Diversity is a challenge
While the major compositional patterns in the Amazon can be analysed well with our dataset, a number of more detailed analyses, such as ‘how many tree species are there in the Amazon’, ‘which species are most threatened and where are they found’ and more detailed analyses of similarities across the region are hampered by an imprecise taxonomy of our data. The latter data also hampers fundamental theoretical questions on distribution and composition. From our current data we can estimate that at ~16,000 tree species inhabit the region but the actual number is unknown due to a large number of morpho-species, that cannot be identified to a valid species name.
Ecology needs a proper taxonomy
While this is by far the largest tree plot database of the tropics, covering practically all 1-ha plots of the Amazon, we have many individuals that cannot be identified to the species level (indets). We are also not entirely sure if individuals from different regions with the same name are in fact the same species, as there may be errors in identification, in Herbaria and Flora's.
On average the ‘indet level’ is low for most of our plots - in general 85 % of the morpho-species have a valid name for the plots. When plots are combined, however, the absolute number of indets increases dramatically. In our Guianas plot data-set to nearly 50% - so only half of the taxa, we can distinguish at species level can actually be given a proper name. The increase occurs as the common species are often repeated on plots but rare species accumulate, each plot brings a few new indets.
We believe that barcoding may solve most (but probably not all) of our identification problems. Our goal is to barcode all morpho-species of all our plots. This is an enormous task and will have to be carried out on close collaboration with all active plot owners.
The sequence data will also allow us to study phylogenetic diversity (even in the absence of valid names), phylogeography and character evolution in trees communities, to name a few. So the theoretical questions that can be tackled with the data also increases when a phylogeny of our species can be constructed.
Experience in barcoding plots has been built up in France (Jerome Chave, Chistopher Baraloto, Daniel Sabatier, Jean-François Molino), Ecuador (Juan Guevara) and RAINFOR.
Defining a barcode or NGS?
There is no universal plant barcode: chloroplast markers (rbcL, matK) have been proposed by BOLD. At the Plant DNA Barcode Project of STRI at least 3 plastid regions are sequenced (rbcl, matK, trnH-pbsA), and 1 neclear region (ITS). In French Guiana eight chloroplast markers, single or in combination, were able to identify c. 70% of the species.
Costs of DNA Barcoding
Prices of DNA barcoding have gone down considerably to a little over 5 euro for sequencing one region. One plate of 95 samples with 23 species for 4 regions amounts to roughly 4050 Euro or 175 per species (97 for the chemicals and sequencing, and 78 for quality control and digital processing)
For the ~500 individuals of our most recent sample this amounts to 16,000 euro. A similar amount will be needed for our new south Guyana samples.
To barcode all of the Guianas sample (~1600 spp) were are looking at 65,000 euro.
Barcoding all Amazonian trees – a major challenge
Finally, to barcode all Amazonian tree species this would be in the order of 6.5 million euros. And this is only 1 sample per species!
Barcoding all Amazonia trees is an ambitious project. We will be able to start this piecewise, and will first search for available DNA-Barcodes to avoid double work, but a full scale barcoding all 16,000 species will require major funding not only to barcode all species but also to locate proper material for all species. Any person or agency willing to fund part of this important work is asked to contact us at Naturalis Biodiversity Center.