Classification and Nomenclature of Viruses

Why a classification system is needed

The diversity of viruses is staggering. Several thousand viruses have been formally recognised, several hundred of which cause disease in humans, and that human fraction is tiny against the background of viruses in the wider environment. High-throughput metagenomic sequencing has since revealed a far larger virosphere than classical methods ever saw, and the catalogue is still growing fast. Viruses are probably the most abundant biological entities on the planet, infecting animals, plants, fungi, protozoa and bacteria alike. A classification system has to impose clear boundaries on what can otherwise look like a continuum of properties, and it has to be flexible enough to absorb a constant stream of newly discovered agents.

Because all viruses share the same fundamental nature regardless of host, virologists use a single universal system, overseen by the International Committee on Taxonomy of Viruses (ICTV). One conceptual point matters more than any other here: viral taxonomy is hierarchical and at most levels reflects evolutionary relationships, but it is deliberately non-systematic. There is no attempt to trace all viruses back to one ancestral root, because the evidence points to several independent origins. Viruses are polyphyletic. This is the key difference from the single tree of cellular life, and it explains why the upper ranks of viral taxonomy are sparse and cautious.

The taxonomic hierarchy

For most of its history the ICTV used a five-rank hierarchy: order, family, subfamily (used rarely), genus and species. That structure was in place from 1991 to 2017 and is the structure most references still present. It also remains the part of the hierarchy a clinician meets daily, because almost all of the conversation in medical virology happens at the family, genus and species level.

In two ratification votes in 2018 and 2019, the ICTV expanded the system into a 15-rank, Linnaean-like hierarchy that can express relationships at every scale, from the narrowest (species) to the very deepest (realm). There are now eight principal ranks and seven derivative sub-ranks, each with a defined suffix:

Principal rank	Suffix	Derivative rank	Suffix
Realm	-viria	Subrealm	-vira
Kingdom	-virae	Subkingdom	-virites
Phylum	-viricota	Subphylum	-viricotina
Class	-viricetes	Subclass	-viricetidae
Order	-virales	Suborder	-virineae
Family	-viridae	Subfamily	-virinae
Genus	-virus	Subgenus	-virus
Species	binomial (see below)	(none)	(none)

Four of the eight principal ranks (order, family, genus, species) carried over from the old scheme; four are new and all sit above order: realm, kingdom, phylum and class. The basal rank is deliberately called the realm, not “domain”: a realm expresses shared genes and proteins among viruses that may have arisen from quite different ancestors, rather than the single-rooted descent that “domain” implies for cellular life. The species rank has no “subspecies” derivative, because no agreed definition of a viral subspecies exists.

The new taxonomic rank “class” is not the same thing as a Baltimore class. They are unrelated concepts that happen to share a word. The taxonomic class is one rung of this hierarchy; a Baltimore class is a genome-and-replication grouping (covered below).

What the ranks group, from the bottom up: a genus brings together viruses with clear evolutionary and biological relationships, usually mirrored in antigenic, host-range and epidemiological features. (Each genus once had a designated type species, but the ICTV abolished that notion in 2021 as a historical artefact that no longer served a clear purpose.) Families contain viruses with broadly similar genome structure, virion morphology and replication strategy. The higher ranks, from order up to realm, express progressively deeper and more distant evolutionary relationships.

The realm is the widest grouping of all, defined by a deeply conserved hallmark (a shared capsid fold, replication enzyme or genome strategy) rather than by descent from one ancestor. Six realms were established in the first wave of 2018–2019, and the framework has kept growing as new lineages are recognised: as of Master Species List 41 (ratified February 2026) there are ten. The complete current list is maintained in the ICTV taxonomy database; the long-established, well-characterised realms include:

• Riboviria: RNA viruses (with the reverse-transcribing viruses) replicating via an RNA-directed polymerase;
• Duplodnaviria: double-stranded DNA viruses sharing the HK97-fold major capsid protein;
• Varidnaviria: DNA viruses sharing the jelly-roll capsid fold;
• Adnaviria: filamentous archaeal viruses with A-form DNA;
• Ribozyviria: small circular RNA agents with self-cleaving ribozymes, the realm that includes hepatitis delta virus.

