Twice a year, influenza experts engage in a critically important forecasting bet: pick three to four vaccine strains now that will match the flu viruses people will face in 8-12 months. The strains used in seasonal influenza vaccines are selected by the World Health Organization (WHO), based on recommendations from a global network of scientists and public health laboratories that monitor influenza viruses.
The challenge is that influenza A, the virus responsible for seasonal flu epidemics, is a moving target. When an individual is exposed to the influenza virus by infection or vaccination, their body develops neutralizing antibodies targeting viral surface proteins, primarily hemagglutinin (HA). HA rapidly evolves, accumulating two to four amino acid changes each year. Those edits can enable the virus to evade immune targeting, which is why vaccine strain selection is so hard to get right.
In a recent study published in Virus Evolution, Caroline Kikawa and colleagues in the Bloom Lab at Fred Hutch used recently collected human blood samples to generate a near-real-time dataset detailing how well human antibodies neutralize the diversity of influenza A viruses in circulation. To do this, they developed a high-throughput assay to test antibodies from 188 human serum samples across multiple cohorts collected between October 2024 and April 2025, spanning a wide range of ages and geographic locations. They assayed these sera against a library of 140 influenza A strains representing the diversity of human H3N2 and H1N1 viruses, two major “subtypes” of human influenza A virus, as of summer 2025.
“How to choose which strains to include in the vaccine is a complex problem.” Lead author Caroline Kikawa shares. “Some of the data we use to make these decisions include measurements of how the immune system perceives different influenza viruses; these data are acquired from “neutralization assays,” where we assay the ability of different sera–the component of blood that contains antibodies–to neutralize different influenza viruses. Our work is an application of a new, sequencing-based neutralization assay which was developed in the Bloom lab that allows us to simultaneously measure neutralization of >100 viruses in close to real-time.”
The key technical trick is barcoding. Each viral strain in the pooled library carries a unique nucleotide “tag,” so the researchers can mix many viruses together, expose that mixture to serial dilutions of a single serum, and then use sequencing to read out the tags and determine how well each virus strain infects cells in the presence of virus-neutralizing antibodies from the serum. From these data, for each virus strain they calculated neutralization titers, or the dilution of serum that reduced virus infectivity by 50%, with a higher titer indicating stronger antibody protection.
The result is a massive dataset: 26,148 neutralization titer measurements capturing how heterogeneous human immunity to influenza can be, both across individuals and across virus strains. So, what did the landscape show? For H3N2, the titers against the most recent cell-derived 2025-2026 vaccine strain were in the same ballpark as titers against many other contemporary strains in the library (representing “today’s” flu viruses). In other words, for most circulating H3N2 flu viruses, people’s antibodies worked as equally well as they did against the current vaccine strain. This suggests the virus hasn’t changed much from the vaccine strain. However, the data also revealed a few small groups of viruses that antibodies struggled to neutralize–early warning signs of strains that could become a bigger problem if they start spreading widely. In contrast, serum from different individuals showed fairly uniform neutralization across all recent H1N1 strains, suggesting less antigenic divergence, in line with H1N1 being known to evolve more slowly than H3N2.
This work represents an exciting proof-of-principle that such assays can produce population-level immune measurements fast enough to potentially inform real-time public health decisions, rather than only retrospective analyses. “We performed this work on a timeline to provide the data publicly in advance of the WHO vaccine choice meeting in September 2025. At that meeting, these data were used to rationalize an updated component of the Southern Hemisphere 2026 seasonal influenza vaccine.” Kikawa adds.
“We’re repeating this effort now for the Northern Hemisphere meeting in February 2026. We’re looking forward to an in-depth post-hoc analyses of all these rich datasets. I’m really excited to dig into the many questions we could explore with these large datasets. Do different age groups have different patterns of susceptibility to certain variants of seasonal influenza? Can we use these data to improve predictions of which variants might emerge or succeed during a given flu season?”
Together, the work illustrates a future where influenza vaccine updates can be guided not only by viral genetics, but by broad, rapidly generated measurements of which flu strains people’s immune systems can currently tackle.
The spotlighted research was funded by the National Institutes of Health NIAID, the Howard Hughes Medical Institute, and the UK Medical Research Council.
Kikawa C, Huddleston JA-O, Loes AN, Turner SA, Lee J, Barr IA-OX, Cowling BJ, Englund JA-O, Greninger AL, Harvey R et al. 2025. Near real-time data on the human neutralizing antibody landscape to influenza virus to inform vaccine-strain selection in September 2025. Virus Evolution. doi: 10.1093/ve/veaf086
Science Spotlight writer Kelly Mitchell is a postdoctoral fellow in the Paddison Lab at Fred Hutch Cancer Center. She utilizes live cell reporters and CRISPR screening to study how glioblastoma cancer cells resist chemotherapy and radiation treatment. She obtained her PhD in cellular biology from Albert Einstein College of Medicine.
For the Media