Refining the DNA barcode for land plants.

T he goal of DNA barcoding is conceptually simple: Find one or a few regions of DNA that will distinguish among the majority of the world’s species, and sequence these from diverse sample sets to produce a large-scale reference library of life on earth (1). This approach can then be used as a tool for species identification and to help in the discovery of new species (2). Since the first DNA barcoding study in 2003 (1), the “animal barcode,” a portion of the mitochondrial gene Cytochrome Oxidase 1, has proved remarkably effective at discriminating among species in diverse groups such as birds, fishes, and insects. In contrast, finding a robust and effective barcode for plants has been more difficult. In 2009, a large consortium of researchers, the “Consortium for the Barcode of Life (CBOL) Plant Working Group,” proposed portions of two coding regions from the plastid (chloroplast) genome—rbcL and matK—as a “core barcode” for plants, to be supplemented with additional regions as required (3). This recommendation was accepted by the international Consortium for the Barcode of Life, but with the important qualifier that further sequencing of additional markers should be undertaken during a “trial period” (4). This trial period was driven by concerns that routine use of a third (or even a fourth) marker may be necessary to obtain adequate discriminatory power and to guard against sequencing failure for one of the markers (matK can be difficult to amplify and sequence). In PNAS, the China Plant Barcode of Life (BOL) Group provides an impressive dataset tackling this question (5) and assesses the potential benefits of supplementing the core barcode for land plants. A total of 46 research teams, coordinated by Li De-Zhu from the Kunming Institute of Botany (Kunming, Yunnan, China), assembled a sequence matrix from 6,286 samples from 1,757 species from 75 seed plant families in China. They sequenced the core barcoding markers (rbcL and matK) along with two other markers: the plastid region, trnH-psbA, and the internal transcribed spacers of nuclear ribosomal DNA (nrDNA ITS). This dataset was used to assess universality (the ease of recovery of barcode sequences), sequence quality (how good the sequences were), and discriminatory power (how effective the sequences were in distinguishing among species). Their findings on the performance of the three plastid markers broadly match the 2009 CBOL Plant Working Group study (3), albeit here based on much larger sample sizes. Each gene region had different strengths and weaknesses: good recovery and sequence quality but low species discrimination for rbcL, better and broadly equal discriminatory power for trnH-psbA and matK,