Skip to main content
NCBI home page
Virus Evolution logoLink to Virus Evolution
. 2026 Feb 16;12(1):veag005. doi: 10.1093/ve/veag005

Introduction and inter-species transmission dynamics of high pathogenicity avian influenza H5N1 viruses in Japan 2021–25

Yik Lim Hew 1, Claire Guinat 2, Manon Couty 3, Diletta Fornasiero 4,5, Takahiro Hiono 6,7,8,9, Norikazu Isoda 10,11,12,13,✉, Yoshihiro Sakoda 14,15,16,17
PMCID: PMC12908673  PMID: 41705171

Abstract

High pathogenicity avian influenza virus impacts poultry and wild birds worldwide. Since the introduction of clade 2.3.4.4b H5N1 isolates in Japan in late 2021, new cases have been reported in domestic birds and poultry each winter. To understand the role of wild birds in introducing these viruses in Japan, a phylodynamic analysis based on geography and host species was conducted using H5N1 isolates in Japan from 2021 to 2025. A total of 892 hemagglutinin gene sequences of H5N1 viruses, collected from birds between June 2021 and May 2025, along with additional sequences that were highly similar to those of Japanese isolates, were obtained from a public database. The role of wild birds in the transmission dynamics of H5N1 isolates in Japan across four winter seasons (2021–25) was assessed using a Bayesian phylodynamic approach with a Multi-Type Birth-Death model. Phylodynamic analysis revealed that the clade 2.3.4.4b comprised three distinct subgroups, G2b, G2c, and G2d, which were prevalent during the winter seasons. Isolates from G2b and G2c were linked to common ancestral strains from North Asia and Northeast Asia, respectively. Meanwhile, G2d, the dominant strain in Japan from 2021 to 2025, shared an ancestral strain from the Northwest America. During winters 2023–25, the ancestral strain was traced back to Northeast Asia, indicating a shift in the viral origin. This transition suggests an increase in virus migration events and expansion of host diversity, implying that Japan may function as a hub for intercontinental virus introductions, receiving multiple independent viral entries from North Asia, Northeast Asia, and Northwest America. Additionally, waterfowl and raptors played a role in introducing viruses into Japan, while poultry and crows generally serving as dead-end hosts. However, the continuous introduction of H5N1 isolates into Japan each winter can alter the disease transmission pattern observed in crows, resulting in more virus spillover from crows to other hosts, such as poultry and charadriiformes. These findings emphasize the importance of continuous monitoring and prompt information sharing to understand the global dynamics of viruses better.

Keywords: H5N1, high pathogenicity avian influenza virus, phylodynamic, Japan, wild birds

Introduction

High pathogenicity avian influenza (HPAI) viruses of the H5N1 subtype (HPAIV) in clade 2.3.4.4b pose a significant threat to wild birds and poultry worldwide. These viruses originate from the A/goose/Guangdong/1/1996 (H5N1: Gs/Gd) lineage and have been spreading in poultry and wild birds since 2003. This lineage has evolved into multiple clades with multiple antigenic variants ‘via’ genetic mutations and reassortments with other avian influenza viruses (WHO/OIE/FAO H5N1 Evolution Working Group 2008). Among these, clade 2.3.4.4b has rapidly expanded since 2016 (Pohlmann et al. 2022). In late spring 2021, continued circulation of clade 2.3.4.4b HPAIVs in Europe led to the emergence of multiple reassortant genotypes through genetic reassortment between H5N8 HPAIVs and locally circulating low pathogenicity avian influenza viruses (European Food Safety Authority, European Centre for Disease Prevention, Control, European Union Reference Laboratory for Avian Influenza et al. 2021). This resulted in repeated outbreaks in poultry, wild birds, and, more recently, mammals across Asia, Europe, Africa, and the Americas (Kandeil et al. 2023, Ruiz-Saenz et al. 2023), and even in Antarctica (Banyard et al. 2024). From late 2021 onward, clade 2.3.4.4b HPAIVs showed increased prevalence and diversification in the Far East region in Asia, including Japan, South Korea, Russia, and China, likely driven by internal and intercontinental bird migrations (Isoda et al. 2022, Lee et al. 2023, Yang et al. 2023).

After H5N1 HPAIV entered Japan in November 2021, successive seasonal outbreaks have occurred among poultry and wild birds. The hemagglutinin (HA) genes of H5N1 and H5N8 HPAIVs circulating in Japan in winter 2021–22 belonged to group 2 (G2), which arose from the H5 HPAIVs that circulated in Europe in late 2020 (Baek et al. 2021). G2 is further divided into multiple subgroups (G2a–G2e) based on a naming system specific to Japan, and the origins of these subgroups have been previously described (Hew et al. 2024a, Isoda et al. 2025). Among them, G2b and G2d were primarily identified in Japan during winter 2021–22 (Takadate et al. 2023). In the following winter season, the outbreak worsened, coinciding with the emergence of multiple G2 subgroups (G2b, G2c, and G2d), indicating increased viral diversity (Hew et al. 2024a). Despite a temporary decline in the outbreak numbers in winter 2023–24, a resurgence occurred in the following winter 2024–25, with viruses in the G2d subgroup remaining prevalent. Unlike in Europe, H5 HPAIVs have not been detected in Japan during the summer, supporting the hypothesis of annual winter introductions from neighbouring countries (Isoda et al. 2025). Moreover, H5N1 HPAIV cases have been reported across multiple host species, including the large-billed crow (Corvus macrorhynchos), white-naped cranes (Grus vipio), white-tailed sea eagle (Haliaeetus albicilla), slaty-backed gull (Larus schistisagus), and Galliformes, such as chicken, emu, quail, and guinea fowl (Takadate et al. 2024, Hew et al. 2024a, Isoda et al. 2025, Esaki et al. 2025a). Considering the repeated H5N1 HPAIV outbreaks in Japan across multiple host species and the detection of viruses from multiple G2 subgroups, understanding the origin and transmission dynamics of these viruses in different species is imperative. H5N1 HPAIVs were consistently detected in Japan from 2021 to 2025, but other subtypes, such as H5N8 and H5N5 HPAIVs, were limited in Japan. Similar to the pattern observed in Europe, H5N1 HPAIVs replaced H5N8 HPAIVs in 2021–22, and H5N5 HPAIVs were introduced from Europe, mainly in winter 2023–24, ‘via’ a distinct bird migration route (Takadate et al. 2023, Hew et al. 2024b).

Recent phylodynamic analyses have revealed continuous viral exchanges among East Asian countries such as China, South Korea, and Japan, reflecting persistent lineage turnover and gene flow within wild bird populations along the East Asian–Australasian Flyway (Seo et al. 2025). Furthermore, H5N1 viruses closely related to East Asian lineages have been found in North America, including Alaska and Canada, indicating sporadic transcontinental dispersal ‘via’ the Trans-Beringian route or the West Pacific Flyway (Lee et al. 2015, Alkie et al. 2022). Although these introductions have contributed only marginally to the overall viral diversity in North America, this underscores the role of overlapping migratory flyways in linking continental bird populations (Signore et al. 2025). Overall, these findings demonstrate how phylodynamic approaches can reveal intercontinental connections and the evolutionary processes driving the global spread of H5N1 HPAIV. In this study, a Bayesian phylodynamic framework was used to analyse a comprehensive sequence dataset of 892 H5N1 HPAIV HA gene sequences collected in Japan from 2021 to 2025, together with additional sequences from other countries that are closely related to Japanese isolates, obtained from a public database. Using this approach, the geographic origins of virus introductions and the transmission patterns between bird species can be inferred. By combining sequence data with host, temporal, and geographic metadata, phylodynamic models can effectively reconstruct viral spread and evolution (Guinat et al. 2021, Featherstone et al. 2022, Carnegie et al. 2023), helping establish epidemiological links between wild and domestic bird populations as well as facilitating national surveillance and control strategies.

