On Thursday, December 8, 2017, Valentin Vasselon, a PhD student in Biodiversity, Ecology, Environment at the Centre Alpin de Recherche sur les Réseaux Trophiques des Écosystèmes Limniques (CARRTEL), will defend his thesis "Barcoding and bioindication: development of diatom metabarcoding for assessing river quality". The defense will take place at 9 am, in the UMR CARRTEL/INRA videoconference room, in Thonon-les-Bains.
thesis summary
Diatoms are microscopic unicellular algae that are excellent indicators of the ecological status of the environment in which they are found. As part of the Water Framework Directive (WFD), diatom communities are used to assess the quality of watercourses. For this purpose, quality indices based on species sensitivity to pollution are calculated from the composition and relative abundance of diatom taxa. Species identification is generally carried out using a microscope, which is not only complex, but can be time-consuming and costly when large numbers of samples have to be processed. A new method developed recently enables species to be identified not on the basis of morphological variability, but on the basis of genetic variability, using short DNA sequences (or DNA barcodes). Combined with high-throughput sequencing technologies, this molecular approach, known as metabarcoding, makes it possible to identify all the species present in an environmental sample, and to process several hundred samples in parallel.
These advantages make metabarcoding an attractive alternative to morphological identification for WFD purposes. Although several studies have demonstrated the ability of this approach to correctly identify diatom species found in environmental samples, the lack of reliability in quantifying the relative abundances of species limits the calculation of reliable indices and the use of metabarcoding as a bioindication tool. The aims of this thesis work were therefore (i) to identify and optimize the biases impacting the relative quantification of diatoms in metabarcoding; (ii) to apply the molecular approach on a large scale, on environmental samples from river networks in order to compare the quality assessments obtained by the morphological and molecular approaches.
First, we assessed the quantification biases associated with DNA extraction on pure diatom cultures and environmental samples. Although the choice of extraction method affected the quality and quantity of extracted DNA, as well as the relative abundances of certain species obtained by metabarcoding, community composition and quality scores were not significantly affected. We therefore decided to use the GenElute method, which produces the largest quantities of DNA at the lowest cost. Secondly, using qPCR experiments on pure diatom cultures, we showed that the number of copies of the rbcL gene (used as a DNA barcode) is proportional to cell biovolume, resulting in an overestimation of species with large biovolumes in metabarcoding. Based on this correlation, a correction factor was proposed and applied to metabarcoding data from artificial communities and environmental samples, making it possible to obtain relative species abundances comparable to those obtained by microscopy, and to improve the reliability of quality scores.
Finally, the application of the molecular approach to WFD river monitoring networks in Mayotte and mainland France has shown that metabarcoding is a faster and more cost-effective alternative to the morphological approach, while still providing a good assessment of river quality. Our work confirms that the molecular approach can be used to assess river quality. However, further studies are required before routine application and implementation in the WFD can be envisaged, particularly in terms of standardization and normalization of the methods used in the molecular approach.