Utilizing virus genomic surveillance to predict vaccine effectiveness.

As new vaccines are being developed for fast-evolving viruses, determining when and how to update them, and what data should inform these decisions, remains a significant challenge. We developed a model to inform these vaccine updates in near real-time and applied it to SARS-CoV-2 by quantifying the relationship between vaccine effectiveness (VE) and genetic distance from mRNA vaccine formulation sequences using 10,156 genomes from Connecticut (April 2021-July 2024) and data from over one million controls, employing a two-stage statistical approach. We showed a strong inverse correlation between spike gene amino acid distance and VE; every 10 amino acid substitutions away from the vaccine sequences resulted in a 15.4% (95% credible intervals (CrI): -2.0%, 34.6%) reduction in VE. Notably, this framework allows us to quantify the anticipated impact of emerging variants on VE, as demonstrated by the predicted 43.4% (95% CrI: -5.7%, 90.1%) drop in VE for the 2023/24 vaccine following the emergence of JN.1 variants based on sequence data alone. By linking amino acid substitutions to VE, this approach leverages genomic surveillance to monitor population-level protection and inform timely vaccine updates.