Materials and methods

Sample collection and virus isolation

Active and passive surveillance is conducted annually at the Laboratory of Microbiology, Hokkaido University, Sapporo, Japan, to monitor the circulation of HPAIV in Hokkaido, the northernmost prefecture of Japan. Detailed information about sampling sites for active surveillance and methods for passive surveillance has been published previously (Hew et al. 2024a, Isoda et al. 2025). Tracheal and cloacal swab samples were collected from bird carcasses and immediately mixed with a virus transport medium, consisting of Minimum Essential Medium (Shimadzu Diagnostics Corporation, Tokyo, Japan) containing 10 mg/ml streptomycin (Meiji Seika Pharma, Tokyo, Japan), 10 000 U/ml penicillin G (Meiji Seika Pharma, Tokyo, Japan), 250 U/ml nystatin (Sigma-Aldrich, St. Louis, MO), 0.3 mg/ml gentamicin (MSD, Tokyo, Japan), and 0.5% bovine serum albumin fraction V (Roche, Basel, Switzerland). This mixture was then inoculated into 10-day-old embryonated eggs; the virus was isolated from the allantoic fluid and confirmed using a hemagglutination assay (Yamamoto et al. 2011).

Genome sequencing

To evaluate the pathogenicity of the isolates, viral RNA was extracted from the allantoic fluid using either TRIzol LS Reagent (Thermo Fisher Scientific, Waltham, MA) or the QIAamp Viral RNA Mini Kit (QIAGEN, Hilden, Germany). The HA gene was amplified by polymerase chain reaction (PCR) with a specific primer set to confirm the presence of multiple basic amino acid residues at the HA cleavage site, a molecular marker of HPAIV (Perdue 2008). The PCR products were sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific, Waltham, MA) and analysed using the 3500 Genetic Analyser (Thermo Fisher Scientific, Waltham, MA). Then, next-generation sequencing was performed using the Flongle adaptor (Oxford Nanopore Technologies, Oxford, England) to analyse the genetic background of the HPAIV isolates. The primers used to amplify all eight gene segments, namely polymerase basic 2, polymerase basic 1, polymerase acidic, HA, nucleoprotein, neuraminidase, matrix, and non-structural, of these isolates have been described previously (Hew et al. 2024a). Libraries for Oxford Nanopore were prepared with the NEB Ultra II End Repair/dA-Tailing Module (New England Biolabs, Ipswich, MA) and sequenced on the Flongle adaptor using the Ligation Sequencing Kit V14 (Oxford Nanopore Technologies, Oxford, England). The sequencing reads were then mapped and assembled using FluGAS version 2 (World Fusion, Tokyo, Japan).

Phylogenetic tree inferences of H5N1 HPAIV in Japan

The HA sequences of H5N1 HPAIV (n = 141) obtained by the Laboratory of Microbiology, Hokkaido University, were combined with other sequences acquired from Japan between 2021 and 2025 (n = 309) by the Global Initiative on Sharing All Influenza Data (GISAID). To identify genetically similar sequences from other countries, the Basic Local Alignment Search Tool in GISAID was used to compare the Japanese sequences with global HA sequences to identify the H5N1 HPAIV sequences with high similarity to the Japanese isolates for further analysis. All HA sequences were aligned using Multiple Alignment using Fast Fourier Transfor (MAFFT) v7.526 (Katoh and Standley 2013) and manually inspected and trimmed with AliView v1.30 (Larsson 2014). Sequences that were identical (100% sequence homology) and sequences shorter than 1700 base pairs were excluded. The HA gene sequences used in this study contained only the open reading frame region. Maximum likelihood phylogenies were inferred using IQ-Tree 2 (Minh et al. 2020) with the Hasegawa-Kishino-Yano model + Gamma-distributed rates with 4 categories (HKY + G4) nucleotide substitution model (Shapiro et al. 2006). These phylogenies were used to identify monophyletic Japanese clades and evaluate the temporal signal in TempEst v1.5.3 (Rambaut et al. 2016). For visualization and annotation of phylogenies, the R package ggtree v3.14.0 was used (Yu 2020).

Bayesian phylodynamic of H5N1 HPAIV in Japan

A Bayesian phylodynamic approach was used to infer the introduction and inter-species transmission dynamics of H5N1 HPAIV. The Multi-Type Birth-Death (MTBD) model was applied to the phylogenetic tree to explicitly account for structured populations representing different hosts or geographic regions, with separate transmission, sampling, and removal rates (Kühnert et al. 2016, Vaughan and Stadler 2025).

The MTBD model was chosen over simpler birth–death models and discrete phylogeographic approaches because it enables joint estimation of type-specific epidemiological parameters, including the effective reproductive number (Re), becoming-uninfectious rate, sampling proportion, and migration rates between demes. This allows for the inference of virus spread over time in a way that accurately reflects how infections actually occur (Barido-Sottani et al. 2020, Zhukova et al. 2023). Unlike discrete phylogeographic models, which estimate ancestral geographic states without linking to transmission dynamics, MTBD directly models migration as part of the epidemiological process and can adjust for uneven sampling across hosts or regions.

HA sequences were categorized into four epizootic periods (June 2021–May 2022, June 2022–May 2023, June 2023–May 2024, and June 2024–May 2025) and further organized by geographic origin and avian host species. Geographic demes included Europe, Japan, Northeast Asia (South Korea and China), North Asia (Russia), and Northwest America (Canada: British Columbia and Alberta; USA: Alaska). Host demes were grouped as anatidae (ducks, swans, geese), charadriiformes (gulls and curlews), poultry (chickens, ducks, emus, quails, guinea fowl), raptors (eagles, falcons, hawks), cranes, and crows; species with particularly low sample sizes in Japan were combined into a ‘Japan birds’ category to avoid overparameterization. Supplementary Table S1 provides the detailed sequence counts and subgroup distributions. The MTBD model was implemented in BEAST v2.6.3 with the Birth-Death-Migration-Model-Prime (BDMM-Prime) package, using the Broad-platform Evolutionary Analysis General Likelihood Evaluator (BEAGLE) library for computational efficiency (Ayres et al. 2012). Analyses employed an HKY + G4 nucleotide substitution model (Shapiro et al. 2006) and a relaxed lognormal molecular clock (Drummond et al. 2006). Priors for all key model parameters are summarized in Supplementary Tables S2–S5, including the becoming-uninfectious rate, which corresponds to a median infectious period of 7 days (Imperia et al. 2023), and migration rates, which are expected to average about two migration events per lineage per year (Jourdain et al. 2007).

To mitigate potential sampling bias caused by uneven surveillance efforts among hosts and regions (such as the overrepresentation of poultry outbreaks compared to wild bird detections), deme-specific sampling proportions were calculated within the MTBD framework. For each deme, the sampling proportion reflects the likelihood that an infection was represented in the dataset. It was estimated as the ratio of analysed sequences to the total number of reported cases in the Food and Agriculture Organization’s EMPRES Global Animal Disease Information System (Supplementary Tables S2–S5). This method partially compensates for uneven sampling coverage by using priors informed by relative surveillance effort and calibrated to avoid overfitting demes with small sample sizes. The MTBD model assumes that transmission, becoming-uninfectious, and sampling rates stay constant within each analysis period and that all relevant demes are included in the dataset; therefore, estimated parameters should be considered as time-averaged values across sampled populations (Barido-Sottani et al. 2020). Phylodynamic parameters and posterior trees were estimated using Markov chain Monte Carlo over 100 million steps, sampling every 10 000 iterations. The first 10% of samples were discarded as burn-in, and convergence was assessed in Tracer v1.7, ensuring all effective sample sizes exceeded 200 (Rambaut et al. 2018). Maximum clade credibility (MCC) trees were reconstructed with TreeAnnotator v1.10.4 (Suchard et al. 2018) and visualized in R v2.4.2 using the ggtree package (Yu 2020).

