Introduction

The hemagglutinin (HA) proteins of seasonal influenza viruses continuously evolve to erode human neutralizing antibody immunity such that people are re-infected with an influenza A virus roughly every 5 years (Kucharski et al. 2018, Ranjeva et al. 2019). The influenza A subtypes that co-circulate in humans, H3N2 and H1N1 (descended from the 2009 H1N1 pandemic strain), acquire three to four and two to three amino acid substitutions in their HA proteins per year, respectively (Smith et al. 2004, Bedford et al. 2014).

To keep pace with this evolution, updates to the strains in the seasonal influenza vaccine are considered biannually with the goal of ensuring the vaccine strains are well matched to circulating strains. Recommendations of which strains to include in updated vaccines are made at vaccine-composition meetings, with a meeting in late winter to choose the strain for the upcoming Northern Hemisphere influenza season, and a meeting in early autumn to choose the strain for the upcoming Southern Hemisphere influenza season (World Health Organization 2024, n.d.-a, n.d.-b). The recommendations are based on a combination of viral sequence data reflecting the relative prevalence of different viral strains in the human population and serological data quantifying the antigenic properties of strains (Luksza and Lässig 2014, Neher et al. 2016, Huddleston et al. 2020, Shi et al. 2025). Historically, the serological data consisted of hemagglutination inhibition assays performed with sera from ferrets infected with defined viral strains (Smith et al. 2004, Jorquera et al. 2019). However, there has been increasing use of measurements made using human sera (Ampofo et al. 2015, Fonville et al. 2016, World Health Organization 2024, n.d.-a, n.d.-b, Kikawa et al. 2025) as there is growing recognition that the neutralizing antibody specificities of humans are more complex than that of singly infected ferrets (Linderman et al. 2014, Fonville et al. 2016, Cobey and Hensley 2017, Lee et al. 2019).

Following a period of low influenza genetic diversity during and after the COVID-19 pandemic, influenza evolution prior to the September 2025 vaccine-strain selection meeting was characterized by the rapid emergence and spread of multiple H3N2 variants carrying mutations at important antigenic sites (Huddleston et al. 2025, Sabaiduc et al. 2025). The September 2025 meeting recommended for the 2026 Southern Hemisphere vaccine an updated H3N2 strain from subclade J.2.4 and an updated H1N1 strain from subclade D.3.1 (World Health Organization n.d.-b) (throughout this paper, we use a new dynamic nomenclature system for influenza subclades; Neher et al. 2026). However, by the end of the 2025, additional antigenically divergent daughter subclades had emerged and become predominant for both subtypes: H3N2 subclade K (derived from J.2.4 with HA1 mutations K2N, S144N, N158D, I160K, Q173R, and T328A) and H1N1 subclade D.3.1.1 (derived from D.3.1 with HA1 mutations R113K, A139D, E283K, and K302E). Very recent experimental studies have shown that subclade K is antigenically distinct (Kirsebom et al. 2025, Cheng et al. 2026, Dee et al. 2026, Ikonen et al. 2026, Liu et al. 2026, Separovic et al. 2026, Wang et al. 2026), with human sera having approximately two-fold lower titres to subclade K than other recent H3N2 strains (Liu et al. 2026, Wang et al. 2026), although the exact magnitude of the titre decrease varied across cohorts and studies (Guiomar et al. 2026, Wilson et al. 2026). Additional HA mutations have since arisen in subvariants of both subclades K and D.3.1.1, although the effects of these new mutations have not been experimentally characterized prior to the current study.

Here, we characterize the neutralization by human sera of current human seasonal H3N2 and H1N1 strains to provide data to inform the upcoming decision of which strains to include in the 2026–2027 Northern Hemisphere vaccine. We do this by using a recently developed sequencing-based neutralization assay (Loes et al. 2024, Kikawa et al. 2025, Kikawa et al. 2026) that can rapidly measure titres of many human sera against many viral strains. We previously used this assay to generate data prior to the September 2025 vaccine-strain selection decision (Kikawa et al. 2025). This study takes a similar approach but uses an updated set of H3N2 and H1N1 strains representative of those currently circulating in the human population (including multiple subvariants with subclades K and D.3.1.1) and a new set of human sera collected mostly in October to November 2025. We measure a total of 27 409 neutralization titres to provide a near real-time portrait of human neutralizing antibody immunity to seasonal influenza.