2008.002G.U.v1.Definition-sp.pdf

The proposal by Gibbs and Gibbs to remove the word polythetic from the current definition of virus species is not in line with the current ICTV practice of using multiple demarcation criteria for delineating different species taxa (1). Characters useful for discriminating between individual species in a genus are host range, cell and tissue tropism, pathogenicity, mode of transmission, antigenic properties, etc. However, any one of these characters can be altered by a few mutations, which is the reason why species taxa are defined polythetically i.e. collectively by a consensus group of covariant characters, and not by a single shared character that would be sufficient to define a species (1,2). This allows a virus mutant with an altered host range, symptom or nucleotide at a certain position in the sequence to remain a member of a particular species. Other reasons for rejecting the proposal of Gibbs and Gibbs have been published (2-5). The current definition of virus species endorsed by the ICTV applies to virus species as the lowest category in the hierarchical system of virus classification (i.e. the category of all virus species taxa). However, the same term “species” is also used for the 1950 individual species taxa recognized in the 8th ICTV report. As emphasized by all taxonomists (6), the two meanings of species should not be confused since it would amount to confounding the element gold (the class of gold atoms) with the concept of element (the class of all elements). Whereas the 1950 species taxa are defined and delineated by a combination of species demarcation criteria selected by ICTV Study Groups, the ICTV definition of the species category only makes the meaning of this class concept intelligible; the definition does not provide the means for identifying and recognizing the concrete objects that are the members, i.e. the referents, of each class of species taxon. The reference to “replicating lineage” in the definition of the category species was included to make it clear that species are not defined only by overall phenotypic similarity but also by the internal cohesion found in a lineage of direct ancestry and descent. The members of a lineage will always share certain inherited characters but, contrary to the claim of Gibbs and Gibbs (7), this does not contradict the fact that species taxa are defined polythetically. Lineages can be divided up in a variety of ways and some of them will correspond to species while others will correspond to genera and families. Including the concept of lineage in the definition of the species category does not mean that it is necessary to know which parts of the genome sequence are retained in the members of a particular species taxon; other parts will be retained in the case of individual genera and families. The definition of the species category elucidates the meaning of the concept but this definition should not be confused with the task of identifying the individual virus members of a species taxon using diagnostic characters (2-5). Virus identification consists in allocating a virus to a species but this requires that the species had to be created and defined polythetically by taxonomists beforehand (2). Only by comparing the properties of many members of established species is it possible to discover which diagnostic character or set of characters will discriminate between the members of different species. It may turn out that a nucleotide motif or the reactivity with a particular monoclonal antibody may suffice for identifying a virus as a member of a species. However, the presence of such a diagnostic character does not imply that a single character is sufficient to define a species monothetically as claimed by Gibbs and Gibbs. If a species could be defined in this way, identifying a virus as a member of a species would not be very informative. It is the frequent combined occurrence of many defining characters in the members of a species that makes it possible to predict many of the properties of a newly discovered virus, once it has been identified as a member of the species. Characters are the traits and attributes use to describe and recognize a particular individual. Characters are often confused with parts, as shown by the use of the term “feature” to refer to both attributes and parts (8). A character or trait used in the definition of species taxa should not be confused with a part of a virion such as a particular nucleotide at one position in the genome sequence, or a nucleotide motif or the entire genome sequence (2). A part of a thing is a thing and not a property (9) and a monothetic species class, allegedly "defined" by a nucleotide sequence, would thus be based on a single phenotypic character, namely a particular chemical sequence inside a virion. This leads to a classification of genome sequences rather than of viruses. References: 1. Ball L.A. (2005) 8th ICTV Report, Elsevier-Academic Press, pp 3-8. 2. Van Regenmortel M.H.V. (2006) Virologists, taxonomy and the demands of logic. Archives of Virology 151: 1251-1255. 3. Van Regenmortel M.H.V. (2003) Viruses are real, virus species are man-made, taxonomic constructions. Archives of Virology 148: 2481-2488. 4. Van Regenmortel M.H.V. (2007) Virus species and virus identification : Past and current controversies. Infection, Genetics and Evolution 7: 133-144. 5. Van Regenmortel M.H.V. (2008) Virus species. In : Encyclopedia of Virology, 3rd Edition (Mahy B.W.J. and Van Regenmortel M.H.V., eds), Elsevier, pp 401-406. 6. Mayr E. (1982) The Growth of Biological Thought. Diversity, Evolution and Inheritance. Harvard Univ Press, Cambridge, Mass. 7. Gibbs A.J. and Gibbs M.J. (2006) A broader definition of the virus species. Archives of Virology 151: 1419-1422. 8. Ghiselin M.T. (1997) Metaphysics and the Origin of Species. New-York State Univ Press, New-York, p. 201. 9. Mahner M. and Bunge M. (1997) Foundations of Biophilosophy. Springer, Berlin, p.11.