Results

Overview of H5N1 HPAIV epizootics in Japan 2021–25

During winter 2021–22, there were 22 poultry outbreaks in the first half of the season and 83 wild birds later in the season (Supplementary Fig. S1b). By the following winter, there was an increase with 136 wild bird cases and 88 poultry outbreaks, mostly early in the season and peaking in midwinter. During winter 2023–24, the wild bird cases remained high (n = 136), but only 10 poultry outbreaks occurred. In winter 2024–25, wild bird cases surged to 302, along with 47 poultry outbreaks, with wild bird cases showing a significant rise at the end of winter. Throughout this period, no infections were detected in Japan during the summer months (June–August).

Genetic analysis

To investigate potential links, an additional 47 European H5N1 HPAIV sequences from 2020 to 2021 were included in the analysis. Phylogenetic analysis of the H5N1 HPAIV HA genes revealed that Japanese isolates from winter 2021–22 were genetically distinct from European isolates collected in 2020–21 (Supplementary Fig. S2). Some Japanese isolates clustered with European viruses from 2021 to 2022, considered descendants of the 2020–21 European strains. These clusters also included viruses from Northwest America collected in 2021–22. Additionally, a subset of Japanese isolates from 2021 to 2022 clustered with viruses from North Asia and Northeast Asia. Similar clustering patterns were observed in subsequent winter seasons (2022–23, 2023–24, and 2024–25), with Japanese isolates remaining genetically linked to viruses from Northwest America, Northeast Asia, and North Asia, while no European H5N1 isolates were detected in Japan after 2022. This pattern suggests ongoing interregional migration events involving Northeast Asia, Northwest America, and North Asia in each winter season.

Introduction of H5N1 HPAIV into Japan 2021–23

During winter 2021–22, two independent introductions of H5N1 HPAIV into Japan were observed based on the MCC tree, which were identified as the G2b and G2d subgroups (Fig. 1a). The G2b subgroup was predominantly detected in southern Japan through sporadic outbreaks (Supplementary Fig. S1b) at the beginning of the epidemic and probably originated from Northeast Asia (e.g. South Korea and China) because the ancestral strains trace back to this region. Conversely, the G2d subgroup was predominantly detected in northern Japan, specifically in Hokkaido, in the middle of the epidemic. It was most likely introduced from Northwest America (such as the USA and Canada) based on the ancestral strains, leading to multiple outbreaks. For the migration per lineage per year, the mean migration rates indicated Northwest America {1.6, 95% highest posterior density [HPD] (0.7, 2.8)} and Northeast Asia [0.7, 95% HPD (0.3, 1.1)] as major potential sources for introduction into Japan. Conversely, North Asia [1.9, 95% HPD (0.4, 4.0)], Northwest America [2.2, 95% HPD (0.5, 4.7)], and Northeast Asia [2.5, 95% HPD (0.7, 5.0)] were inferred to be major potential destination regions from Japan (Fig. 1b). However, the number of inferred migration events showed that most of the viral movements to Japan occurred from Northwest America [30.2, 95% HPD (16, 43)], with fewer events originating in Northeast Asia [10.5, 95% HPD (6.0, 16.0)] (Fig. 1c). This suggests that while some routes exhibit high potential migration rates, they may have contributed to fewer introductions, possibly due to shorter lineage time spent in Japan or sparse sampling. Indeed, the viruses spent comparable amounts of time across the following regions: Northeast Asia [6.5, 95% HPD (5.4, 7.8)], followed by Europe [6.2, 95% HPD (4.8, 7.7)], and Northwest America [4.0, 95% HPD (2.4, 5.5)]. However, the viruses spent the least amount of time in Japan [1.5, 95% HPD (1.0, 2.1)] and North Asia [0.6, 95% HPD (0.3, 1.0)] (Fig. 1d).

Figure 1.

Figure 1

Open in a new tab

(a) The maximum clade credibility (MCC) tree was constructed using the hemagglutinin gene of the H5N1 high pathogenicity avian influenza virus (HPAIV) from clade 2.3.4.4b in Japan from June 2021 to May 2022; it was supplemented with other H5N1 HPAIV samples from different regions during the same period, represented by different colours for the internal and external nodes of the MCC tree; heatmaps showing (b) the average per-lineage virus migration rates and (c) the number of inferred virus migrations between different geographic regions in winter 2021–22; (d) the pie chart displays the percentage of time the viruses spent in each geographic region during winter 2021–22.

H5N1 HPAIV was reintroduced in winter 2022–23, and was categorized into three distinct subgroups (G2b, G2c, and G2d) based on the MCC tree. The G2c subgroup was primarily seen in southern Japan, likely originating from Northeast Asia. Viruses in the G2d subgroup were primarily detected in northern Japan, in Hokkaido Prefecture (Supplementary Fig. S1b), and were likely introduced from Northwest America. The G2b subgroup, which was detected in Hokkaido in autumn 2022, likely originated from North Asia (Fig. 2a). The mean migration rates indicated that North Asia [1.9, 95% HPD (0.7, 3.3)] and Northeast Asia [1.5, 95% HPD (0.5, 2.6)] were major potential sources for introduction into Japan, while Northeast Asia [1.7, 95% HPD (0.4, 5.1)] and North Asia [1.4, 95% HPD (0.3, 2.7)] were potential destination regions for viral lineages originating from Japan (Fig. 2b). Regarding the number of inferred migration events, a high number of viral movements from North Asia [21.7, 95% HPD (6.0, 38.0)], followed by Northeast Asia [10.8, 95% HPD (2.0, 2.0)] to Japan, was observed but with fever inferred events involving Northwest America [7.9, 95% HPD (2.0, 13.0)] (Fig. 2c). Based on the detection of multiple subgroups of the viruses in Japan, it can be deduced that the amount of time these viruses spend in Japan [11.7, 95% HPD (8.6, 15.0)] is comparable to that in other regions such as Northwest America [10.7, 95% HPD (8.8, 12.9)], followed by Northeast Asia [9.3, 95% HPD (7.1, 11.9)]. The viruses seem to spend the least amount of time in North Asia [5.7, 95% HPD (2.6, 8.9)] (Fig. 2d).

Figure 2.

Figure 2

Open in a new tab

(a) The maximum clade credibility (MCC) tree was constructed using the hemagglutinin gene of the H5N1 high pathogenicity avian influenza virus (HPAIV) from clade 2.3.4.4b in Japan from June 2022 to May 2023; it was supplemented with other H5N1 HPAIV samples from different regions during the same period, represented by different colours for the internal and external nodes of the MCC tree; heatmaps showing (b) the average per-lineage virus migration rates and (c) the number of inferred virus migrations between different geographic regions in winter 2022–23; (d) the pie chart displays the percentage of time viruses spent in each geographic region during winter 2022–23.

Shift in the geographic origins of H5N1 HPAIV during winters 2023–25

H5N1 HPAIVs from the G2c and G2d subgroups were reintroduced during the winters 2023–24 and 2024–25, with a shift in the geographic origin of the viruses in the G2d subgroup. While introductions in winters 2021–22 and 2022–23 were primarily linked to ancestral strains from Northwest America, those in the later winters were more likely to have originated from Northeast Asia (Figs 3a and 4a). In winter 2023–24, North Asia [1.6, 95% HPD (0.5, 3.2)], Northeast Asia [1.4, 95% HPD (0.7, 2.3)], and Northwest America [1.3, 95% HPD (0.4, 2.5)] had similar rates of virus introduction into Japan (Fig. 3b). Meanwhile, Northeast Asia [2.2, 95% HPD (0.4, 4.4)] was inferred as the major potential destination from Japan. Based on the number of inferred migration events, most of the viral movement was observed from Northeast Asia [65.8, 95% HPD (50.0, 80.0)] to Japan with fewer inferred events involving Northwest America [4.0, 95% HPD (0.0, 10.0)] and North Asia [2.1, 95% HPD (0.0, 6.0)] (Fig. 3c), as evidenced by the substantial duration of time these viruses spend in Northeast Asia [11.8, 95% HPD (8.8, 15.6)]. However, they spend the least amount of time in Japan [2.4, 95% HPD (1.4, 3.6)], followed by Northwest America [1.1, 95% HPD (0.3, 2.0)], and North Asia [0.7, 95% HPD (0.2, 1.3)] (Fig. 3d).

