H3Nx influenza A viruses circulate across multiple host species, including birds, humans, swine, equine, and canine populations. Despite frequent host switching by influenza viruses, the mechanisms that drive host-specific adaptation versus maintenance in reservoir hosts remain incompletely understood. This study used a large genomic dataset to quantify how evolutionary processes differ between reservoir (avian) and spillover (mammalian) hosts, testing whether reassortment or directional selection predominates in different host types.
The authors assembled a dataset of 13,295 H3Nx virus sequences. They developed a computational pipeline to (1) quantify host-specific adaptive evolution and (2) map reassortment events onto phylogenetic trees while incorporating measures of statistical uncertainty. The pipeline enabled comparison of reassortment rates, persistence of reassortant lineages, and signals of directional selection across host groups. The authors also made code and supporting files available through their GitHub repository.
Analyses revealed clear evidence of adaptive evolution in mammalian H3Nx lineages, particularly in the surface glycoprotein genes HA (hemagglutinin) and NA (neuraminidase). Mammalian viruses—across humans, swine, equine, and canine hosts—showed directional selection in these segments, consistent with ongoing adaptation after establishment in mammalian hosts. The study reports that adaptive evolution is a characteristic feature of H3Nx evolution once viruses circulate in mammalian populations.
By contrast, avian H3Nx lineages displayed very little evidence of directional selection. Instead, evolution in birds was characterized by frequent reassortment among genome segments. Avian viruses generated many novel reassortant lineages, reflecting a reassortment-dominant evolutionary strategy within the avian reservoir rather than selection-driven change in individual segments.
A substantial fraction of avian reassortant lineages are short-lived. The authors report that between 29.8% and 47.4% of all avian reassortant lineages are purged within the first year of circulation. Reassortment in birds did not show a detectable fitness benefit at the lineage level, given the high rate of rapid turnover and lack of directional-selection signatures. These dynamics suggest that while reassortment frequently generates diversity in birds, most reassortant combinations do not establish for the long term in the avian population.
Reassortment patterns differed by segment and host type. The study found that NA reassorted more frequently than expected in avian viruses but reassorted less frequently than expected in swine. Such segment-specific differences imply that selective constraints or ecological factors acting on particular segments vary by host, shaping which segment combinations are generated and retained.
Reassortment events were enriched in the context of mammalian host switches but not enriched for avian host switches. This enrichment suggests that reassortment may be particularly beneficial for mediating transitions between mammalian hosts. In contrast, avian-to-avian or avian host transitions did not show the same association with reassortment, reinforcing the view that reassortment serves different evolutionary roles across host types.
Collectively, the results support a model in which host ecology and biology govern influenza evolutionary strategy. In the avian reservoir, H3Nx evolution is dominated by frequent reassortment, producing many transient lineage variants with rapid turnover. Upon spillover and subsequent circulation in mammalian hosts, evolution shifts toward measurable adaptive evolution in key surface proteins (HA and NA), with reassortment playing a different or more targeted role—particularly in facilitating mammalian host switches and in some hosts such as swine where reassortants tend to persist.
The authors note that code and data supporting the analyses are available via the linked GitHub repository. Funding sources declared include Pew Charitable Trusts, the Margaret Q. Landenberger Research Foundation, the National Institute of Allergy and Infectious Diseases (NIAID/NIH), and the United States Department of Agriculture Agricultural Research Service. The authors declared no competing interests in the source.
All facts above are taken directly from the source report. Specific methodological details, including exact statistical thresholds, model parameters, or full pipeline implementation steps, were not exhaustively reported in the source text summarised here; readers should consult the authors’ repository and the full preprint for full technical details and supplementary materials.