Because viruses are polyphyletic, this is not one tree but several, each rooted in its own realm.

In practice, only genus and species assignment is obligatory. The higher ranks are populated only where the evidence justifies them, so most viruses are not yet placed at all 15 levels, and the tally is heavily weighted toward the species end: a 2020 count recorded 6 realms, 10 kingdoms, 17 phyla, 39 classes, 59 orders, 189 families, around 2,200 genera and over 9,100 species. Treat any such figure as a snapshot of a fast-moving target and consult the current ICTV release (the Master Species List is republished roughly every year) for up-to-date numbers.

And it is genuinely a moving target, in more than its membership: the rank framework itself is still being revised. The six realms of 2019 had become ten by Master Species List 41 (February 2026), and the number of recognised species has roughly doubled over the same span, past 17,500. Some of that change is structural rather than additive. In 2025, for example, the realm Monodnaviria proved not to be monophyletic and was broken up: it was renamed Floreoviria for the one kingdom it kept, while the other three moved out to form the new realms Efunaviria, Volvereviria and Pleomoviria. The live ICTV taxonomy browser at ictv.global/taxonomy is the authoritative current picture (the complete realm list, the latest counts, and any newly ratified ranks), updated after each ratification vote. No static snapshot can stay ahead of it, so anything load-bearing is best verified against the live database.

Each virus’s realm, family and genus assignment lives in its individual profile in the A–Z index.

The criteria used to classify a virus

In the era before molecular biology, five properties were given roughly equal weight, because they had already been measured for many viruses: the type of nucleic acid (DNA or RNA); virion size; virion morphology; virion stability (behaviour across pH, temperature, lipid solvents and detergents); and virion antigenicity. This framework was practical, and it still informs rapid benchtop identification, such as recognising an adenovirus by negative-contrast electron microscopy before confirming it serologically.

Today three primary criteria delineate the main taxa:

1. the type, character and nucleotide sequence of the genome;
2. the strategy of replication; and
3. the structure of the virion.

Genome sequencing now often comes first. Against the reference genomes held in public databases such as GenBank, a sequence can place an isolate into a specific taxon almost immediately, which is why partial sequencing is a routine early step in identifying anything new or unexpected.

The “replication strategy” axis is formalised in the Baltimore classification, which sorts viruses into seven groups by the route they take from genome to messenger RNA. The ICTV families table is itself laid out along these genome-and-replication lines: double-stranded DNA, single-stranded DNA, reverse-transcribing, double-stranded RNA, negative-sense single-stranded RNA, and positive-sense single-stranded RNA.

Defining a species

Species is simultaneously the most important and the most contested rank. The ICTV’s working definition is that a species is a monophyletic group of viruses whose properties can be distinguished from those of other species by multiple criteria. The distinguishing criteria are set per family by the relevant Study Group, and may include natural and experimental host range, cell and tissue tropism, pathogenicity, vector specificity, antigenicity, and the degree of relatedness of genomes or genes.

The reason a species is defined on multiple criteria is that viral taxonomy is polythetic: a taxon is defined by a set of shared properties, no single one of which is essential for membership. No lone feature decides a species, which is exactly what lets the definition tolerate the natural variation found within one.

Two structural rules now tidy the species rank. Since 2018, every species must belong to a genus (a requirement that, surprisingly, was not enforced before), and since the 2021 binomial mandate every species name is binomial (see Nomenclature, below).

Below the species level sit the lineages that often matter most in clinical and epidemiological practice: serotypes, genotypes, subtypes, variants, escape mutants and vaccine strains. The naming conventions here vary from virus to virus and lie outside the ICTV’s remit.

The species is a man-made taxonomic construct, whereas the virus is the actual physical entity. Much historical confusion flows from blurring the two.

Nomenclature: species names, formal taxa, and vernacular names

Three names that look similar must be kept apart: the formal species name, the formal higher-taxon name, and the everyday vernacular virus name.