Figure 3.

Figure 3

Open in a new tab

(a) The maximum clade credibility (MCC) tree was constructed using the hemagglutinin gene of the H5N1 high pathogenicity avian influenza virus (HPAIV) from clade 2.3.4.4b in Japan from June 2023 to May 2024; it was supplemented with other H5N1 HPAIV samples from different regions during the same period, represented by different colours for the internal and external nodes of the MCC tree; heatmaps showing (b) the average per-lineage virus migration rates and (c) the number of inferred virus migrations between different geographic regions in winter 2023–24; (d) the pie chart displays the percentage of time the viruses spent in each geographic region during winter 2023–24.

Figure 4.

Figure 4

Open in a new tab

(a) The maximum clade credibility (MCC) tree was constructed using the hemagglutinin gene of the H5N1 high pathogenicity avian influenza virus (HPAIV) from clade 2.3.4.4b in Japan from June 2024 to May 2025; it was supplemented with other H5N1 HPAIV samples from different regions during the same period, represented by different colours for the internal and external nodes of the MCC tree; heatmaps showing (b) the average per-lineage virus migration rates and (c) the number of inferred virus migrations between different geographic regions in winter 2024–25; (d) the pie chart displays the percentage of time the viruses spent in each geographic region during winter 2024–25.

In winter 2024–25, introductions into Japan were again mainly from Northeast Asia [1.0, 95% HPD (0.5, 1.6)] and Northwest America [1.6, 95% HPD (0.4, 3.2)], with most inferred migration events traced to Northeast Asia [38.0, 95% HPD (29.0, 46.0)], and only limited inferred events involving the Northwest America [0.6, 95% HPD (0.0, 2.0)] (Fig. 4b and c). Viruses also spent considerably more time in Northeast Asia [8.3, 95% HPD (6.0, 10.9)] than in Japan [1.6, 95% HPD (1.0, 2.2)] or Northwest America [0.7, 95% HPD (0.3, 1.2)] (Fig. 4d), which suggest that Northeast Asia remained the dominant source and reservoir of viruses during these consecutive winters.

Inter-species transmission dynamics of H5N1 HPAIV in Japan during winters 2021–25

In winter 2021–22, the analyses focused on avian hosts within the G2d subgroup, including raptors, crows (Corvidae), anatidae (e.g. ducks and geese), and poultry. The mean Re within the host groups, which indicates the average number of secondary infections within a host group, exceeded 1 in poultry [3.1, 95% HPD (0.9, 5.7)] and crows [1.3, 95% HPD (0.1, 2.8)] (Fig. 5a), suggesting sustained transmission among these populations. Conversely, the mean Re between host groups, called inter-species transmission, was lower, indicating more limited inter-species spread. The highest mean Re was from raptors [1.8, 95% HPD (0.0, 6.4)] and anatidae [1.6, 95% HPD interval (0.0, 5.7)] to crows (Fig. 5b, Supplementary Fig. S3a). This is consistent with the number of inferred inter-species transmission events, which was the highest in raptors [8.1, 95% HPD (0.0, 29.0)], followed by anatidae [8.0, 95% HPD (0.0, 29.0)] to crows (Fig. 5c). In contrast, transmission from crows to other host groups was rare, with few inferred events to raptors, poultry, or anatidae. This directional asymmetry was also reflected in the MCC tree, where most crow isolates formed a monophyletic clade, indicating limited transmission to other species (Supplementary Fig. S3b). In winter 2022–23, the G2c and G2d subgroups were analysed separately. In the G2c subgroup, the mean Re within the host groups exceeded 1 only in the cranes [1.6, 95% HPD (1.0, 2.2)], while this value was below 1 in the other host groups (Fig. 6a). The mean Re between host groups remained consistently below 1, suggesting very limited inter-species transmission (Fig. 6b, Supplementary Fig. S4a). In the G2d subgroup, the mean Re values within and between host groups were also below 1 (Fig. 6d and e, Supplementary Fig. S5a), supported by the low number of inferred inter-species transmission events (Fig. 6f).

Figure 5.

Figure 5

Open in a new tab

(a) The ridge plot for each host type (x-axis) shows the density distribution of the mean effective reproductive number (Re) (x-axis) within the hosts during winter 2021–22; (b) the chord diagram displays transmission events between different bird species in Japan and other regions during the winter 2021–22, indicating the mean Re among host groups; the line thickness reflects the magnitude of the mean Re between hosts, and the arrows show the direction of host transmission; (c) the heatmap displays the inferred cross-species transmissions among hosts in Japan and between other regions during winter 2021–22.

Figure 6.

Figure 6

Open in a new tab

Ridge plots for each host type (x-axes) showing the density distribution of the mean effective reproductive number (Re) (x-axes) within hosts in the (a) G2c and (d) G2d subgroups during winter 2022–23; chord diagrams displaying the transmission events between different bird species in Japan in the (b) G2c and (e) G2d subgroups and other regions during winter 2022–23, indicating the mean Re among host groups; the line thickness reflects the magnitude of the mean Re between hosts, and the arrows show the direction of transmission; heatmaps denoting the inferred cross-species transmissions among hosts in Japan in the (c) G2c and (f) G2d subgroups, and between other regions during winter 2022–23.

In winter 2023–24, the mean Re within host groups in the G2d subgroup exceeded 1 in several bird species, with the highest value observed in crows [4.8, 95% HPD (1.0, 8.2)], followed by poultry [2.5, 95% HPD (0.1, 5.3)], cranes [2.0, 95% HPD (0.2, 4.0)], and anatidae [1.7, 95% HPD (0.1, 3.8)] (Fig. 7a), indicating ongoing within-group transmission. Although the mean Re between host groups remained below 1, it was closer to this threshold compared to that seen in previous winters, for instance, from raptors to crows [0.9, 95% HPD (0.0, 2.3)] and anatidae [0.9, 95% HPD (0.0, 2.4)] (Fig. 7b, Supplementary Fig. S6a). This pattern aligns with the number of transmission events estimated between those groups (Fig. 7c) and is also reflected in the MCC tree (Supplementary Fig. S6b). In winter 2024–25, the mean Re within host groups in the G2d subgroup exceeded 1 in several bird species, with the highest in cranes [9.5, 95% HPD (5.5, 14.1)], followed by crows [6.2, 95% HPD (0.2, 10.4)], poultry [2.1, 95% HPD (0.0, 15.1)], anatidae [1.1, 95% HPD (0.0, 3.2)], and charadriiformes [1.2, 95% HPD (0.0, 3.4)] (Fig. 8a). Meanwhile, the mean Re between host groups exceeded 1 from raptors to anatidae [1.2, 95% HPD (0.0, 3.4)] and charadriiformes [1.1, 95% HPD (0.0, 3.3)], as well from anatidae to raptors [1.0, 95% HPD (0.0, 2.8)] and charadriiformes [1.0, 95% HPD (0.0, 2.9)]. Interestingly, transmission from crows to charadriiformes [1.1, 95% HPD (0.0, 3.2)] also showed a mean Re exceeding 1 (Fig. 8b, Supplementary Fig. S7a). This pattern matches the estimated number of transmission events observed between these groups (Fig. 8c), as shown in the MCC tree (Supplementary Fig. S7b).

