How many viruses infect humans?
Published counts run from 219 to 1,131, depending on the evidence required
Matic Broz, PhD9 min
The commonly cited answer is that about 270 viruses are known to infect humans. That figure comes from a 2022 review, not a continuously updated inventory. Scientists keep finding viruses and revising how species are classified.
A broader 2022 atlas found 1,131 virus species in human samples. Not all 1,131 are proven to reproduce in human cells or cause disease. Each study sets its own threshold for what counts as evidence.
To see where a current database count falls between those figures, we matched Virus-Host DB records for humans to the species list ICTV released in July 2026. That gives 404 virus species with a recorded human host, and about half of them are also recorded in other vertebrates.
How many viruses infect humans?
The best concise estimate is about 270 known human-infecting viruses, based on a 2022 review.[1]
The newest detailed census covers RNA viruses only. It lists 239 species with published evidence of natural human infection as of December 2024, up from 214 in 2018. Thirteen additions were newly discovered species. The others reflect taxonomic changes and revised inclusion rules.[2]
That census sets a high evidence bar. A virus must be recognized as a species by the International Committee on Taxonomy of Viruses and supported by a peer-reviewed report of natural human infection. Detection in cultured human cells or deliberate experimental infection does not qualify. A virus does not have to cause symptoms to count.[2]
Most of these RNA viruses also infect other animals. Of the 239 species, 147 are strictly zoonotic, meaning known human cases come from non-human hosts, and only 24 are currently known to infect humans alone. The remaining species include viruses that can spread in limited outbreaks or circulate endemically in people.[2]
Why do estimates range from 219 to 1,131?
Published totals range from 219 to 1,131 because the studies count different kinds of evidence, use different taxonomies, and cover different dates.
A 2012 literature survey counted 219 recognized virus species with evidence of natural human infection. It found species from 23 families and noted that new human viruses were being identified at a rate of three or four per year.[3]
The 2022 review put the concise estimate at about 270 known human-infecting viruses.[1] A separate 2022 Human Virus Database study used a broader rule, collecting animal-virus sequences isolated or detected in human tissues, fluids, and excreta. That atlas contained 1,131 species from 54 families, including 433 DNA viruses, 656 RNA viruses, and 42 reverse-transcribing viruses.[4]
The chart compares published totals under different inclusion rules, not growth over time.[1][3][4]
The atlas identifies viral signals that may deserve further study. Its total is not a count of 1,131 confirmed human pathogens. Finding viral sequence in a blood, stool, or tissue sample does not prove that the virus productively infects human cells.
These species totals are also different from the number of virus particles in one person. The human virome includes viruses that infect our bacteria, while inherited viral DNA includes ancient retroviral sequences that are now part of the human genome.
How many virus species have a recorded human host today?
Matching Virus-Host DB to the July 2026 ICTV species list, we count 404 virus species with a recorded human host, about 2% of the 17,554 species ICTV lists.[5][6] A record means that a virus genome was linked to Homo sapiens in RefSeq, GenBank, UniProt or the literature. It does not show that the virus replicates in people or causes disease.
The species fall in five realms (Riboviria 220, Floreoviria 146, Varidnaviria 22, Duplodnaviria 13, Ribozyviria 3) and 38 families. Two families supply a quarter of the total: Anelloviridae with 51 species and Papillomaviridae with 49. Fifty of those 100 species rest on database annotation alone, with no disease name or paper attached to the human record.
| Genome type | Species | Disease annotated | Literature cited | Annotation only | Also in other vertebrates |
|---|---|---|---|---|---|
| dsDNA | 98 | 34 | 20 | 44 | 26 |
| ssDNA | 83 | 1 | 51 | 31 | 19 |
| Reverse-transcribing | 16 | 4 | 6 | 6 | 11 |
| dsRNA | 13 | 4 | 5 | 4 | 8 |
| ssRNA(+) | 84 | 35 | 37 | 12 | 69 |
| ssRNA(-) and ambisense | 110 | 39 | 53 | 18 | 71 |
| All species | 404 | 117 | 172 | 115 | 204 |
The count falls between the published totals because it uses a different threshold from either. The 2022 review of disease-causing viruses asks for established disease, so it leaves out viruses that are recorded in people without known illness. The atlas of 1,131 species gathered animal-virus sequences detected in human tissue, fluid or excreta, which is not the same as a host record. Our count asks only that the database links the virus to a human host and that ICTV recognizes the species with a vertebrate host. The 117 species with a disease name attached are not a count of pathogens, because that field is filled only where a curated disease entry exists.