Species names are now binomial. After decades of inconsistent and often confusing formats, the ICTV mandated in 2021, with the transition essentially completed at the February 2025 ratification vote, that every virus species name take a binomial form: a genus name ending in -virus, followed by a single-word species epithet, written in italics with the genus capitalised. The epithet may be Latinised (Orthoebolavirus zairense) or freeform (Simplexvirus humanalpha1), and uses only the 26 letters of the Latin alphabet, numerals and hyphens. Measles virus, for instance, is the species Morbillivirus hominis. This mirrors the Linnaean convention used across the rest of biology, a format virologists first reached for almost a century ago and have now, in effect, returned to.

Higher taxa, from realm down to genus, are capitalised and italicised, with the rank label preceding the name: “the family Paramyxoviridae”, “the genus Morbillivirus”, “the realm Riboviria”.

Vernacular names are the working names that clinicians and researchers actually use: lowercase, not italicised, and entirely outside the ICTV’s remit. “Measles virus”, “poliovirus” and “hepatitis B virus” are vernacular names, not species names, and they are settled informally by the virologists who study each agent.

Two ambiguities follow:

• A vernacular name is not a species name. “Measles virus” (the agent) and Morbillivirus hominis (its species) are related but distinct: the physical virus versus the taxonomic category. The older practice of writing a species name almost identical to the vernacular name, differing only in italics, is exactly the ambiguity the binomial mandate set out to end.
• The same root at family and genus level. “Bunyavirus” might mean the family Bunyaviridae, the genus Orthobunyavirus, or a particular species. The fix is to state the rank explicitly.

Viro Wiki gives the formal taxon or binomial species name where it carries weight, and otherwise uses the English vernacular.

Practical groupings clinicians use

Distinct from the formal taxonomy, viruses are routinely grouped by how they spread and where they replicate. These categories cut across families and are clinically useful precisely because they track transmission and tropism rather than ancestry:

• Enteric viruses are acquired by ingestion (faecal–oral) and replicate mainly in the gut, remaining localised there. They span Picornaviridae (genus Enterovirus), Caliciviridae, Astroviridae, Coronaviridae, Reoviridae (Rotavirus) and others.
• Respiratory viruses are acquired by inhalation or fomites and stay localised to the respiratory tract: Orthomyxoviridae, Paramyxoviridae, Coronaviridae, Adenoviridae, certain Picornaviridae, and more.
• Arboviruses (arthropod-borne) replicate in a blood-feeding arthropod and are then transmitted by bite to a vertebrate, in which viraemia rises high enough to reinfect feeding arthropods, perpetuating the cycle. Genuine replication in the vector is the defining feature; mere mechanical carriage on mouthparts (the “flying pin”) does not qualify. Found across Togaviridae, Flaviviridae, Rhabdoviridae, Bunyaviridae and Reoviridae.
• Blood-borne viruses are transmitted by transfusion, shared injecting equipment and other parenteral routes, and several are also sexually transmitted: HBV, HCV, HDV, HIV-1 and HIV-2, HTLV-1 and HTLV-2.
• Hepatitis viruses are grouped only because the liver is their main target; hepatitis A, B, C, D and E belong to five entirely unrelated families.
• Oncogenic viruses establish persistent infection and can transform host cells towards malignancy. Those with oncogenic potential are found in Herpesviridae, Adenoviridae, Papillomaviridae, Polyomaviridae, Hepadnaviridae, Retroviridae and Flaviviridae.

Proving that a virus causes a disease

Classification also underpins the harder question of causation. The Henle–Koch postulates, drawn up for bacteria, do not transfer cleanly to viruses: many viruses cannot be grown in culture or animals, disease may appear in only a fraction of those infected, the same syndrome can have several causes, and the agent may even be gone by the time disease appears (a “hit and run” relationship). The postulates were therefore revised for viruses by Thomas Rivers (1937) and again by Alfred Evans (1982).

Fredricks and Relman (1996) reframed them for the sequencing era. Their molecular guidelines ask, in essence: is the viral sequence found in most or all cases (strength); is it confined to diseased rather than healthy tissue (specificity); does its copy number fall with cure and rise with relapse (response to treatment); does infection precede disease (temporality); does the biology make sense (plausibility); is there more virus in more severe or more diseased tissue (biological gradient); and are the findings reproducible across laboratories (consistency)?