Figure 7.

Figure 7

Open in a new tab

(a) The ridge plot for each host type (x-axis) shows the density distribution of the mean effective reproductive number (Re) (x-axis) within hosts during winter 2023–24; (b) the chord diagram displays transmission events between different bird species in Japan and other regions during winter 2023–24, indicating the mean Re among host groups; the line thickness reflects the magnitude of the mean Re between hosts, and the arrows show the direction of host transmission; (c) the heatmap displays the inferred cross-species transmissions among hosts in Japan and between other regions during winter 2023–24.

Figure 8.

Figure 8

Open in a new tab

(a) The ridge plot for each host type (x-axis) shows the density distribution of the mean effective reproductive number (Re) (x-axis) within hosts during winter 2024–25; (b) the chord diagram displays transmission events between different bird species in Japan and other regions during winter 2024–25, indicating the mean Re among host groups; the line thickness reflects the magnitude of the mean Re between hosts, and the arrows show the direction of host transmission; (c) the heatmap displays the inferred cross-species transmissions among hosts in Japan and between other regions during winter 2024–25.

Discussion

This study examined the geographic origins and transmission patterns of H5N1 HPAIVs belonging to clade 2.3.4.4b across different bird species, which have been repeatedly introduced to Japan each winter from 2021 to 2025. The results showed that the HA gene sequences in Japan diverged into multiple subgroups, such as G2b, G2c, and G2d, as reported previously (Hew et al. 2024a). The G2b subgroup detected in winter 2021–22 likely originated from Northeast Asia. When it re-emerged in winter 2022–23, it was probably introduced from North Asia. H5N1 HPAIV in the G2c subgroup, first identified in Japan in winter 2022–23, was likely introduced from Northeast Asia, and reappeared in winter 2024–25, again originating from Northeast Asia, indicating consistent virus movement between Northeast Asia and Japan (Seo et al. 2024, Hew et al. 2024a, Si et al. 2025). The presence of G2b and G2c in Japan in winters 2021–25 further indicates that HPAIV spreads along the East Asian-Australasian flight route, primarily through wild migratory birds during autumn. Multiple phylogenetic studies support a primary west-to-east transmission pattern, with East Asian viruses being more closely related to those from Europe and Central Asia than to North American lineages (Lee et al. 2015, The Global Consortium for H5N8 and Related Influenza Viruses 2016, Meng et al. 2019).

In contrast, the G2d subgroup was first identified in Japan during winter 2021–22 and has been detected annually since then (Isoda et al. 2022, 2025, Hew et al. 2024a). This study shows that during winters 2021–23, common ancestral strains were found in North Asia, Northwest America, and Japan, as evidenced by frequent viral migration events among these regions. Notably, phylodynamic analyses in this study indicate that the ancestral strains of the G2d subgroup may have originated from viruses circulating in Northwest America in winters 2021–23, suggesting a trans-Beringian introduction of viruses from North America to Japan. Such a transmission route has historically been considered unlikely due to limited migratory bird overlap and sparse surveillance in critical regions such as Alaska and Siberia (Ito et al. 1995, Lee et al. 2015, The Global Consortium for H5N8 and Related Influenza Viruses 2016). However, a distinct cluster in the MCC tree for winter 2022–23 further supports this hypothesis by revealing North Asia as an intermediate region before the virus transmission into Japan. Despite the possibility of a direct introduction from Northwest America to Japan, the limited availability of genomic data from North Asia constrains definitive conclusions. Nevertheless, previous studies have shown genetic similarities between HPAIVs detected in Northwest America and those isolated in Japan, indicating possible transmission through migratory birds along the West Pacific flight path (Alkie et al. 2022). Additionally, a bidirectional viral flow might exist, as viruses detected in East Asia in autumn 2022 were closely related to those identified in North America, although current evidence is limited (Kang et al. 2023). From winter 2023–24, the likely geographical origin of the G2d subgroup appeared to have shifted, with evidence suggesting its Northeast Asian origin. This implies that the changes in transmission patterns for this group of viruses may result from ongoing viral migration events among North Asia, Northwest America, and Northeast Asia, with North Asia serving as a major viral hub, given its role as a key breeding ground for many migratory birds and its location at the intersection of multiple flight paths (Zeng et al. 2024). This may promote the spread of G2d viruses to Northeast Asia, as shown by the increased viral migration events from North Asia. Furthermore, the extended presence of the virus in Northeast Asia in winters 2023–25 suggests a potential local presence of G2d viruses in that region. These findings suggest that Japan could act as a hub for intercontinental virus introductions, as multiple virus entries from different regions, including North Asia, Northwest America, and Northeast Asia, have been detected, reflecting its strategic location along the East Asia–Australasian and West Pacific migratory flyways. This is particularly concerning given that, despite the seemingly distinct introduction routes for each subgroup (G2b, G2c, and G2d), their entry patterns appear to change across the years, reflecting the unpredictable nature of HPAIV spread and complicating control strategies.

This study revealed variations in the transmission patterns of H5N1 HPAIVs across different bird species (Fig. 9). Poultry showed consistent within-group Re during several seasons, notably during winters 2021–22 (Re = 3.1), 2023–24 (Re = 2.5), and 2024–25 (Re = 2.1), suggesting sustained intraspecies transmission. Farm-to-farm transmission of HPAIV is relatively rare in Japanese HPAIV outbreaks, according to the post-outbreak investigations by the national authorities. The high mean Re observed in poultry may result from duplicate sequence submissions from multiple laboratories in Japan for the same poultry cases. This might result in the reporting of multiple sequences for a single event, leading to an overestimation of the mean Re values within the poultry population. However, HPAI outbreaks occurred continuously at high rates in poultry farms during winter 2024–25, which theoretically matches the high mean Re in these birds. Overall, ‘between-group’ transmission from poultry to wild birds remained rare throughout, suggesting that poultry most likely act as recipients rather than drivers of inter-species transmission. Crows emerged as a key host group with divergent roles across the seasons. In the winters of 2021–22, 2023–24, and 2024–25, the mean Re within crows exceeded 1 (Re = 1.3, 4.8, and 6.2, respectively), indicating sustained intraspecies transmission. While onward transmission from crows to other species remained limited early on, an increasing transmission potential was observed in later seasons, including to charadriiformes in winter 2024–25. These findings suggest a shift in the role of crows from primarily dead-end hosts (as seen with G2c) to potential intermediate hosts (notably for G2d), possibly driven by ecological factors such as regional overwintering behaviour in Hokkaido and the environmental persistence of viruses. Nevertheless, crows may still function as opportunistic transmitters, as evidenced by the higher number of inferred events between crows and poultry in the winters of 2022–23 and 2024–25, based on the presence of dead crows near infected poultry farms (Ministry of Agriculture, Forestry and Fisheries, Japan 2023). Experimental studies also indicate variable susceptibility to different H5 HPAIV subgroups (Hiono et al. 2016), which warrants further investigation for the G2b, G2c, and G2d subgroups, as shown by the differences in pathogenicity in chickens across these subgroups (Hew et al. 2024a).

Figure 9.

Figure 9

Open in a new tab

Graphical abstract depicting the transmission events between hosts in (a) the G2c subgroup during winter 2022–23 and (b) the G2d subgroup during winters 2021–25, based on the mean effective reproductive number (Re); the arrow indicates the direction of transmission; the solid and dotted lines represent mean Re greater than or equal to 1.0 and Re <1.0, respectively.