Marc van Regenmortel’s 968 word ‘comment’ on our proposal to ditch the word “polythetic” in the ICTV definition of a virus species is mostly an irrelevant filibuster.  The only crucial issue is the meaning of the word polythetic.  Does it merely mean “variable” as Marc explains, at length?  Or does it describe a particular and restricted type of variability, as most of us believe?  The paper on which our proposal is based (Gibbs & Gibbs A broader definition of ‘the virus species’.  Arch. Virol. 151(7):1419-22, 2006) describes the difference between variable and polythetic, and gives references to further mainstream reading available on the WWW.

Adrian Gibbs in his comment is correct in pointing out that his proposal to remove the word “polythetic” from the current ICTV definition of virus species hinges on the meaning of the word “polythetic”. He maintains that his 2006 article (Arch Virol 151, 1419-1422) clarifies that meaning whereas his paper, in fact, obfuscates important distinctions that need to be made.

Adrian claims that removing the term “polythetic” from the definition would allow virus species to be defined more broadly by sets of any type of character. He makes this claim because he wants to promote genome-based characters which he erroneously regards as monothetic characters useful for defining species. However, there are no such things as monothetic or polythetic characters and any type of property can be used to define a monothetic or polythetic class. These terms actually describe two types of classes with different distributions of properties but not two types of properties. In a polythetic class, each member of the class shares a large but unspecified number of properties, each property is present in a large but unspecified number of members of the class and no property is necessarily present in all the members of the class. A monothetic class differs from a polythetic one in that it is defined by properties that are both necessary and sufficient for establishing membership in the class. This means that all the members of a monothetic class must necessarily share one or more characters.

Whereas genera and families are monothetic taxa defined by stable properties such as virion morphology, species are polythetic taxa defined by properties such as host range and pathogenicity that are easily modified by mutation and are therefore not necessarily present in every member of a species. This leads Adrian to claim that I use the term polythetic to mean “variable” whereas I have repeatedly explained that a polythetic class is defined by the absence of a given property in individual members of the class.

Adrian also seems to have difficulty understanding that to identify a virus as a member of a species, that species must have been created by taxonomists beforehand (Van Regenmortel, Systematic and Applied Microbiol 33, 1-6, 2010). If hundreds of isolates of a rod-shaped virus are sequenced and a variety of common nucleotide motifs are detected in different isolates (see Arch Virol 151, 1419-1422), there is no way to ascertain which motif on its own can be used to define either a particular variant, strain, species or genus. Nucleotide motifs, like monoclonal antibodies that can differentiate tobacco mosaic virus from tomato mosaic virus (Arch Virol 95, 191-203, 1997) are diagnostic markers useful for identifying which viruses are members of a pre-existing species. In no way are such markers actually species-defining properties that would allow the creation of monothetic taxa.

Adanson was the first to use the polythetic principle of overall similarity in taxonomy. Adrian is quoting  18th Century French Adansonian sentences (see Arch Virol 151, 1419-1422) in an effort to convince us that Adanson was trying to define taxa more monothetically. How a taxon can be “more” monothetic is meaningless.  A taxon is either monothetic or it is not, in the latter case it is polythetic.  

The undesirable consequence of defining virus species monothetically using a single arbitrarily chosen percentage of pair-wise sequence identity between pairs of individual viral genome sequences is demonstrated in the case of the genus Begomovirus (italics). Using a pair-wise sequence identity of less than 89% in the DNA-A genome component of begomoviruses as a single demarcation criterion, more than 180 different species have been created (see Arch Virol 150, 2151-2179, 2005 and Arch Virol 154, 1181-1188, 2009). However, in many cases they do not correspond to biologically distinguishable entities that would justify the label species. Relying on a single genome-based character leads to a classification of genomes rather than of viruses.