Skip to main content
Dryad

Ecogeographic drivers of the spatial spread of highly pathogenic avian influenza outbreaks in Europe and North America, 2016–2022

Cite this dataset

Gass, Jonathon et al. (2023). Ecogeographic drivers of the spatial spread of highly pathogenic avian influenza outbreaks in Europe and North America, 2016–2022 [Dataset]. Dryad. https://doi.org/10.5061/dryad.4qrfj6qdk

Abstract

H5Nx highly pathogenic avian influenza (HPAI) viruses of clade 2.3.4.4 have caused outbreaks in Europe among wild and domestic birds since 2016 and were introduced to North America via wild migratory birds in December 2021. We examined the spatiotemporal extent of HPAI viruses across continents and characterize ecological and environmental predictors of virus spread between geographic regions through constructing a Bayesian phylodynamic generalized linear model (phylodynamic-GLM). Findings demonstrate localized epidemics of H5Nx throughout Europe in the first several years of the epizootic, followed by a singular branching point where H5N1 viruses were introduced to North America, likely via stopover locations throughout the North Atlantic. Once in the US, H5Nx viruses spread at a greater rate between US-based regions and no evidence demonstrated spread back to any European region. We establish that geographic proximity is a predictor of virus spread between regions, which implies that inter-continental transport across the Atlantic Ocean is relatively rare. Increase in mean ambient temperature over time was predictive of reduced H5Nx virus spread, which may reflect the effect of climate change on declines in host species abundance, decreased persistence of the virus in the environment, or changes in migratory patterns due to ecological alterations. Our data provide new knowledge about the spread and directionality of H5Nx virus dispersal in Europe and North America during an actively evolving inter-continental outbreak, including predictors of virus movement between regions, which will contribute to surveillance and mitigation strategies as the outbreak unfolds, and in future instances of uncontained avian spread of HPAI viruses.

Methods

Publicly available H5Nx sequence data and phylogenetic analysis

All publicly available avian-derived (domestic and wild) H5Nx HA segment sequences from Europe and North America between 2016 and 2022 were downloaded from the Influenza Research Database (IRD) on May 12, 2022 (n=321). We added 170 publicly available H5Nx HA sequences from 2021–2022 downloaded from GISAID on May 15, 2022, as these were unavailable on IRD at the time of sequence acquisition, and 15 unpublished H5N1 HA sequences from avian surveillance in Massachusetts, USA by our research group in 2022 (described elsewhere), totaling 546 HA sequences. Metadata for each sequence was collected, including sampling date, season, host species, and geographic sampling location. Only IAV sequences from wild avian species or environmental matrices were included. Duplicate sequences, sequences with less than 75% unambiguous bases, all vaccine derivative and recombinant sequences, and sequences with unavailable isolation date, location, or host species were excluded, resulting in 506 sequences. Downsampling was performed to ensure relative evenness of geographic state groupings while preserving genetic diversity of the dataset, using geographic state and year for random stratification. To root and historically time-calibrate the tree, H5 subtype HA avian sequences from IRD were downloaded for the period 1979–2015 from Europe and North America and randomly downsampled by year, resulting in 33 historic sequences. These sequences were ‘masked’ to ensure their contribution to the tree structure but not to quantification of diffusion rates or the GLM. The total downsampled dataset, including the outgroup (GISAID sequences from North America (n=170), unpublished Massachusetts sequences acquired by our group (n=15), publicly available H5Nx sequences from Europe 2016-2022 (n=162)), and historic sequences (n=33) resulted in a total of 380 sequences (Supplementary table 1). Multiple sequence alignments were performed using MUSCLE in Geneious Prime 2022.05.14 (https://www.geneious.com) and trimmed to the open reading frame.

Usage notes

Notepad++, BEAST v.1.10.4, RStudio

Funding

National Institute of Allergy and Infectious Diseases, Award: HHSN272201400008C