Raptors significantly contribute to inter-species transmission, with the highest between-host Re observed in the earlier seasons, especially toward crows (Re = 1.8 in winter 2021–22; Re = 0.9 in winter 2023–24) and toward anatidae and charadriiformes (Re = 1.1 in winter 2024–25). Transmission from raptors to anatidae is generally considered uncommon because these two species occupy different ecological niches. Viruses are not easily detectable in waterfowl, which are considered to be asymptomatic carriers. In contrast, raptors are highly susceptible to HPAIV infection because they sometimes scavenge on infected birds, including waterfowl. This typically causes early death (Fujimoto et al. 2022), leading to earlier detection and reporting of their HPAIV infection. The difference in disease notifications after HPAIV infection among anatidae and raptors could also indirectly affect the interpretation of the disease dynamics. Furthermore, crows are also susceptible to HPAIV infection due to scavenging or habitat sharing with waterfowl. However, they are usually not assessed until their mass die-off, unlike that in the raptors. Although raptors are typically regarded as spillover hosts, acquiring infection through predation or scavenging infected birds, the results suggest that they may contribute to local transmission to other host groups before succumbing to the virus infection. While their limited mobility after infection restricts their long-distance spread (Hirschinger et al. 2025), their role in local amplification and transmission, especially to crows and other wild birds, warrants further study. Anatidae played a dual role across the seasons. Notably, the ‘within-group’ Re values were often above 1 (Re = 1.7 in winter 2023–24, Re = 1.1 in winter 2024–25), suggesting sustained transmission within wild waterfowl populations, these birds were recurrently identified as sources of cross-species transmission, especially to crows, raptors, cranes, and charadriiformes, consistent with their capacity for long-distance migration even when infected or asymptomatic (Lv et al. 2022). In contrast, cranes showed a progressive increase in epidemiological significance throughout the study period. In winter 2022–23, they were the only host group in the G2c subgroup with a mean Re above 1 (Re = 1.6). Their transmission potentials increased further in winter 2023–24 (Re = 2.0) and peaked in winter 2024–25 (Re = 9.5) within the G2d subgroup, likely because of virus introduction ‘via’ shared habitats with anatidae (Esaki et al. 2025b), which facilitated cross-species transmission. However, transmission from cranes to other wild birds remained rare, indicating that cranes are more often recipients than drivers of inter-species spread.

Although this study offers valuable insights into the geographic origins and transmission dynamics of H5N1 HPAIV in Japan, several limitations remain. Surveillance in Japan primarily depends on deceased birds or environmental samples rather than live-captured birds, which may overlook subclinical infections and asymptomatic carriers and restrict real-time understanding of virus circulation. Delays in data sharing and inconsistent surveillance efforts across countries further limit knowledge of the wider HPAIV situation. In phylodynamic analyses, sampling biases and limited sequence availability can impact accurate reconstruction of virus transmission. The MTBD model was appropriate for this study because it connects the phylogenetic tree with the underlying epidemiological processes of transmission, migration, and sampling among structured populations. MTBD enables estimation of specific transmission and migration rates, helping to distinguish repeated introductions from local persistence and partially accounting for uneven sampling across host types (Barido-Sottani et al. 2020). While discrete-trait phylogeographic approaches can describe lineage movements, they do not offer mechanistic insights into transmission dynamics (De Maio et al. 2015). However, the limitation of MTBD is that it assumes that transmission, removal, and migration rates are constant within each analysis period and is computationally demanding (Kühnert et al. 2016, Barido-Sottani et al. 2020). Therefore, the estimated rates should be viewed as averages over the period rather than exact figures. Nonetheless, the observed clustering and virus migration patterns among regions and hosts remained consistent across all analyses, supporting the robustness of the inferred transmission and migration dynamics (Figs 1–5, Supplementary Figs S3–S7). Consequently, ongoing efforts to improve global surveillance, including sampling live-captured wild birds and sharing data promptly, are essential for refining phylodynamic inferences and better understanding of H5 HPAIV transmission dynamics.

Supplementary Material

veag005_Supplemental_Figures_Tables
veag005_supplemental_figures_tables.zip (5.5MB, zip)

Acknowledgements

We thank Mayumi Endo and Fumihito Takaya for their technical support; the Japan Ministry of the Environment and Hokkaido Prefecture for their cooperation; and the authors and laboratories that submitted the sequences to GISAID EpiFlu database, which were used in this study. All GISAID data submitters can be reached directly through the GISAID website (https://www.gisaid.org). We also thank the Genotoul Bioinformatics Platform, Toulouse Occitanie (Bioinfo Genotoul, https://doi.org/10.15454/1.5572369328961167E12), for providing computing and storage resources.

Contributor Information

Yik Lim Hew, Laboratory of Microbiology, Department of Disease Control, Faculty of Veterinary Medicine, Hokkaido University, Kita 18, Nishi 9, Kita-Ku, Sapporo 060-0818, Hokkaido, Japan.

Claire Guinat, Université de Toulouse, École Nationale Vétérinaire de Toulouse (ENVT), Institut National de Recherche pour l'Agriculture, l'Alimentation et l'Environnement (INRAE), Interactions Hôtes-Agents Pathogènes (IHAP), 23 chemin des capelles, 31300, Toulouse, France.

Manon Couty, Université de Toulouse, École Nationale Vétérinaire de Toulouse (ENVT), Institut National de Recherche pour l'Agriculture, l'Alimentation et l'Environnement (INRAE), Interactions Hôtes-Agents Pathogènes (IHAP), 23 chemin des capelles, 31300, Toulouse, France.

Diletta Fornasiero, Université de Toulouse, École Nationale Vétérinaire de Toulouse (ENVT), Institut National de Recherche pour l'Agriculture, l'Alimentation et l'Environnement (INRAE), Interactions Hôtes-Agents Pathogènes (IHAP), 23 chemin des capelles, 31300, Toulouse, France; Istituto Zooprofilattico Sperimentale delle Venezie, Viale dell'Università 10, 35020 Legnaro (Padova), Italy.

Takahiro Hiono, Laboratory of Microbiology, Department of Disease Control, Faculty of Veterinary Medicine, Hokkaido University, Kita 18, Nishi 9, Kita-Ku, Sapporo 060-0818, Hokkaido, Japan; One Health Research Center, Hokkaido University, Kita 18, Nishi 9, Kita-tu, Sapporo 060-0818, Hokkaido, Japan; International Collaboration Unit, International Institute for Zoonosis Control, Hokkaido University, Kita 20, Nishi 10, Kita-Ku, Sapporo 001-0020, Hokkaido, Japan; Hokkaido University Institute for Vaccine Research and Development (HU-IVReD), Hokkaido University, Kita 21, Nishi 11, Kita-Ku, Sapporo 001-0021, Hokkaido, Japan.

Norikazu Isoda, Laboratory of Microbiology, Department of Disease Control, Faculty of Veterinary Medicine, Hokkaido University, Kita 18, Nishi 9, Kita-Ku, Sapporo 060-0818, Hokkaido, Japan; One Health Research Center, Hokkaido University, Kita 18, Nishi 9, Kita-tu, Sapporo 060-0818, Hokkaido, Japan; International Collaboration Unit, International Institute for Zoonosis Control, Hokkaido University, Kita 20, Nishi 10, Kita-Ku, Sapporo 001-0020, Hokkaido, Japan; Hokkaido University Institute for Vaccine Research and Development (HU-IVReD), Hokkaido University, Kita 21, Nishi 11, Kita-Ku, Sapporo 001-0021, Hokkaido, Japan.

Yoshihiro Sakoda, Laboratory of Microbiology, Department of Disease Control, Faculty of Veterinary Medicine, Hokkaido University, Kita 18, Nishi 9, Kita-Ku, Sapporo 060-0818, Hokkaido, Japan; One Health Research Center, Hokkaido University, Kita 18, Nishi 9, Kita-tu, Sapporo 060-0818, Hokkaido, Japan; International Collaboration Unit, International Institute for Zoonosis Control, Hokkaido University, Kita 20, Nishi 10, Kita-Ku, Sapporo 001-0020, Hokkaido, Japan; Hokkaido University Institute for Vaccine Research and Development (HU-IVReD), Hokkaido University, Kita 21, Nishi 11, Kita-Ku, Sapporo 001-0021, Hokkaido, Japan.

