Evaluating metabarcoding to analyse diet composition of species foraging in anthropogenic landscapes using Ion Torrent and Illumina sequencing

Marie-Amélie Forin-Wiart (1) , Marie-Lazarine Poulle (2) , Sylvain Piry (3) , Jean-François Cosson (3) , Claire Larose (4) , Maxime Galan (3)

1 DEPE-IPHC - Département Ecologie, Physiologie et Ethologie
2 PROTAL - Protozooses Transmises par l'Alimentation (Cryptosporidiose, Giardose et Toxoplasmose) : Mode de Contamination et Pathogénie (PROTAL) - EA 3800
3 UMR CBGP - Centre de Biologie pour la Gestion des Populations
4 Diana Pet Food Business

DNA metabarcoding of faecal samples is being successfully used to study the foraging niche of species. We assessed the ability of two benchtop high-throughput sequencing (HTS) platforms, to identify a large taxonomic array of food items from domestic cats Felis silvestris catus, including prey and human-related food taxa (pet food and leftovers leaving undetectable solid remains in faeces). Scats from a captive feeding trial (n=41) and from free-ranging individuals (n=326) were collected and analysed using a cytb mini-barcode in independent PCR replicates on the Ion PGM and the MiSeq platforms. Outputs from MiSeq were more sensitive and reproducible than those from Ion PGM due to a higher sequencing depth and sequence quality on MiSeq. DNA from intact prey taxa was detected more often (82% of the expected occurrences) than DNA from pet food (54%) and raw fish and meat (31%). We assumed that this variability was linked to different degree of DNA degradation: The Ion PGM detected significantly less human-linked food, birds, field voles, murids and shrews in the field-collected samples than the MiSeq platform. Pooling the replicates from both platforms and filtering the data allowed identification of at least one food item in 87.4% of the field-collected samples. Our DNA metabarcoding approach identified 29 prey taxa, of which 25 to species level (90% of items) including 9 rodents, 3 insectivores, 12 birds and 1 reptile and 33 human-related food taxa of which 23 were identified to genus level (75% of items). Our results demonstrate that using HT5 platforms such as MiSeq, which provide reads of sufficiently high quantity and quality, with sufficient numbers of technical replicates, is a robust and non-invasive approach for further dietary studies on animals foraging on a wide range of food items in anthropogenic landscapes.

Domaines: Sciences de l'environnement

Marie-Amélie Forin-Wiart, Marie-Lazarine Poulle, Sylvain Piry, Jean-François Cosson, Claire Larose, et al.. Evaluating metabarcoding to analyse diet composition of species foraging in anthropogenic landscapes using Ion Torrent and Illumina sequencing. Scientific Reports, 2018, 8 (1), pp.17091. ⟨10.1038/s41598-018-34430-7⟩. ⟨hal-02107284⟩