The potential of seedbank digital information in plant conservation

Societal Impact Statement Seedbanks are vital for biodiversity conservation, but their potential remains underutilised due to a limited understanding of the intraspecific genetic diversity they hold. By leveraging digitised data associated with seedbank collections, such as sampling locations, number of maternal plants and seed traits, we can attempt the estimation of genetic variation and identify gaps in collections, enabling better prioritisation of species for conservation efforts. These advancements can inform policy targets like those of the Kunming‐Montreal Global Biodiversity Framework, promoting more effective conservation strategies. Digitisation and emerging machine‐learning technologies offer scalable, cost‐efficient solutions to enhance conservation knowledge, ensuring biodiversity resilience for future generations. Summary Seedbank collections hold significant untapped potential for advancing conservation science and practice, but the intraspecific genetic diversity (i.e. diversity within a species) stored in worldwide seedbank collections remains largely unknown, hindering the effective use of seeds for both informing and implementing in situ interventions. As producing genetic data is time‐consuming and expensive, other data associated with seedbank collections can greatly enhance our understanding of the genetic variation stored in seed collections when genetic data are unavailable. Information such as the location of sampling sites, estimated population size and the number of mother plants from which seeds were collected can facilitate the estimation of the genetic diversity captured in the collections. This information can also be used to estimate the sampling effort required to fill gaps in seedbank collections to better represent genetic diversity, through comparison with existing baselines from species where genetic diversity is characterised, and through simulations. Digitisation of the data associated with seedbank collections makes the approaches above practicable at scale. In addition, digital images of the seeds themselves may identify intraspecific phenotypic variation and can, therefore, be used to prioritise populations for future genetic studies. In this article, we explore the potential of digitised information made available by seedbanks for improving our understanding of the intraspecific genetic diversity preserved in collections. We describe possible improvements that might enhance the predictive power of digital information for genetic studies, and discuss the challenges and opportunities associated with these.