Conflict of interest

The authors declare no competing interests.

Funding

This work was supported by the Japan Agency for Medical Research and Development (grant no. JP223fa627005). It was also partially funded by the Japan International Cooperation Agency under the Science and Technology Research Partnership for Sustainable Development (grant no. JP23jm0110019). Additional support came from the Japan Science and Technology Agency’s Support for Pioneering Research Initiated by the Next Generation (grant no. JPMJSP2119), the regulatory research projects for food safety, animal health, and plant protection funded by the Ministry of Agriculture, Forestry, and Fisheries of Japan (grant no. JPJ008617.23812859), the Doctoral Program for World-Leading Innovative and Smart Education sponsored by the Japan Ministry of Education, Culture, Sports, Science and Technology, and the WISE Grant-in-Aid for graduate students through the Program for One Health Frontier at the Graduate School of Excellence, Hokkaido University (grant no. PH36210001). This research was performed by the Environment Research and Technology Development Fund (JPMEERF20254004) of the Environmental Restoration and The Conservation Agency is provided by the Ministry of the Environment of Japan.

Data availability

All gene sequences of the high pathogenicity avian influenza viruses used in this study are available on the GISAID database and provided in Supplementary Table S6.

Code availability

The XML files used for the discrete and continuous phylogeographic analyses are available from the GitHub repository (https://github.com/YLHew/MTBD-HPAIV_Japan.git).

