Captive breeding programs for wildlife species typically rely on pedigrees to inform genetic management. Although pedigree-based breeding strategies are quite effective at retaining long-term genetic variation, management of zoo-based breeding programs continues to be hampered when pedigrees are poorly known. The objective of this study was to evaluate two options for generating single nucleotide polymorphism (SNP) data to resolve unknown relationships within captive breeding programs. We generated SNP data for a zoo-based population of addax (Addax nasomasculatus) using both the Illumina BovineHD BeadChip and double digest restriction site-associated DNA (ddRAD) sequencing. Our results demonstrated that estimates of allele sharing (AS) between pairs of individuals exhibited low variances. Average AS variances were highest when using 50 loci (SNPchipall = 0.00159; ddRADall = 0.0249), but fell below 0.0003 for the SNP chip dataset when sampling ≥250 loci and below 0.0025 for the ddRAD dataset when sampling ≥500 loci. Furthermore, the correlation between the SNPchipall and ddRADall AS datasets was 0.88 (95%CI = 0.84 – 0.91) when subsampling 500 loci. Collectively, our results indicated that both SNP genotyping methods produced sufficient data for accurately estimating relationships, even within an extremely bottlenecked population. Our results also suggested that analytic assumptions historically integrated into the addax pedigree are not adversely impacting long-term pedigree-based management; kinships calculated from the analytic pedigree were significantly correlated (p >>0.001) with AS estimates. Overall, our conclusions are intended to serve as both a proof of concept and a model for applying molecular data to the genetic management of captive breeding programs.

addax_Stacks_filesStacks 1.01 catalog output files with summary statistics. This file must be unzipped.Allele_Sharing_ResultsAverage pair-wise allele sharing estimates in csv format. The zipped file includes allele sharing estimates and the variance using 500 subsampled loci for both ddRAD and SNP chip approaches. One thousand bootstrap replicates with replacement were performed. Results with minor allele filtering (snpCHIP2 and ddRAD2) and without minor allele filtering (snpCHIPall and ddRADall) are included.