MobiSeq to investigate relatedness within an amphibian species

We used MobiSeq as a new reduced representation library generation technique to investigate the relationship of 87 tadpoles (Dendrobates tinctorius) collected from 17 phytothelmata in the Nouragues Nature Reserve, French Guiana. We collected tail tissue by cutting a small part of the tail tip from 87 tadpoles. We included 5 duplicates to test the reliability of the method, giving a total sample size of 92. The duplicates were F01 with F46, F02 with F84, F03 with F74, F04 with F54 and F05 with F50. DNA was extracted using a salting-out method. We designed specific primers for repetitive elements as MobiSeq uses transposable elements (TEs) in DNA for target enrichment PCRs. Library preparation was performed at the Department of Interdisciplinary Life Sciences, Ecological Genetics Laboratory, University of Veterinary Medicine, Vienna. We used different approaches to call genotypes based on single nucleotide polymorphisms (SNPs). First, their sequences were trimmed and aligned de novo using the stacks pipeline, followed by filtering for quality and missingness. Second, we mapped the sequences to a newly derived reference genome of Dendrobates tinctorius, followed by filtering and calling SNPs with the stacks pipeline. Third, we mapped the sequences to a newly derived reference genome of Dendrobates tinctorius, followed by filtering and calling of SNPs using the ANGSD pipeline developed for MobiSeq. Comparison of these approaches revealed significant differences in genotype missingness and number of informative SNPs. In addition, the degree of relatedness between individuals is relatively low, which might be due to limited information given by sampling tadpoles only (no adults and therefore potential parents were sampled to increase full likelihood probabilities of relatedness). We observed strong effects of analytical pipelines. We recommend careful primer design for transposable elements, ideally with a reference genome of the focal species, and mapping of sequences to a reference genome. Ideally, the ANGSD pipeline is used for further processing of the sequences (trimming, mapping, SNP calling).