A pragmatic protocol for seed viability monitoring in ex situ plant genebanks

Abstract Seed viability monitoring is standard practice in genebanks and is performed to assess whether seeds of an accession are of good quality or need to be regenerated to prevent loss of the accession. A precise assessment of seed viability is labor-intensive, subject to sampling error and biological variation, while it also accelerates seed depletion. Balancing a proper viability assessment against costs is a conundrum, that needs to be resolved for genebanks to efficiently preserve germplasm. After the Centre for Genetic Resources, the Netherlands (CGN) concluded that the repeatability of their outsourced germination assessments was too low, it was decided to perform the germination assays in-house. We here present data that suggest an improved consistency in our new viability assays. However, improved viability testing cannot by itself resolve the assessment conundrum. To formulate a pragmatic testing and regeneration protocol, we set new testing intervals (based on empirical data) and redefined regeneration thresholds (by defining reliability requirements). Seed lots that are found to be close to- or below regeneration thresholds may forego regeneration at the recommendation of a curators’ experience and expert judgement. Curators can schedule such seed lots for future testing at shortened intervals. By allowing for such flexibility, it is likely that better judgement calls can be made during future testing, using fewer seeds, reducing the danger of losing accessions and of unnecessary regenerations.