Washington DC

It may be too early to use a technology called DNA barcoding to speed the identification of species, says an analysis released last week.

DNA barcoding involves collecting and comparing genetic sequences from many species. Some proponents believe that bypassing the slower techniques of traditional taxonomy will identify unknown species and categorize the world's biodiversity more quickly. But a study by Christopher Meyer and Gustav Paulay of the University of Florida in Gainesville shows that barcoding works well only for species that are already much studied.

DNA barcoding identifies well known species such as this chestnut cowrie, but it comes a cropper on others. Credit: B. D. COLE/CORBIS

The scientists examined a database of marine snails called cowries, which have been studied since the nineteenth century because of their valuable shells. By analysing the sequence of a gene called cytochrome oxidase 1 from a cowrie specimen, then comparing it with the entire database of genetic sequences, the scientists correctly identified cowrie species with less than 4% error.

But Meyer and Paulay hit trouble when they reconfigured their database to examine how barcoding works for less well studied groups. They used a method pioneered by other barcoding scientists to pinpoint new species. This method compares diversity among members of the same species and diversity between different species. They showed that in species that have been less studied, the barcoding method misidentified unknown specimens up to 20% of the time (C. P. Meyer and G. Paulay PLoS Biol. 3, e422; 2005).

Some scientists view the high success rate for known species as a triumph for barcoding. They also say that the field is very young, and is still working out the best ways to deal with organisms about which there is little information. “This was a very ambitious test for the technique,” says David Schindel, executive secretary of the Consortium for the Barcode of Life, a group of 93 barcoding practitioners hosted by the Smithsonian Institution in Washington DC. “Chris and Gustav have done a very good job of presenting the possible procedural problems that barcoding will encounter, and we are working out protocols for more sophisticated analysis of barcode data,” says Schindel.

Meyer stresses that the concept of barcoding remains solid. Indeed, his museum is a member of the barcoding consortium. But, he says, his study shows that barcoding will not supplant traditional taxonomy. Meyer says he hopes his study will help convince funding agencies to support old-fashioned taxonomy — a discipline that seems to be slowly disappearing — as well as DNA barcoding.