The scale of the challenge is staggering. More than 17 million SARS-CoV-2 genome sequences have been publicly shared since the pandemic began, a figure that dwarfs anything virologists had previously contemplated. Before 2020, few researchers would have predicted that a single virus would accumulate even 100,000 genomes. When Rambaut and colleagues proposed the Pango system early in the pandemic, they anticipated that roughly 200 active lineages would suffice, warning that more would obscure rather than clarify discussion. In practice, between 200 and 1,000 distinct lineages circulated in any given month for much of the pandemic. That the system survived this explosion of data owes much to community curation and to tools such as Pangolin, Nextclade, UShER and Taxonium, which made assigning new genomes to lineages a routine computational task.

The core logic of a Pango lineage system is deceptively simple. A lineage is defined as a set of virus sequences forming an epidemiologically relevant cluster on a phylogenetic tree—specifically, all sequences descending from the most recent common ancestor of a curated set of designated reference genomes. The framework distinguishes between designation, in which high-quality sequences are formally assigned to define a lineage, and assignment, in which new genomes are classified against those definitions using phylogenetic placement, defining mutations or machine-learning classifiers. Designation is a deliberate, curated act; assignment is inferential and carries some uncertainty, which the authors argue must be acknowledged when results are interpreted, particularly for lower-quality or incomplete sequences.

What makes a cluster worthy of a name? The new framework lays out a set of signals that can justify designation, tailored to the biology and public health context of each virus. These include introduction into a new region with evidence of onward transmission, a change in phenotype, a notable constellation of mutations with possible functional relevance, recombination or reassortment, or evidence of rapid growth relative to other lineages. For SARS-CoV-2, early lineage definitions leaned heavily on geographical clustering while travel restrictions limited human movement; as global travel resumed, the criteria shifted toward growth advantage, recombination and mutations associated with immune escape. Any new lineage must also be genetically distinct from its ancestor by one or more mutations, show evidence of onward community transmission, and be supported by enough high-quality genomes to evaluate these criteria—though the exact thresholds are deliberately left flexible, to be set per virus and revised as circumstances change.

The architecture of lineage names encodes ancestry directly. Founding lineages are labelled with single letters at an initial time point chosen to capture the diversity relevant to public health; descendant sublineages append dot-separated numbers, so the 75th sublineage of BA.2 becomes BA.2.75. To keep names readable as diversity accumulates, an alias is introduced at a defined hierarchical level—after three levels for SARS-CoV-2, after two for mpox virus—replacing the prefix with the next available letter while the full ancestry is preserved in metadata. Circulating recombinants receive X-prefixed codes, as with SARS-CoV-2 lineage XBB, a recombinant of BJ.1 and BA.2.75, and designation can proceed even when the parental lineages remain uncertain. For reassorting viruses with segmented genomes, the system can be anchored to a single representative segment, typically the fastest-evolving one most relevant to transmission or antigenicity, with reassortment events in other segments capable of triggering new designations.

Critically, the authors stress that Pango is not a universal taxonomy and should not replace existing schemes. The International Committee on Taxonomy of Viruses classifies viruses down to the species rank, and below that level the field has long used a patchwork of terms—type, subtype, genotype, group and clade—each with its own conventions. A Pango system occupies an intermediate granularity: finer than persistent clade nomenclatures suited to long-term evolutionary patterns such as antigenic turnover, and coarser than phylotypes, a dot-separated notation encoding every node in a phylogeny for fine-scale local investigations such as hospital outbreaks. For endemic pathogens whose diversity is deep and largely historic, the authors caution that a fine-scale dynamic system may add complexity without public health benefit. Yet it becomes valuable when new selective pressures are expected—following vaccine or antiviral introduction, for instance—or when eradication programmes demand close tracking of contemporary transmission, as with poliovirus or measles virus.