References

  1. Alkie  TN, Lopes  S, Hisanaga  T  et al.  A threat from both sides: multiple introductions of genetically distinct H5 HPAI viruses into Canada via both East Asia-Australasia/Pacific and Atlantic flyways. Virus Evol  2022;8:veac077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ayres  DL, Darling  A, Zwickl  DJ  et al.  BEAGLE: an application programming interface and high-performance computing library for statistical phylogenetics. Syst Biol  2012;61:170–3. 10.1093/sysbio/syr100 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baek  Y-G, Lee  Y-N, Lee  D-H  et al.  Multiple reassortants of H5N8 clade 2.3.4.4b highly pathogenic avian influenza viruses detected in South Korea during the winter of 2020–2021. Viruses  2021;13:490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Banyard  AC, Bennison  A, Byrne  AMP  et al.  Detection and spread of high pathogenicity avian influenza virus H5N1 in the Antarctic region. Nat Commun  2024;15:7433. 10.1038/s41467-024-51490-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Barido-Sottani  J, Vaughan  TG, Stadler  T. A multitype birth–death model for Bayesian inference of lineage-specific birth and death rates. Syst Biol  2020;69:973–86. 10.1093/sysbio/syaa016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Carnegie  L, Raghwani  J, Fournié  G  et al.  Phylodynamic approaches to studying avian influenza virus. Avian Pathol  2023;52:289–308. [DOI] [PubMed] [Google Scholar]
  7. De Maio  N, Wu  C-H, O’Reilly  KM  et al.  New routes to phylogeography: a Bayesian structured coalescent approximation. PLoS Genet  2015;11:e1005421. 10.1371/journal.pgen.1005421 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Drummond  AJ, Ho  SYW, Phillips  MJ  et al.  Relaxed phylogenetics and dating with confidence. PLoS Biol  2006;4:e88. 10.1371/journal.pbio.0040088 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Esaki  M, Okuya  K, Tokorozaki  K  et al.  Highly pathogenic avian influenza A(H5N1) outbreak in endangered cranes, Izumi Plain, Japan, 2022–23. Emerg Infect Dis  2025a;31:937–47. 10.3201/eid3105.241410 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Esaki  M, Okuya  K, Tokorozaki  K  et al.  Surveillance of avian influenza viruses in the Izumi plain reveals the role of wild ducks in the introduction of H5N1 HPAIVs during the 2023/24 winter season. Comp Immunol Microbiol Infect Dis  2025b;123:102389. 10.1016/j.cimid.2025.102389 [DOI] [PubMed] [Google Scholar]
  11. European Food Safety Authority, European Centre for Disease Prevention, Control, European Union Reference Laboratory for Avian Influenza, Adlhoch  C, Fusaro  A  et al.  Avian influenza overview September – December 2021. EFSA J  2021;19:94. 10.2903/j.efsa.2021.7108 [DOI] [Google Scholar]
  12. Featherstone  LA, Zhang  JM, Vaughan  TG  et al.  Epidemiological inference from pathogen genomes: a review of phylodynamic models and applications. Virs Evol  2022;8:veac045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fujimoto  Y, Ogasawara  K, Isoda  N  et al.  Experimental and natural infections of white-tailed sea eagles (Haliaeetus albicilla) with high pathogenicity avian influenza virus of H5 subtype. Front Microbiol  2022;13:1007350. 10.3389/fmicb.2022.1007350 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Guinat  C, Vergne  T, Kocher  A  et al.  What can phylodynamics bring to animal health research?  Trends Ecol Evol  2021;36:837–47. 10.1016/j.tree.2021.04.013 [DOI] [PubMed] [Google Scholar]
  15. Hew  LY, Isoda  N, Takaya  F  et al.  Continuous introduction of H5 high pathogenicity avian influenza viruses in Hokkaido, Japan: characterization of viruses isolated in winter 2022–2023 and early winter 2023–2024. Transbound Emerg Dis  2024a;2024:1–18. 10.1155/2024/1199876 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hew  YL, Hiono  T, Monne  I  et al.  Cocirculation of genetically distinct highly pathogenic avian influenza H5N5 and H5N1 viruses in crows, Hokkaido, Japan. Emerg Infect Dis  2024b;30:1912–7. 10.3201/eid3009.240356 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hiono  T, Okamatsu  M, Yamamoto  N  et al.  Experimental infection of highly and low pathogenic avian influenza viruses to chickens, ducks, tree sparrows, jungle crows, and black rats for the evaluation of their roles in virus transmission. Vet Microbiol  2016;182:108–15. [DOI] [PubMed] [Google Scholar]
  18. Hirschinger  J, Höfle  U, Sánchez-Cano  A  et al.  Multidisciplinary tracking of highly pathogenic avian influenza A(H5N1) outbreak in griffon vultures, southern Europe, 2022. Emerg Infect Dis  2025;31:1589–99. 10.3201/eid3108.241456 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Imperia  E, Bazzani  L, Scarpa  F  et al.  Avian influenza: could the H5N1 virus be a potential next threat?  Microbiol Res  2023;14:635–45. [Google Scholar]
  20. Isoda  N, Onuma  M, Hiono  T  et al.  Detection of new H5N1 high pathogenicity avian influenza viruses in winter 2021–2022 in the Far East, which are genetically close to those in Europe. Viruses  2022;14:2168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Isoda  N, Hiono  T, Hew  YL  et al.  Dynamics of high pathogenicity avian influenza virus infection with multiple introductions in a crow flock in an urban park in Hokkaido, Japan. Comp Immunol Microbiol Infect Dis  2025;121:102367. [DOI] [PubMed] [Google Scholar]
  22. Ito  T, Okazaki  K, Kawaoka  Y  et al.  Perpetuation of influenza A viruses in Alaskan waterfowl reservoirs. Arch Virol  1995;140:1163–72. [DOI] [PubMed] [Google Scholar]
  23. Jourdain  E, Gauthier-Clerc  M, Bicout  D  et al.  Bird migration routes and risk for pathogen dispersion into western Mediterranean wetlands. Emerg Infect Dis  2007;13:365–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kandeil  A, Patton  C, Jones  JC  et al.  Rapid evolution of A(H5N1) influenza viruses after intercontinental spread to North America. Nat Commun  2023;14:3082. 10.1038/s41467-023-38415-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kang  Y-M, Heo  G-B, An  S-H  et al.  Introduction of multiple novel high pathogenicity avian influenza (H5N1) virus of clade 2.3.4.4b into South Korea in 2022. Transbound Emerg Dis  2023;2023:1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Katoh  K, Standley  DM. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol  2013;30:772–80. 10.1093/molbev/mst010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kühnert  D, Stadler  T, Vaughan  TG  et al.  Phylodynamics with migration: a computational framework to quantify population structure from genomic data. Mol Biol Evol  2016;33:2102–16. 10.1093/molbev/msw064 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Larsson  A. AliView: a fast and lightweight alignment viewer and editor for large datasets. Bioinformatics  2014;30:3276–8. 10.1093/bioinformatics/btu531 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lee  D-H, Torchetti  MK, Winker  K  et al.  Intercontinental spread of Asian-origin H5N8 to North America through Beringia by migratory birds. J Virol  2015;89:6521–4. 10.1128/JVI.00728-15 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lee  S, Cho  AY, Kim  T  et al.  Novel highly pathogenic avian influenza A(H5N1) clade 2.3.4.4b virus in wild birds, South Korea. Emerg Infect Dis  2023;29:1475–8. 10.3201/eid2907.221893 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lv  X, Li  X, Sun  H  et al.  Highly pathogenic avian influenza A(H5N8) clade 2.3.4.4b viruses in satellite-tracked wild ducks, Ningxia, China, 2020. Emerg Infect Dis  2022;28:1039–42. 10.3201/eid2805.211580 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Meng  W, Yang  Q, Vrancken  B  et al.  New evidence for the east–west spread of the highly pathogenic avian influenza H5N1 virus between central Asian and east Asian-Australasian flyways in China. Emerg Microbes Infect  2019;8:823–6. 10.1080/22221751.2019.1623719 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Minh  BQ, Schmidt  HA, Chernomor  O  et al.  IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol Biol Evol  2020;37:1530–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ministry of Agriculture, Forestry and Fisheries, Japan . Livestock Infectious Disease Surveillance Monthly Report. Tokyo, Japan: Ministry of Agriculture, Forestry and Fisheries, 2023.
  35. Perdue  ML. Molecular determinants of pathogenicity for avian influenza viruses. In: Avian Influenza. Ames, IA, USA: Blackwell Publishing, 2008, 23–41. 10.1002/9780813818634.ch2 [DOI] [Google Scholar]
  36. Pohlmann  A, King  J, Fusaro  A  et al.  Has epizootic become enzootic? Evidence for a fundamental change in the infection dynamics of highly pathogenic avian influenza in Europe, 2021. mBio  2022;13:e00609–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rambaut  A, Lam  TT, Max Carvalho  L  et al.  Exploring the temporal structure of heterochronous sequences using TempEst (formerly Path-O-Gen). Virus Evol  2016;2:vew007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Rambaut  A, Drummond  AJ, Xie  D  et al.  Posterior summarization in Bayesian phylogenetics using tracer 1.7. Syst Biol  2018;67:901–4. 10.1093/sysbio/syy032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ruiz-Saenz  J, Martinez-Gutierrez  M, Pujol  FH. Multiple introductions of highly pathogenic avian influenza H5N1 clade 2.3.4.4b into South America. Travel Med Infect Dis  2023;53:102591. [DOI] [PubMed] [Google Scholar]
  40. Seo  Y-R, Cho  AY, Si  Y-J  et al.  Evolution and spread of highly pathogenic avian influenza A(H5N1) clade 2.3.4.4b virus in wild birds, South Korea, 2022–2023. Emerg Infect Dis  2024;30:299–309. 10.3201/eid3002.231274 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Seo  Y-R, Cho  AY, Kim  D-J  et al.  Transmission dynamics of highly pathogenic avian influenza A(H5N1) and A(H5N6) viruses in wild birds, South Korea, 2023–2024. Emerg Infect Dis  2025;31:1561–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Shapiro  B, Rambaut  A, Drummond  AJ. Choosing appropriate substitution models for the phylogenetic analysis of protein-coding sequences. Mol Biol Evol  2006;23:7–9. [DOI] [PubMed] [Google Scholar]
  43. Si  Y-J, Kim  D-J, Lee  S-H  et al.  New incursions of H5N1 clade 2.3.4.4b highly pathogenic avian influenza viruses in wild birds, South Korea, October 2024. Front Vet Sci  2025;11:1526118. 10.3389/fvets.2024.1526118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Signore  AV, Giacinti  J, Jones  MEB  et al.  Spatiotemporal reconstruction of the North American A(H5N1) outbreak reveals successive lineage replacements by descendant reassortants. Sci Adv  2025;11:eadu4909. 10.1126/sciadv.adu4909 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Suchard  MA, Lemey  P, Baele  G  et al.  Bayesian phylogenetic and phylodynamic data integration using BEAST 1.10. Virus Evol  2018;4:vey016. 10.1093/ve/vey016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Takadate  Y, Tsunekuni  R, Kumagai  A  et al.  Different infectivity and transmissibility of H5N8 and H5N1 high pathogenicity avian influenza viruses isolated from chickens in Japan in the 2021/2022 season. Viruses  2023;15:265. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Takadate  Y, Mine  J, Tsunekuni  R  et al.  Genetic diversity of H5N1 and H5N2 high pathogenicity avian influenza viruses isolated from poultry in Japan during the winter of 2022–2023. Virus Res  2024;347:199425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. The Global Consortium for H5N8 and Related Influenza Viruses . Role for migratory wild birds in the global spread of avian influenza H5N8. Science  2016;354:213–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Vaughan  TG, Stadler  T. Bayesian phylodynamic inference of multitype population trajectories using genomic data. Mol Biol Evol  2025;42:msaf130. 10.1093/molbev/msaf130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. WHO/OIE/FAO H5N1 Evolution Working Group . Toward a unified nomenclature system for highly pathogenic avian influenza virus (H5N1). Emerg Infect Dis  2008;14:e1–1. 10.3201/eid1407.071681 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Yamamoto  N, Sakoda  Y, Motoshima  M  et al.  Characterization of a non-pathogenic H5N1 influenza virus isolated from a migratory duck flying from Siberia in Hokkaido, Japan, in October 2009. Virol J  2011;8:65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Yang  J, Zhang  C, Yuan  Y  et al.  Novel avian influenza virus (H5N1) clade 2.3.4.4b reassortants in migratory birds, China. Emerg Infect Dis  2023;29:1244–9. 10.3201/eid2906.221723 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Yu  G. Using ggtree to visualize data on tree-like structures. Curr Protoc Bioinformatics  2020;69:e96. 10.1002/cpbi.96 [DOI] [PubMed] [Google Scholar]
  54. Zeng  J, Du  F, Xiao  L  et al.  Spatiotemporal genotype replacement of H5N8 avian influenza viruses contributed to H5N1 emergence in 2021/2022 panzootic. J Virol  2024;98:e01401–23. 10.1128/jvi.01401-23 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Zhukova  A, Hecht  F, Maday  Y  et al.  Fast and accurate maximum-likelihood estimation of multi-type birth–death epidemiological models from phylogenetic trees. Syst Biol  2023;72:1387–402. 10.1093/sysbio/syad059 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

veag005_Supplemental_Figures_Tables
veag005_supplemental_figures_tables.zip (5.5MB, zip)

Data Availability Statement

All gene sequences of the high pathogenicity avian influenza viruses used in this study are available on the GISAID database and provided in Supplementary Table S6.

Articles from Virus Evolution are provided here courtesy of Oxford University Press

RESOURCES