Developing a DNA metabarcoding method to identify diet taxa in Neotropical foxes Provisionally Accepted

Natalia Mannise^{1, 2} Mariana Cosse^2* Gonzalo Greif³ Nadia Bou² Carlos Robello³ Susana Gonzalez² Andres Iriarte^1*

• ¹Facultad de Medicina, Universidad de la República, Uruguay
• ²Instituto de Investigaciones Biológicas Clemente Estable (IIBCE), Uruguay
• ³Pasteur Institute of Montevideo, Uruguay

Pampas and crab-eating foxes are medium-sized canids living in sympatry in the middle east of South America. Studies on the diet composition of these species provide a deep understanding of their ecological roles in the ecosystem structure and regulation. Using the metabarcoding technique, we analyzed the diet of both fox species in order to identify the vertebrate taxa included as food items. A fragment of the 12S ribosomal gene of the mtDNA was amplified using DNA extracted from 27 scat samples collected in south-central Uruguay during cold (June 2015) and warm (January -April 2016) seasons. A fox DNA blocking primer was designed to minimize the host amplicon products, and pooled samples were sequenced through paired-end reads (100 bp library) on a MiSeq Illumina Platform. The generated sequences were compared to a reference database built with sequences available in GenBank. In concordance with previous studies using traditional methods, we found that the most common food taxon were rodents. Qualitative differences in diet composition between both fox species were identified. Armadillo species were only found in pampas fox diet, while a greater variety of amphibians and birds were detected in crab-eating fox feces. Additionally, an innovative approach to differentiate between real and artifact sequences was employed. This method was based on comparing mutations at conserved and non-conserved positions within the secondary structure of the 12S rRNA, combined with network sequence reconstruction. Our results demonstrate the efficacy of the methodology in detecting the food species present in both fox diets, enabling the evaluation of intraspecific diversity among these species and facilitating the discarding of sequencing errors. This makes the methodology applicable to a wide range of studies.