Abstract
BackgroundDNA barcoding of rain forest trees could potentially help biologists identify species and discover new ones. However, DNA barcodes cannot always distinguish between closely related species, and the size and completeness of barcode databases are key parameters for their successful application. We test the ability of rbcL, matK and trnH-psbA plastid DNA markers to identify rain forest trees at two sites in Atlantic central Africa under the assumption that a database is exhaustive in terms of species content, but not necessarily in terms of haplotype diversity within species.Methodology/Principal FindingsWe assess the accuracy of identification to species or genus using a genetic distance matrix between samples either based on a global multiple sequence alignment (GD) or on a basic local alignment search tool (BLAST). Where a local database is available (within a 50 ha plot), barcoding was generally reliable for genus identification (95–100% success), but less for species identification (71–88%). Using a single marker, best results for species identification were obtained with trnH-psbA. There was a significant decrease of barcoding success in species-rich clades. When the local database was used to identify the genus of trees from another region and did include all genera from the query individuals but not all species, genus identification success decreased to 84–90%. The GD method performed best but a global multiple sequence alignment is not applicable on trnH-psbA.Conclusions/SignificanceBarcoding is a useful tool to assign unidentified African rain forest trees to a genus, but identification to a species is less reliable, especially in species-rich clades, even using an exhaustive local database. Combining two markers improves the accuracy of species identification but it would only marginally improve genus identification. Finally, we highlight some limitations of the BLAST algorithm as currently implemented and suggest possible improvements for barcoding applications.
Highlights
Resources for descriptive taxonomy and biodiversity inventories are much needed to manage the consequences of global change on the world’s biodiversity [1]
Taxonomic identification can be difficult, especially for tropical trees: individuals from the same species can vary morphologically according to their age or growing conditions, and closely related species can be morphologically very similar
Material for DNA extraction consisted of 5–50 cm2 of leaf tissue immediately dried in silica-gel. These samples were included in the African rain forest tree DNA samples collection of the Universite Libre de Bruxelles, in the Evolutionary Biology and Ecology laboratory
Summary
Resources for descriptive taxonomy and biodiversity inventories are much needed to manage the consequences of global change on the world’s biodiversity [1]. Taxonomic identification can be difficult, especially for tropical trees: individuals from the same species can vary morphologically according to their age or growing conditions, and closely related species can be morphologically very similar. Reproductive organs such as flowers and fruits are often needed to achieve accurate identification of morphologically similar species, but are often unavailable during field surveys [5]. The most common errors involved incorrectly splitting rare morphological variants of common species and incorrectly lumping geographically segregated morphologically similar species These errors had a measurable impact on ecological analyses. We test the ability of rbcL, matK and trnH-psbA plastid DNA markers to identify rain forest trees at two sites in Atlantic central Africa under the assumption that a database is exhaustive in terms of species content, but not necessarily in terms of haplotype diversity within species
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
