Taxonomic similarity does not predict necessary sample size for ex situ conservation: a comparison among five genera.

Effectively conserving biodiversity with limited resources requires scientifically informed and efficient strategies. Guidance is particularly needed on how many living plants are necessary to conserve a threshold level of genetic diversity in ex situ collections. We investigated this question for 11 taxa across five genera. In this first study analysing and optimizing ex situ genetic diversity across multiple genera, we found that the percentage of extant genetic diversity currently conserved varies among taxa from 40% to 95%. Most taxa are well below genetic conservation targets. Resampling datasets showed that ideal collection sizes vary widely even within a genus: one taxon typically required at least 50% more individuals than another (though Quercus was an exception). Still, across taxa, the minimum collection size to achieve genetic conservation goals is within one order of magnitude. Current collections are also suboptimal: they could remain the same size yet capture twice the genetic diversity with an improved sampling design. We term this deficiency the 'genetic conservation gap'. Lastly, we show that minimum collection sizes are influenced by collection priorities regarding the genetic diversity target. In summary, current collections are insufficient (not reaching targets) and suboptimal (not efficiently designed), and we show how improvements can be made.

A nuclear DNA barcode for eastern North American oaks and application to a study of hybridization in an Arboretum setting.

DNA barcoding has proved difficult in a number of woody plant genera, including the ecologically important oak genus Quercus. In this study, we utilized restrictionsite-associated DNA sequencing (RAD-seq) to develop an economical single nucleotide polymorphism (SNP) DNA barcoding system that suffices to distinguish eight common, sympatric eastern North American white oak species. Two de novo clustering pipelines, PyRAD and Stacks, were used in combination with postclustering bioinformatic tools to generate a list of 291 potential SNPs, 80 of which were included in a barcoding toolkit that is easily implemented using MassARRAY mass spectrometry technology. As a proof-of-concept, we used the genotyping toolkit to infer potential hybridization between North American white oaks transplanted outside of their native range (Q. michauxii, Q. montana, Q muehlenbergii/Q. prinoides, and Q. stellata) into a horticultural collection surrounded by natural forests of locally native trees (Q. alba and Q. macrocarpa) in the living collection at The Morton Arboretum (Lisle, IL, USA). Phylogenetic and clustering analyses suggested low rates of hybridization between cultivated and native species, with the exception of one Q. michauxii mother tree, the acorns of which exhibited high admixture from either Q. alba or Q. stellata and Q. macrocarpa, and a hybrid between Q. stellata that appears to have backcrossed almost exclusively to Q. alba. Together, RAD-seq and MassARRAY technologies allow for efficient development and implementation of a multispecies barcode for one of the more challenging forest tree genera.