Predictability of antigenic evolution for H3N2 human influenza A virus

Area of Science:

• Virology
• Immunology
• Computational Biology

Background:

• Influenza A virus poses a significant public health risk due to rapid antigenic evolution.
• Effective vaccine development requires accurate prediction of viral antigenic drift.
• Existing empirical methods lack insight into mutation effects and evolutionary mechanisms.

Purpose of the Study:

• To develop a theoretical method for predicting H3N2 human influenza A virus antigenic evolution.
• To estimate antigenic distance by evaluating de novo mutations.
• To understand the mechanisms driving influenza A virus antigenic evolution.

Main Methods:

• Developed a theoretical model to predict antigenic evolution of H3N2 influenza A virus.
• Estimated antigenic distance using hemagglutination inhibition (HI) titers.
• Incorporated amino acid properties (volume, isoelectric point, accessibility) and locations (receptor-binding sites, glycosylation sites) in hemagglutinin 1 (HA1) into antigenic models.

Main Results:

• The best theoretical model showed relatively low prediction accuracy for dominant strains.
• Identified a tendency for amino acid sites with greater potential impact on antigenicity to evolve.
• Observed a higher likelihood of amino acid changes with larger potential effects occurring.

Conclusions:

• Natural selection may play a role in enhancing the antigenic evolution of H3N2 influenza A virus.
• The theoretical model provides insights into potential drivers of viral antigenic change.
• Further refinement of predictive models is needed for improved vaccine strain selection.