The framework is also explicit about when it should not be used at all. Pathogens must be ‘measurably evolving’—accumulating detectable genetic diversity over the timescale of interest—for a lineage system to be informative. Many RNA viruses meet this bar, but double-stranded DNA viruses generally evolve too slowly unless host-mediated processes such as APOBEC3 editing elevate their mutation rates, as has occurred in mpox virus. Zoonotic outbreaks driven by repeated independent spillovers, such as the 2018 Lassa fever outbreak in Nigeria or re-emergences of yellow fever virus in Brazil, do not warrant lineage systems, nor do self-limiting, geographically confined transmission chains. Dynamic nomenclatures are most valuable for sustained human-to-human outbreaks spreading across multiple regions, exemplified by the global clade IIb mpox outbreak of 2022, where Pango-style lineages complemented existing clade labels. Independent spillovers giving rise to separate sustained outbreaks would each warrant their own lineage system.

To demonstrate the framework on a virus with a long history of human circulation, the team built a Pango lineage system for chikungunya virus, a mosquito-borne alphavirus first identified in 1952 and since reported in 114 countries. Working from roughly 3,300 publicly available near-full-length genomes, they filtered the data to sequences sampled from 2016 onwards and designated 21 lineages across eight founding lineages, labelled A through H. The conventional three-clade geography-based system maps onto this scheme: the West African clade corresponds to lineages A and B, the Asian clade to lineage C, and the East/Central/South African clade to lineages D through H, with twelve hierarchical sublineages defined across the founding lineages. The authors argue that geography-based labels have become misleading as phylogeographic patterns shift, and that naming conventions for chikungunya have lacked standardization. Sequences predating 2016 fall outside the system under the designation O, but can be incorporated as new founding lineages if they reappear in contemporary transmission—a flexibility illustrated by a 2021 mosquito sample from China that falls outside all current lineages.

Lessons from the pandemic loom over the entire enterprise. Pango labels proved poorly suited to public communication, which is why the World Health Organization introduced Greek-letter names for variants of concern such as Alpha and Omicron; yet after Omicron no further Greek letters were assigned, and the WHO has continued to rely on Pango lineages for tracking variants of interest and concern. The pandemic also revealed structural vulnerabilities. Global sequencing capacity remains deeply unequal, with high-income countries overrepresented in public databases while epidemiologically important regions contribute few sequences. The authors argue that designation systems must compensate by applying stricter thresholds in well-sampled countries and giving greater weight to data from underrepresented regions, and that the people closest to the data must be empowered to propose and flag new lineages through transparent public forums such as GitHub.

Perhaps the most sobering recommendation concerns sustainability. The Pango system was conceived as dynamic and time-limited, expected to be most useful for the duration of the pandemic, and the authors argue that such systems should exist only as long as they retain clear utility and support. When an epidemic pathogen transitions to endemicity, a more stable, persistent nomenclature with lower maintenance burden may serve better. Yet no organization currently exists to fund the establishment and maintenance of lineage systems, even as public health agencies, universities and industry all benefit from them. The authors propose a model that leverages distributed expertise across public health laboratories and academia, coordinated through central governance and modelled on established surveillance networks. As viral genomic data continue to shape outbreak control, they conclude, coherent and accessible nomenclature will be essential for translating complex genomic information into public health action—and the principles of utility, scalability, transparency and equity laid out here offer a foundation for building it.

Subject of Research: Generalization of the Pango viral lineage nomenclature system for genomic epidemiology and surveillance

Article Title: A generalized lineage nomenclature for viral genomic epidemiology

Article References: Colquhoun, R., Hinrichs, A. S., Hill, V., Ruis, C., Roemer, C., Pybus, O. G., Rambaut, A., & O’Toole, Á. (2026). A generalized lineage nomenclature for viral genomic epidemiology. Nature Microbiology. https://doi.org/10.1038/s41564-026-02468-3

Image Credits: AI Generated

DOI: 10.1038/s41564-026-02468-3

Keywords: Pango lineages, viral genomics, genomic epidemiology, SARS-CoV-2, chikungunya virus, phylogenetics, nomenclature, outbreak surveillance, mpox virus, viral evolution, public health, recombination