Sequence-based antigenic change prediction by a sparse learning method incorporating co-evolutionary information.

Area of Science:

• Virology
• Computational Biology
• Immunology

Background:

• Influenza antigenic drift, driven by mutations at antigenic sites, necessitates continuous vaccine updates.
• Current methods for identifying antigenic variants often focus on individual mutation sites.
• The cumulative impact of multiple simultaneous mutations on influenza antigenicity is not fully understood and may not be additive.

Purpose of the Study:

• To develop and apply a novel computational method, AntigenCO, for identifying and quantifying both single and co-evolutionary sites driving influenza antigenic drifts.
• To assess the performance of AntigenCO in predicting antigenic variants compared to existing single-site analysis methods.

Main Methods:

• Development of the AntigenCO computational method to analyze co-evolutionary patterns at influenza antigenic sites.
• Application of AntigenCO to historical influenza data to identify key sites contributing to antigenic drift.
• Comparative analysis of AntigenCO's predictive accuracy against methods relying on single-site mutations.

Main Results:

• AntigenCO accurately identified single and co-evolutionary sites responsible for historical influenza antigenic drifts.
• The developed method achieved a prediction accuracy of up to 90.05% for antigenic variants.
• AntigenCO significantly outperformed traditional methods based on single-site analyses.

Conclusions:

• AntigenCO provides a more accurate approach to antigenic variant identification by considering co-evolving sites.
• This computational tool can enhance influenza surveillance and aid in selecting optimal vaccine candidates.
• Quantifying the impact of simultaneous mutations is critical for understanding and predicting influenza evolution.