Half of the 404 species, 204, are also recorded in at least one other vertebrate. The share is highest for RNA viruses, at 69 of 84 positive-strand species and 71 of 110 negative-strand and ambisense species, and lowest for DNA viruses, at 26 of 98 double-stranded and 19 of 83 single-stranded species. A shared host indicates zoonotic potential, not how often a virus passes from animals to people.
How many virus families infect humans?
Twenty-six virus families are commonly listed as causes of human disease.[7]
The 26-family figure comes from a clinical grouping and changes with taxonomy. The 2012 natural-infection survey counted 23 families across all genome types.[3] The 2026 RNA-virus catalogue also counts 23 families, but only for its 239 human-infective RNA virus species.[2]
The broader Human Virus Database found 54 families because it included virus species detected in human samples under more inclusive rules.[4] Family names and boundaries also change when taxonomists split, merge, or reassign groups. The evidence rule and genome scope belong beside any family count.
How many types of flu virus affect humans?
Three types of influenza virus affect humans: influenza A, B, and C.[8]
Four influenza types are recognized in total: A, B, C, and D. Types A and B cause seasonal epidemics in people. Type C usually causes mild illness and is not associated with human epidemics. Type D mainly infects cattle and is not known to cause illness in people.[8][9]
Influenza A is divided into subtypes by two surface proteins. Scientists recognize 18 hemagglutinin, or H, subtypes and 11 neuraminidase, or N, subtypes. Eight H subtypes and six N subtypes have been detected in humans, while more than 130 H-N combinations have been identified in nature, mostly in wild birds.[8][9]
Only A(H1N1) and A(H3N2) routinely circulate in people. Influenza B is divided into lineages rather than subtypes. B/Victoria still circulates, while B/Yamagata has not been detected since March 2020.[9]
CDC groups flu viruses further into clades and subclades by comparing gene sequences. Those relationships can be examined with a multiple sequence alignment and a phylogenetic tree.
Current U.S. flu vaccines are trivalent. They target an A(H1N1) virus, an A(H3N2) virus, and a B/Victoria-lineage virus.[10]
Methodology
The species count in Table 1 is our analysis of two public resources, both downloaded on October 23, 2026. Virus-Host DB pairs NCBI taxonomy IDs for viruses and hosts, drawing on RefSeq, GenBank, UniProt and curated literature; the file we used is dated October 4, 2026 and reports RefSeq release 237 and GenBank release 273.0.[5] The ICTV Virus Metadata Resource release MSL41.v1 of July 29, 2026 lists 17,554 species and gives each isolate a host source.[6]
We kept the 1,496 Virus-Host DB records whose host taxid is 9606 (Homo sapiens), which cover 533 NCBI species names. Each was matched to an ICTV species by exact name, with one hand-made mapping: NCBI's Orthoflavivirus encephalitidis, the Neudoerfl strain of tick-borne encephalitis virus, is ICTV's Orthoflavivirus neudoerflense. A fallback match on shared accession numbers found nothing, because Virus-Host DB lists mostly RefSeq accessions and the VMR lists GenBank ones.
Records that match no ICTV species are not counted: 128 records under 110 NCBI names, mostly provisional "sp." names, named strains and groups NCBI labels "unclassified", such as human gut microviruses and endogenous retrovirus K. A species that ICTV has since split or renamed can also fail to match, so the count may miss a few; the unmatched names are listed in the data folder.
ICTV's host-source column then screened the list. We kept species whose VMR host source includes vertebrates and set aside 19 whose does not: 14 with other hosts (5 bacterial, 1 protist and 8 unknown or sewage-derived, mostly gut phages and small circular DNA viruses) and 5 with invertebrate hosts only.
Each kept species takes its strongest evidence class across its human records. Disease annotated means a record carries a disease name in Virus-Host DB or its KEGG link; literature cited means a PubMed ID but no disease name; annotation only means neither. These classes describe how the database supports the host record, not a review of the underlying papers, and the disease field is incomplete. A species counts as also recorded in other vertebrates when any Virus-Host DB record for the same ICTV species names a non-human host within Vertebrata. We did not separate natural from experimental hosts.
Genome type and realm come from the VMR. Reverse-transcribing combines single-stranded RNA and double-stranded DNA reverse-transcribing viruses, and ambisense species are grouped with negative-strand RNA viruses. Virus-Host DB only covers viruses with complete genomes in RefSeq or GenBank, so species known from serology or short fragments are absent, and a later download can give different counts. We did not match our 207 RNA-genome species (excluding the 16 reverse-transcribing ones) against the 239 in the 2026 RNA-virus catalogue, so the two lists are not reconciled species by species.[2]
The analysis script, species table, summary counts, unmatched names and provenance notes with file checksums reproduce the counts.