HIV and AIDS is the canonical worked example. HIV is found in nearly every case meeting the clinical definition of AIDS; it concentrates in immune target tissues and is absent in healthy people except future progressors; antiretroviral therapy lowers circulating virus and restores immune function; infection precedes disease in mother-to-child, needlestick and transfusion settings; HIV kills CD4+ T cells and macrophages while SIV causes AIDS in inoculated macaques; viral load predicts the rate of progression; and the findings reproduce worldwide.

Two further landmarks show the same reasoning applied to cancer. A prospective study of 45,000 children by the IARC established that EBV infection and exceptionally high antibody titres preceded Burkitt’s lymphoma by 7 to 54 months, that the EBV genome is present in the tumour cells, and that EBV induces lymphoma in primates. In Taiwan, Beasley and colleagues showed that persistent hepatitis B infection raised the risk of primary liver cancer roughly 100-fold, and not the risk of other cancers.

For many virus–disease links, “risk factor” describes the relationship better than an absolute “cause”, because not everyone infected develops the disease and not every case of the disease is caused by that virus. With today’s sensitive assays, careful work is needed to separate a genuine causative role from an incidental co-infection of no clinical significance.

Sequencing and the moving target of taxonomy

Whole viral genomes can now be sequenced in hours, and multiple-sequence alignment with phylogenetic-tree construction is routine whenever a new or unexpected isolate appears. These data continually reshape ideas about the origin and evolution of medically important viruses. Deep conserved sequences justify the handful of orders that exist (Herpesvirales, Mononegavirales, Nidovirales, Picornavirales), and the analyses suggest few further orders will be warranted. At the finer scale, relationships within families and genera are being clarified all the time.

Reassortment complicates the picture in instructive ways. The 2009 H1N1 pandemic virus turned out to be a reassortant carrying genes from four ancestral lineages: North American swine, North American avian, human, and the swine influenza typically found in Asia and Europe. Some members of the Bunyaviridae are natural reassortants too.

The payoff of the whole enterprise is practical. A robust but flexible taxonomy lets a virologist relate a newly found agent to known viruses and anticipate its likely properties, and infer evolutionary relationships that guide diagnosis, research and disease control. The system is scaffolding, and it is meant to keep evolving.

References and recommended reading

• Burrell CJ, Howard CR, Murphy FA. Classification of Viruses and Phylogenetic Relationships. In: Fenner and White’s Medical Virology, 5th edition. Academic Press / Elsevier; 2017. Chapter 2, pp. 15–25. DOI: 10.1016/B978-0-12-375156-0.00002-3. The primary source for the classification principles, criteria, epidemiological groupings and causation framework.
• International Committee on Taxonomy of Viruses Executive Committee (Gorbalenya AE, Krupovic M, Kuhn JH, et al.). The new scope of virus taxonomy: partitioning the virosphere into 15 hierarchical ranks. Nat Microbiol 2020;5(5):668–674. DOI: 10.1038/s41564-020-0709-x. The source for the 15-rank hierarchy that superseded the older five-rank structure.
• Siddell SG, Smith DB. Binomial names for virus species: the rediscovery of an old idea. J Gen Virol 2025;106:002102. DOI: 10.1099/jgv.0.002102. The current account of the binomial species-name mandate (ratified 2021, completed at the February 2025 ICTV vote).
• Pfeiffer JK, Condit RC, Schoggins JW. Principles of Virology. In: Fields Virology, 7th edition, Volume 4: Fundamentals. Wolters Kluwer; 2023. Chapter 2. Current reference text; source for the realm concept, the 2020 taxon census, the polythetic principle, and the 2021 abolition of type species.
• International Committee on Taxonomy of Viruses. Virus Taxonomy: the Master Species List and online Taxonomy Browser (MSL41, ratified February 2026). ictv.global/taxonomy. The live, authoritative source for the current realm list and taxon counts, updated after each ratification vote.
• Krupovic M, Varsani A, Roux S, et al. Reorganization of the realm Monodnaviria (ICTV Taxonomy Proposal 2025.002G; ratified into MSL41). The source for the Monodnaviria to Floreoviria split and the three new realms, cited as the worked example of structural taxonomic change.