Cross-reactivity against drifted influenza A (H3N2) subclade K is predicted by pre-existing antibody signatures from recent vaccine targets

Abstract Seasonal influenza vaccine strains are selected months before the flu season begins, risking a vaccine mismatch. To that end, quickly predicting protection against newly emerged influenza strains is a public health preparedness need. Further, understanding immunological correlates of protection against mismatched targets is necessary for improving annual influenza vaccination compositions and/or platforms. Using a modified systems-predicted neutralizing antibody (SNAb) approach, we quantified predicted influenza virus neutralization responses in a human cohort pre- and post-vaccination during the 2025/2026 influenza season, during which the vaccine-mismatched H3N2 subclade K accounted for most influenza virus infections. This was done using the previous 10 years of vaccine targets, including the components of the 2025/2026 vaccination, as well as the 2026/2027 vaccine targets. We identified that H3N2 subclade K responders were predicted by baseline SNAb profiles, particularly to the H3N2 components of the 2024/2025 and 2025/2026 influenza vaccine. Systems serological analysis supported this at the multivariate level, with even greater accuracy in predicting mismatched responses from pre-vaccination humoral signatures. Lastly, we employed this approach to predict who would likely be high responders to the 2026/2027 influenza vaccine H1 component, and found that antibody features selected as predicting a strong neutralization response were disproportionately enriched for the H1 component of the 2025/2026 influenza vaccination. Our results support a model in which systems-based approaches can identify signatures predicting strong influenza vaccination responses before the influenza season begins.