In Search of Species-Specific SNPs in a Non-Model Animal (European Bison (Bison bonasus))—Comparison of De Novo and Reference-Based Integrated Pipeline of STACKS Using Genotyping-by-Sequencing (GBS) Data

SIMPLE SUMMARY: The European bison is a dramatically low-diversified species, commonly analyzed using cattle-dedicated tools. Our aim was to compare two genotyping-by-sequencing (GBS) pipelines: de novo and reference pipeline, using the STACKS software and to reveal the maximum possible number of species-specific SNPs for our further project on European bison health. Therefore, we compared two genotyping-by-sequencing (GBS) pipelines: de novo (non-reference based) and a reference-based, using the STACKS software. We found a higher number of polymorphic loci from the reference pipeline in comparison to the de novo one. Next, we compared the results of the reference pipeline for the draft genome of European bison and completely annotated the Bos taurus genome. Higher numbers of polymorphic loci were revealed in European bison than in Bos taurus through the reference pipeline. We observed a possible effect of PCR duplicates on GBS data, as previously reported with the RADSeq approach. We recommend using a reference pipeline without PCR duplicates as a more efficient tool for species with low genetic diversity. ABSTRACT: The European bison is a non-model organism; thus, most of its genetic and genomic analyses have been performed using cattle-specific resources, such as BovineSNP50 BeadChip or Illumina Bovine 800 K HD Bead Chip. The problem with non-specific tools is the potential loss of evolutionary diversified information (ascertainment bias) and species-specific markers. Here, we have used a genotyping-by-sequencing (GBS) approach for genotyping 256 samples from the European bison population in Bialowieza Forest (Poland) and performed an analysis using two integrated pipelines of the STACKS software: one is de novo (without reference genome) and the other is a reference pipeline (with reference genome). Moreover, we used a reference pipeline with two different genomes, i.e., Bos taurus and European bison. Genotyping by sequencing (GBS) is a useful tool for SNP genotyping in non-model organisms due to its cost effectiveness. Our results support GBS with a reference pipeline without PCR duplicates as a powerful approach for studying the population structure and genotyping data of non-model organisms. We found more polymorphic markers in the reference pipeline in comparison to the de novo pipeline. The decreased number of SNPs from the de novo pipeline could be due to the extremely low level of heterozygosity in European bison. It has been confirmed that all the de novo/Bos taurus and Bos taurus reference pipeline obtained SNPs were unique and not included in 800 K BovineHD BeadChip.