Page 8

eFOOD-Lab_International_02_2014

analysis (see Fig. 1b). Nevertheless, experienced persons are able to extract the two sequences and identify both species. Minor species in samples with unbalanced ingredients can be detected when present > 10 % 2, otherwise the peaks in the chromatogram are getting too small. Mixtures with more than two species are not able to be analysed with the barcoding approach by Sanger sequencing as in the chromatogram a complex overlay of all sequence peaks is observed. Metagenome-Analysis/ Metabarcoding For the analysis of complete genomes, next generation sequencing (NGS) methods have been shown to be a useful tool, as they enable higher sequence outputs per time compared to Sanger sequencing. Since the first NGS technique was introduced in 2005 3, an exponential increase of published complete genomes can be observed. Additionally, it was recently shown, that fast sequencing of complete plastid genomes (ultrabarcoding) in plant breeding can be used for the characterisation of cultivars with different attributes 4. Apart from whole genome sequencing approaches, NGS also opened the door for the analysis of environmental samples containing several individuals. For such metagenomic studies, it is not necessary to sequence the complete genomic information of each present organism, but the organism is identified by the analysis of defined sequence regions. In many cases, the same sequence regions are used like for barcoding approaches (= metabarcoding). First publications concerning metagenomic/metabarcoding studies focused on microbial populations in environmental samples. In the meantime, also zoological studies use NGS for the analysis of faecal samples of herbivores as well as carnivores in order to get insights into their diets and habitats 5-6. In the food and feed sector, NGS is so far only used for the analysis of food-borne pathogens and starter cultures in fermentation processes 7-9. Whereas the used barcoding regions mentioned above have a length of not less than 500 bp, for the analysis of environmental, food and feed samples shorter sequences have to be used, as in such samples DNA degradation/fragmentation takes place. Optimal sequence regions for metabarcoding should be as short as possible but should provide enough sequence information for reliable species identification. Pompanon et al. 8 2/14 eFOOD-Lab international Fig. 2. Metabarcoding by NGS. Each species can be identified by the sequence output (cf. Sanger sequencing). Apart from that, the abundance of each species sequence (50.000 times for species/sequence 1, 35.000 times for species/sequence 2 and 15.000 times for species/sequence 3) roughly reflects the relative amount of the species in the sample mixture. (2012) give an overview of regions and primer pairs that were already used in metagenomic studies 6. If necessary, new regions have to be chosen and primer pairs have to be designed depending on the requirements of the sample analysis. Additionally, the setup of an own database with reference sequences might be considered. As known from microbial metagenomic studies, an estimation of the relative amount of each species present is the sample is possible using NGS 10. A prerequisite for such quantitative conclusions is that the universal primers do not cause a PCR bias by a favored amplification of one of the species DNA regions. In contrast to Sanger sequencing, in NGS approaches the parallel amplification of DNA from different species present in a sample is not performed in a DNA pool, but the DNA templates are separated either by fixation on a solid phase ("e.g. bridge amplification, Illumina systems) or by individual amplification using an emulsion PCR and magnetic beads (Roche 454 systems). In both cases, finally the amplified DNA templates from the different species – either captured on magnetic beads ("e.g. Roche 454) or clustered on the solid phase (Illumina) – can be individually sequenced several thousand times. Besides a qualitative species identification by alignment of a consensus sequence of each species cluster, a semi-quantitative estimation of the species amount is possible via the abundance of the species sequence in the sequence output of the NGS experiment (Fig. 2). As NGS platforms provide high sequence data outputs per run (up to 50 gigabases/ run; 11), which are normally not necessary for the metabarcoding analysis of one sample, multiplex analyses of several samples in one run are commonly performed. Short additional sequences (so called “barcodes” or “MIDs = multiplex identifiers”) are introduced to the primers and enable a final assignment of each individual sequencing read to the correct sample. NGS for food and feed routine analysis The applicability of NGS based metabarcoding for species identification in food and feed samples for routine analysis is strongly influenced by the costs and time for analysis and data evaluation. For cost calculation, purchase costs, running costs and the resulting costs/sample have to be considered. Current improvements of the different NGS technologies, the presence of competitive NGS platform providers, and the possibility of sample multiplexing led to an exponential decrease of the sequencing costs/megabase in the last decade (~ 5.000 US-$ in 2001, < 1 US-$ in 2010; www.genome.gov/sequencingcosts/). The total analysis time per sample is not only dependent on the running times of the instruments but also on the evaluation of the huge volume of data output, for which the use of bioinformatical tools is absolutely mandatory 5. As no tools are available for the demands of metabarcoding approaches in food and feed analysis, the adaption of existing bioinformatical tools from medical or environmental studies will be necessary. Analytical service provider, like Eurofins Genomics have realized the potential of NGSbased metabarcoding approaches for the analysis of complex food and feed samples. We have used our know-how and technical infrastructure to initiate the implementation of this technique in routine analysis. This will help to improve the authenticity control of mixed samples like fish and meat products, spices, as well as animal meals and starting cultures and to facilitate the detection of food fraud. Qualit y Mana gement


eFOOD-Lab_International_02_2014
To see the actual publication please follow the link above