Shortcuts in systematics? A commentary on DNA-based taxonomy

The primary aims of taxonomy are to name, circumscribe, describe and classify species. The first goal is convention but the remainder are science. The International Codes of Nomenclature are legislative instruments and nomenclature is simply a mechanism to ensure that a species name is legitimately attached to a type specimen, regardless of scientific status. The type of a species does not serve, as Tautz et al. ([1xDNA points the way ahead in taxonomy. Tautz, D. et al. Nature. 2002; 418: 479CrossRef | PubMedSee all References, 2xA plea for DNA taxonomy. Tautz, D. et al. Trends Ecol. Evol. 2003; 18Abstract | Full Text | Full Text PDF | Scopus (441)See all References], but see [3xThe status of taxonomic literature. Minelli, A. Trends Ecol. Evol. 2003; 18Abstract | Full Text | Full Text PDF | Scopus (22)See all References][3]) assert, as ‘the central reference for comparisons’. The crucial link between names and scientific investigation is species circumscription followed by description. The Codes require Linnaean binomials: a genus name and a species epithet. The rules are totally silent about what constitutes a species; rather this is a key goal of biological investigation. Circumscriptions of genera and species evolve as science progresses.The Linnaean binomial system is not ‘inherently unstable’ but is used to interpret the underlying science. The problem that ‘a name that has been used for a long time thus can suddenly disappear’ (or reappear, for that matter) is a nuisance for everyone. However, if name changes are considered a serious problem, then the conservation and rejection criteria available in the Codes can be invoked.Neither the Botanical nor the Zoological Code has a fixed authoritarian supervisory body; rather, they have committees chosen by a democratic process that oversee changes. The Codes work by consensus and are designed to be open and universally applicable [4xTaxonomy needs evolution, not revolution. Knapp, S. et al. Nature. 2002; 419: 559CrossRef | PubMedSee all References][4]. Tautz et al. ([2xA plea for DNA taxonomy. Tautz, D. et al. Trends Ecol. Evol. 2003; 18Abstract | Full Text | Full Text PDF | Scopus (441)See all References][2], and see [3xThe status of taxonomic literature. Minelli, A. Trends Ecol. Evol. 2003; 18Abstract | Full Text | Full Text PDF | Scopus (22)See all References][3]) advocate universal, centralized, apparently obligatory registration, a concept emphatically rejected by the International Botanical Congress in 1999. Taxonomists from developing nations led the move against centralization, fearful that the wealthier nations were attempting to monopolize information. In our view, registration would eventually strangle systematics, as debate will be discouraged.Most current taxonomy is pursued using low-cost technology. Mandatory introduction of DNA sequences into taxonomy seems to us a retrograde step. In most instances, a quick survey of morphology will serve the same purpose and, although morphology has its problems, DNA has as many pitfalls. A sufficiently different sequence might warrant the description of a new species, as will a sufficiently different morphology. An expensive and centralized DNA-based taxonomy would only add to the North–South divide in taxonomy, and might exclude the many taxonomists who have limited access to sequencing technology.Acknowledging that there need not (or cannot) be universal agreement about which region of the genome to sequence, Tautz et al. suggest using house-keeping genes, especially the ribosomal genes (at least in animals). Although abundant, ribosomal sequences might be an inappropriate choice in the long run, because they have profound alignment problems and are subject to different degrees of concerted evolution.Matching existing Linnaean names with DNA sequences is fraught with hazards. Although it is tempting to create a new starting date for priority for the Botanical and Zoological Codes [5xChallenges for taxonomy. Godfray, H.C.J. Nature. 2002; 417: 17–19CrossRef | PubMed | Scopus (246)See all References][5], it seems futile to replace existing types (from which DNA in some cases cannot be extracted) with neotypes. As indicated by the authors [1xDNA points the way ahead in taxonomy. Tautz, D. et al. Nature. 2002; 418: 479CrossRef | PubMedSee all References][1], expert taxonomists are in short supply and many important groups are neglected. It takes little imagination to envisage the problems broad-scale designation of neotypes would cause.Tautz et al. argue that existing DNA data bases represent bad taxonomy, as ‘there is no guarantee that the correct species names were assigned by the submitter of the sequence, because there are no established taxonomic standards under which such submissions have to be done’. The circumscription of a species is an opinion [6xSee all References][6]. One might ask therefore who is going to decide on ‘taxonomic standards’. Circumscription changes with increasing knowledge; that's the science. Many sequences are deposited in DNA data bases, but, if a specimen has been misidentified, only inspection of the voucher can solve this problem. It is naïve to think that ‘phylogenetic analysis of query sequences, will readily place any sequences from new species’, as if the problems of homoplasy, alignment, and even phylogenetic methods would not add to the unreliability of information in existing data bases.The role of collections in systematics is vital. Collections represent a comparative model of diversity, and therefore, as the authors [1xDNA points the way ahead in taxonomy. Tautz, D. et al. Nature. 2002; 418: 479CrossRef | PubMedSee all References][1] point out, specimens should be retained as intact as possible. Destructive sampling of type specimens for any characters has always been a severe problem and such procedures are accepted only as a last resort. To destroy a sample to extract its DNA might be inadvisable in the light of future needs.With regard to taxonomic and phylogenetic studies, current practise attempts to create higher level classifications that do not conflict directly with cladograms [7xSee all References][7]. Today, phylogenetic methods applied to DNA data are a strong focus of systematics research, and it is difficult to believe that the pendulum might swing in another direction. However, methods do change and the fixed opinions of today might seem redundant in the future.So what is the upshot? Indeed, what has the molecular revolution really achieved for taxonomy? The main advance is that it has provided access to copious data for cladistic analysis and, in our view, has provided new data sets and a new class of characters that can be extremely useful in broad-scale comparisons of everything from bacteria to mammals and plants. Often one is forced to use one, or a few, carefully selected specimens as representatives of a taxon, but experience has repeatedly shown that this can be a major mistake. Deliberately using a single specimen as a representative of the taxon will only create havoc in taxonomy, a fact long realized by taxonomists working with other types of data. Individual bases and DNA sequences are simply characters, tiny fragments of the lifecycle. It seems perverse to us to advocate using a DNA sequence as a mandatory identification tag for a species, even as a first approximation. Therefore, this plea for a DNA-based taxonomy is little more than a cri de coeur for bioinformatics and increased reliance on sequence data. We agree that there is a very strong need for efficient, although perhaps not centralized, repositories for DNA samples. How and where such samples should be curated, and how the curators should be funded, is an open question. We also agree that it would be a good idea to include DNA sequences in the diagnoses of taxa. However, we feel that a DNA-based taxonomy along the suggested lines would catastrophically retard taxonomic activity, and it would certainly not relieve the need for many more taxonomists, especially in the tropics, for the good of taxonomy's many dependent user groups.