VaccineScout: A Computational Framework for Guiding the Composition and Valency of SARS-CoV-2 Vaccine Updates

VaccineScout leverages global SARS-CoV-2 genomic data (14 million genomes, 2021–2025) to show over half of post-selection spike amino acid sequences were already circulating, supporting annual vaccine updates.

• Why it matters: Vaccine updates are still largely driven by serology, and there is no public computational tool to prioritize which variants to include and how many (valency) to target.
• What they did: The team analyzed spike sequences from 2021–2025 and built VaccineScout to rank candidates by minimizing average spike amino acid substitution distances to forecasted variant prevalence, then quantify diminishing returns for added variants.
• The result: Retrospective tests found VaccineScout-recommended candidates matched FDA recommendations or had lower spike distances to future variants and higher fitness, with near-optimal performance versus an oracle using future prevalence.

AI summary of the abstract · may contain errors · not medical advice · read the paper.