Assessing the viability of genebanked seeds from rare, wild plants native to the United States using the D.E.A.D. paradigm.

Premise: Genebanks must maintain viable seeds for decades. Seeds that germinate are clearly alive, but some seeds, often from wild populations, do not germinate because they are dormant, empty, aged, or damaged (D.E.A.D.). This work evaluates the effects of D.E.A.D. factors on genebanked seeds using a unique dataset to improve genebanking practices and standards for ex situ conservation of seed collections. Methods: Seeds from over 100 species were recently collected from the same populations as seeds that were genebanked decades ago. Germination proportion and speed were measured after applying various temperature, chemical, or seed coat abrasion treatments. Viability was further tested using vital staining of samples with a low germination proportion. Proportions of dormant, empty, aged, and damaged seeds were compared between seed cohorts. Results: Germination proportion and speed varied among samples, and cues to stimulate germination of dormant seeds were identified for individual species, leading to a positive correlation between viability metrics of germination and vital staining. Empty seeds primarily contribute to low germination in this study. Aging, indicated by lower and slower germination, was evident in several of the stored samples, compared to those that had been recently harvested. Discussion: This unique approach demonstrates the feasibility of genebanking seeds from diverse endangered plant species using freezer storage. Genebanking methods that are more relevant for crop seeds need to be modified when applied to seeds from wild populations because the sample sizes tend to be small and the seeds tend to germinate slowly and asynchronously.