Effect Measure

Naming a bird flu virus

For years we have been naming flu viruses in a particular way. Now Declan Butler has a news article in Nature observing that the system is being modified for bird flu to be more “politically correct.” What is the system that’s being modified (still in use for seasonal flu)?

It starts off with the type of influenza virus, either A, B or C. The type was originally based on a broad class of antibody response but now is related to genetic markers in two of the eleven proteins (matrix M1 or nuceoprotein, NP). The three types have a common ancestor but only types A and B are of public health importance. Type A flu virus comes in 144 different flavors, all the different combinations of the 16 HA surface proteins and the 9 NA surface proteins (16 times 9 is 144). One of those combinations is H5N1. The usual seasonal flu viruses are combinations (subtypes) H1N1 and H3N2. So if you want to identify a flu virus you have to give its type and subtype, as in A/H5N1. But there is more to naming because even within a subtype there are lots of variations (subtype strains; note that H5N1 is not a strain, it is a subtype).

When a sample is taken from nature (say from a bird or a person) and then propagated in host cells like dog kidney cells, it is called an isolate, a single virus isolated from a clinical or animal specimen. This is done in some laboratory, which labels the isolate according to its own idiosyncratic system. The lab’s serial number or isolate code becomes part of the identification, as does the animal and the geographic area and the year. The only exception to this is when the animal is a human being. If no animal is cited in the name, it was isolated from a human. Here are some examples of full flu virus names:

A/Bar-headed Gs/QH/12/05(H5N1) [type A, subtype H5N1 isolated from a goose (GS) in Qing Hai, China (QH) in 2005, coded with a label numbered 12 in the isolating laboratory]. This virus was shown by genetic analysis to be related to A/Chicken/Shantou/4231/2003. This is a slight variation, with 2005 in place of 05 and omitting H5N1 when the context is clear. The animal is a chicken, the place is Shantou, China and the label on the sample in another laboratory is 4231.

Here’s an H3N2 subtype isolated from a patient: A/Memphis/8/88(H3N2) (i.e., influenza A/H3N2, isolate labeled 8 from Memphis in 1988). When referring to specific isolates in scientific papers often a shorthand is used. In this case, after an initial identification, the virus might be referred to merely as M88 (Memphis 1988, the rest having been noted earlier).

So how is this system being modified and why?

Names of flu viruses appearing in the scientific literature have undergone something of a mutation this year. The group of H5N1 avian flu viruses, once known as ‘Fujian-like’, has morphed into ‘Clade 2.3.4′. And its ‘Qinghai-like’ cousin is now called ‘Clade 2.2′. They are part of a complete revision of H5N1 nomenclature recommended by the World Health Organization (WHO).

The system is now more politically correct. It avoids the “stigmatizing labelling of clades by geographical reference”, according to the WHO. In 2006, when scientists assigned the name ‘Fujian-like virus’ to a vaccine-resistant strain of H5N1 that spread across Asia, China objected strongly to the name. Government officials argued that the southeastern Chinese province of Fujian was tainted by association as the virus spread.

But a WHO official told Nature that discussions for a revised nomenclature were underway before the spat, and that the “impetus was scientific”. “We recognized that naming based on geography was probably not a good idea for other reasons,” she says. (Declan Butler, Nature)

The new designations are based on phylogenetic (“family”) trees for the H5N1 subtype. By analyzing the relationships in the hemagglutinin (H5) surface virus, ten lineages (clades) have been identified with sub branches in several. The clade and geographic designations are not the same. For example, clade 2.3.4 is found not only in Fujian province in China but in Laos, Myanmar and Vietnam. Clade 2.2 is spread over 60 countries on three continents. Bird flu experts hope the new terminology will make relationships clearer. The genetic sequences in GenBank, however, will continue to use the old naming system which gives the geographic origin of the isolate.

The naming business provides a fascinating look into the interplay between science, established convention and politics. As we know from the current dispute between Indonesia and the world scientific community, those same forces and more are at work in determining which viral isolates even exist to have names put on them. How the world works may be independent of our perceptions of it but what we know about the world and how we learn it certainly has a large social component. You don’t have to subscribe to a full blown “social construction of reality” viewpoint to see it. Just look at how a virus is named.


  1. #1 Marissa
    April 25, 2008

    I an see the logic in this new system, but I doubt it would have happened without political prssure. Time ill tell regarding its utility. You’re right Revere, most interesting interplay.

  2. #2 anon
    April 26, 2008

    virus names must be informative and uniform, computer-readable, easily convertable.
    Not appealing or politically correct.
    So it’s important that humans don’t create new formulations to describe
    viruses but rather choose from a list of descriptions.
    No creativity please. Computers don’t like it.
    You should really view it from the computer’s POV. Those are the machines
    who read the viruses for us.
    How does a Computer realize that a virus is human ?
    The first entry could be an animal or a country.

    Naming them “Fujian-like” is easier to remember than clade 2.3.4.
    Hopefully there will be common lists to correlate the old lists with the new ones.
    Why clade 2.3.4 and not clade 57 or clade xyz ? Will it grow to
    clade in some years ?

    I assume that we will still continue to call it Qinghai and Fujian in posts and
    emails, and clade 2.3.4 only in official papers and WHO-press issues.
    Then we will need an additional dictionary to translate those official
    statements, sigh. But maybe an automatic webpage for such translation
    can be installed – should be easy.

  3. #3 anon
    April 26, 2008

    and please include the whole date, not only the year.
    It’s important.

    And the geographical coordinates – but that goes to
    the record, not the name.
    Also a species-code, segment, country-code, province-code
    and properly aligned, please.


    A327: H3N2,segment 7
    88D13: 1988,Dec.13

  4. #4 Jonathon Singleton
    April 26, 2008

    “The genetic sequences in GenBank, however, will continue to use the old naming system which gives the geographic origin of the isolate…” — well, this is sensible! This whole thang kinda reminds me of when a group of people alter interpersonal relationships in order to massage the “arrested developed ego” of the designated ALPHAs of the group. But, most realize the sensible thang is maintaining workable protocols based on the reality of lived existence. Transgenic pathogens such as our deliciously murderous H5N1 do what, fer fracks sake, kill many different species! Why!?!

  5. #5 lugon
    April 26, 2008

    It’s somewhat fun to watch what much less serious folks do in the free software world. Go to http://www.debian.org and find out how different versions are named after characters in a children’s animated movie (likewise for ubuntu linux). Folks who need to know use numbers, and the other two extremes (folks who don’t need to know, and folks who DO know) use names. Short, expressive, memorable names, just like when people give names to their cars.

    It’s always about whatever works, methinks.