Species Discovery versus Species Identification in DNA Barcoding Efforts: Response to Rubinoff

Rubinoff's (2006) essay in Conservation Biology raises some important issues relevant to the DNA barcoding initiative. Some of these issues are valid, others have been discussed recently in the DNA barcoding literature (Blaxter Sperling 2003; Hebert et al. 2004; Moritz Ebach Banbury, 2 September 2003; Smithsonian Institution, Washington D.C., May 2004; CBoL London Meeting, February 2005; for reports on these meetings see http://phe.rockefeller.edu/BarcodeConference/). There were originally two purposes of DNA barcoding discussed and debated at these initial meetings. The first and most controversial aspect is species “discovery” (Matz Lipscomb et al. 2003; Sites DeSalle et al. 2005). To look only at DNA sequences of organisms and then call these organisms species without a hypothesis test completely ignores the intellectual and scientific content of taxonomy (Dunn 2003; Lipscomb et al. 2003; Seberg et al. 2003). On the other hand DNA sequences can be used to augment taxonomy and species delimitation as either corroborating evidence for existing hypotheses or as starting points for further testing by other means (DeSalle et al. 2005). Species “identification” based on DNA sequences does not violate the intellectual content of taxonomy or the scientific basis of taxonomy and species delimitation. This statement makes sense because species identification starts with known species designations on the basis of the work of taxonomists and biologists studying speciation. Species identification through DNA sequencing, then, uses the delimitations of taxonomy to establish a means for identifying unknown organisms by querying an existing database. Rubinoff takes issue with the distance-based “barcode” approach as do several other authors (DeSalle et al. 2005; Meyer Prendini 2005). When a character-based system in which DNA sequence information is used to generate discrete diagnostics (DeSalle et al. 2005) is implemented, the distance problem disappears. In fact, a DNA character-diagnostic approach then can become an integral part of the usual way taxonomists examine morphological or behavioral attributes when describing species. In addition, DNA barcoding can contribute to refining species discovery. Once a set of barcodes is established for a group of organisms, unidentified specimens (i.e., query specimens) can be examined. With an effective DNA identification or barcoding system, typically the query specimens will be identified to species based on the barcodes from a database. Nevertheless, some query specimens will not be identified to species. In these situations the DNA sequence information should raise a red flag about the existing taxonomy. The DNA barcodes, then, offer new hypotheses about the taxonomy of the group. These DNA sequence-based hypotheses are then open for testing and are the stimulus and starting point for further taxonomic revision of a particular group. Once the taxonomic community and researchers in DNA barcoding both get the point that DNA barcoding is an extension and augmentation of taxonomy and can be used for species identification (not discovery), the problems discussed by Rubinoff evaporate. Hebert et al. (2003) suggest that there is enough power in short DNA sequences to provide unique signatures that could be used as character-based diagnostics. Davis and Nixon (1992) point out that species can be diagnosed by fixed and different attributes. These fixed and different characters are the gold standard of a character-based DNA identification system. DeSalle et al. (2005) also point out that DNA sequence positions in a short stretch have even more diagnostic character information than the positions that are fixed and different among species. Combinations of positions that are private and even polymorphic to particular species can be found and used in species identification (DeSalle et al. 2005). This leads me to suggest that a short, single nucleotide sequence the size of the Cytochrome Oxidase I (COI) fragment proposed by the DNA barcoding community is long enough to generate unique character-based DNA diagnostics for even very closely related cryptic species. Should one use the COI gene by itself to generate a phylogenetic hypothesis? The answer to that question based on the last decade of systematic research with multiple gene partitions is a resounding no (Gatesy et al. 2002; Rokas et al. 2003). The final problem of DNA barcoding application in the conservation community discussed by Rubinoff is also well taken. Conservation biologists have to work with the public sector at many junctures. The public does need to be convinced that a conservation approach is scientifically valid. Any scientific approach needs to be viewed as reliable, consistent, honorable, and applicable in courts of law and in national and international governing agencies. On the other hand, conservation biologists require methods that are rapid and informative because conservation biology is a crisis discipline (DeSalle rather, I argue that a barcode include DNA, morphology, behavior, geographic information, and any other attribute of the organism that facilitates its identification. If DNA methods are accessed with ease and rapidity in the field, then there is no reason they should not be used if a species identification system has been established. It is not unreasonable to assume that DNA technology will advance to the stage where field-based diagnostics can be accomplished. This development would greatly facilitate field conservation and biodiversity studies, and the utility of a field-based system of rapid species identification should be of obvious importance in conservation, biodiversity studies, and ecological assessments. One of the hardest criticisms levied by Rubinoff concerns the skewed degree of funding to the DNA barcoding community when classical taxonomy could use the funding. This is a political as opposed to a scientific issue. Nevertheless, there are some underlying scientific points to be made that are relevant to this criticism. The DNA sequences in a species identification system can offer an excellent rationale for organizing and interacting with biodiversity databases. Workers are currently attempting to organize the information emanating from biodiversity studies. These initiatives include the Global Biodiversity Information Facility (GBIF), All Species Foundation, Morphobank, and Treebase, among many. The lack of uniformity of information being reported across these databases is a difficulty (Agosti 2003, 2006). Obtaining, archiving, and using a common DNA sequence or set of DNA sequences across a wide range of taxa with a uniform format for submission, accession, storage of tissues, and informatics would greatly enhance our understanding of biodiversity. DNA barcoding initiatives include such approaches to database construction as an integral component of the barcoding effort. This is a plea not to throw the baby out with the bathwater, as Rubinoff suggests with DNA barcoding. His criticism is valid only if species discovery, exclusively through DNA, becomes the raison d' etre of DNA barcoding. If DNA barcoding is defined and used as a species identification method, it has the potential to augment and draw funds to biodiversity assessment and conservation biology projects in general. I thank R. Hanner for critical reading of an early draft of this paper. The writing of this Comment was supported by the Louis B and Dorothy Cullman Program in Molecular Systematics at the American Museum of Natural History and by a grant from the Lounsberry Foundation.