‘Botany without Borders’ brought together the annual meetings of the Botanical Society of America, the Canadian Botanical Association/L’Association Botanique du Canada, American Fern Society and American Society of Plant Taxonomists. It was held July 26–30 in Vancouver, Canada. It proved to be a highly successful transnational meeting. As expected, presentations covered a huge range of topics. High points included talks by Nobel laureate Carl Wieman on ‘using the tools of science to teach science’, and Loren Rieseberg on ‘speciation genes in plants’. Symposia of note included one in honour of the great pteridologist Gerald Gastony and several excellent sessions on polyploidy. Additional major areas of discussion included systematics, economic botany, teaching and literacy, paleobotany and genomics.
One topic consistently under discussion was DNA barcoding, perhaps because of the Canadian location (given that Canada has invested heavily in barcoding, for instance with the opening of the Biodiversity Institute of Ontario in 2007). No less than 16 poster and talk abstracts contained the word ‘barcoding’ or ‘barcode’. These included case studies on tropical ecosystems (David Erickson), ferns (Li-Yaung Kuo), fungi (Mary Berbee), wild nutmegs (Royce Steeves), medicinal plant products (Andrea Schwarzbach), Lotus (Isidro Ojeda), Salix (Hyosig Won), Salix again (Helena Korpelainen), diatoms (Birgit Gemeinholzer), Inga (Kyle Dexter), Cyperaceae (Jessica Le Clerc-Blain), Cortinarius (Emma Harrower), Carex (Brianna Chouinard), Swertia (Kunjani Joshi) and Acer (Jianhua Li).
Mark Chase (on coding and non-coding barcodes), and David Spooner (on problems and challenges) both gave papers tackling problems with common barcoding approaches. Finally, a poster from Fay-Wei Li provided a potential technical innovation for improved throughput (tissue-direct polymerase chain reaction for barcoding).
David Spooner's talk entitled ‘DNA barcoding: an oversimplified solution to a complex problem’ laid out in stark terms some of the key criticisms of DNA barcoding, using examples from his work on Solanum, while lauding the goal of ‘democratizing taxonomy.’ Unlike the situation in animals, where a short (~650 bp) region of the mitochondrial gene COI is enough to identify many metazoan taxa to the species level, there is no known mitochondrial, chloroplast or nuclear locus that is easy to amplify and align in disparate plant lineages while containing ample sequence variation to distinguish species. Plant organellar genomes evolve several orders of magnitude slower than those of animals, far too slow to record the rapid speciation that is possible in some plant groups. Dr Spooner argued that barcoding is retroactive and relies on well-defined species to function properly, while only a small fraction of the 3.6–100 million estimated species on the planet have been described. Additional biological phenomena that would ‘impede’ barcoding in plants include the high rates of auto- and allopolyploidy, asexual taxa and the high frequency of hybridization between many plant species. These and other issues lead the speaker to conclude that ‘there is no level of sequence divergence in any single or limited set of genes that is either necessary or sufficient for species identification.’
Similarly, Helena Korpelainen presented data showing the difficulties of using DNA barcoding in the genus Salix, where trnH-psbA shows essentially no genetic structure. She suggested an alternate approach using microsatellites, whose faster rates of evolution reveal much more genetic structure. Additionally, these nuclear loci can be used to identify hybrids and polyploids. On the other hand, the set of microsatellite loci would only be useful in this genus, and their use would be difficult to automate.
Mark Chase voiced concerns about using non-coding DNA for barcoding, pointing out the difficulty with aligning non-coding sequences due to length variation and rearrangements. Additionally, the high frequency of mononucleotide repeats makes sequencing difficult in these regions. As an example, most monocotyledons have a large (20+ bp) poly-A repeat in the commonly used trnH-psbA region. Moreover, Dr Chase argued that the problem is not that we just have not found the right region to sequence — there has been so much work done already on this problem that ‘if there had been a good marker out there we would already have it.’
Because of these and other concerns, many participants of the workshop on ‘applying modern genomic tools to the management and characterization of plant genetic resources’ came to the consensus that barcoding will never work in plants in the same way that it works in animals. However, alternate approaches to the goals were put forward during the wide-ranging discussion. Participants pointed out that the whole chloroplast genome contains about as much information as the short mitochondrial barcoding sequence used in animals. Furthermore, massive parallel sequencing (MPS) may soon make it possible to sequence whole plastid genomes as a barcode — so-called ‘ultrabarcoding’. Rich Cronn and Aaron Liston of Oregon State University reported an optimization of the Solexa system that allowed whole plastid sequencing for a core sequencing cost of around $100. This cost will probably fall as new technologies become available, particularly if multiplexing becomes efficiently implemented for Applied Biosystem's SOLiD sequencers.
An alternative approach suggested by participants in the workshop was to make use of chip-based single-nucleotide polymorphism (SNP) detection methods to interrogate either the entire plastid genome (whole plastome genotyping), or 10 to 100 nuclear genes. It was pointed out that even with whole plastome genotyping or ultrabarcoding, what is being identified is not a species but a haplotype. Rather than a single species, the user may be given the list of species that are associated with a given haplotype, and additional information can be used to make a final identification. Loren Rieseberg in his talk to the workshop (‘population genetic challenges and the potential of modern genomics technologies for the management and characterization of plant genetic resources’) discussed various approaches, including the idea that the ultimate in DNA barcoding may come from the nuclear genome through the use of new technologies to examine sequence variation in conserved orthologous sequence (COS) genes.
Pragmatic approaches were reflected in several talks throughout the meeting, where speakers had used DNA barcoding to solve specific problems. David Erickson pointed out that we should keep our goals for barcoding realistic. He presented results of a study of trees on Barro Colorado Island (BCI), Panama, where successfully identifying genus would be very useful. The 295 tree species on BCI can be quite difficult for even experts to identify in the field, but as most of the 189 genera have only one species present on the island, the usual problems of barcoding can be simplified using this geographical information. Using rbcL and trnH-psbA, he was able to identify 97% of the 1035 samples he examined to species level and 100% to genus level. The problematic samples all belonged to four diverse genera — Psychotria, Ficus, Inga and Piper. This work is already being used for ecological studies on BCI, and makes new types of studies possible (e.g. by identifying roots).
Royce Steeves discussed a different use for DNA barcoding. He presented work on wild nutmeg trees of the genus Compsoneura. This genus presents extraordinary challenges for those wishing to identify specimens in the field, as most of the taxonomically informative characters rely on close examination of the male flowers, which are smaller than a grain of rice. In addition to being difficult at the best of times and absent much of the time in males, as these trees are dioecious these characters are always absent in females. A combination of trnH-psbA and matK was used to correctly identify 94.7% of the samples to species, compared to an estimated rate of only 40% based on characters available in the field.
Despite these successes, DNA barcoding will likely never live up to the hype of its most ardent promoters, and plant DNA barcoding in particular is fraught with problems. Nevertheless, it has already shown promise in many applications, particularly when other information is available, such as geography in the case of BCI. Whatever the problems, DNA barcoding in plants is clearly here to stay and there is consequently an urgent need to rise to the scientific challenges it presents. Judging from the successful applications of plant DNA barcoding discussed at this conference, this process is well underway.