tools are essential for monitoring and managing the spread of specific HCV genotypes.