Join on WhatsApp
Join on Telegram
Rabat – Scientists estimate that nearly two in five plant species face extinction and over 31% of the world’s 390,000 known vascular plants are under threat. As our planet faces accelerating climate change and widespread habitat loss, scientists and conservationists are racing against time to safeguard the genetic diversity of plants.
Seed banks and living collections in botanic gardens, known collectively as ex situ conservation, offer a crucial backup against the decline or extinction of species in the wild. Yet despite the existence of more than 1,700 gene banks worldwide housing some seven million seed samples, recent analyses suggest that less than 20% of wild plant species are adequately represented ex situ.
This gap puts both ecosystems and human food systems at risk, since crops and their wild relatives provide the genetic building blocks for future resilience.
At University Mohammed VI Polytechnic (UM6P), a research university in Morocco known for its applied science and open-innovation model, doctoral researcher Marwa El Graoui, based at the College of Agriculture and Environmental Sciences (CAES), has developed an open-source framework that supports efforts to fill this critical gap. Her method combines publicly available data, geospatial modeling, and a simple software package to empower anyone to assess how well genetic diversity is being saved, and to prioritize which species need urgent attention.
Vigna as a conservation model
El Graoui’s interest in developing a new conservation metric grew out of her doctoral thesis at UM6P, where she spent the last couple of years mapping wild Vigna species across Africa. “It just made sense to choose that genus,” she explains. “It has a lot of species, each with a different geographical distribution, so we were able to find areas of improvement in our methods because we would stumble upon species with different spatial contexts”.
Although her focus was on Vigna, El Graoui emphasizes that the approach applies to any plant group for which occurrence data and climate layers exist.
At the heart of her framework is the idea that geographic and environmental distance between samples can serve as proxies for genetic differentiation, which is crucial when direct genomic data are lacking.
Drawing on the classic theory of isolation by distance by Sewall Wright, which holds that populations farther apart tend to diverge genetically, and on the concept of environmental differentiation, El Graoui translated both spatial separation and habitat differences into comparable distance values. This allowed her to estimate how much genetic diversity remains represented among stored seeds, even for species without published genetic sequences.
Defining a species’ range accurately proved equally challenging. Simple circular buffers around known collection points often under-estimate distribution, while species distribution models (SDMs) can over-predict suitable habitat, including areas where the species has never been recorded.
To strike a balance, El Graoui and her co-author, Professor Robert J. Hijmans, devised an “adjusted range” method. They applied inclusion buffers to capture documented occurrence zones, then overlaid exclusion buffers to remove implausible areas predicted by the SDM. “We called it the adjusted range… we didn’t want to be either under-predicting or over-predicting,” El Graoui notes.
An open-source toolkit in R
Working with open data from repositories like GBIF brought both opportunity and unexpected hurdles. While thousands of georeferenced records were readily available, El Graoui found that many samples flagged as ex situ collections were actually marked as living specimens or wild occurrences. “This small mistake meant we thought species were well conserved, when in fact they were very poorly conserved in ex situ repositories,” she recalls. Correcting these mislabels dramatically shifted the conservation status of nearly every Vigna species in her study. This underscores the urgent need for improved data standards in seed banks and biodiversity platforms.
To make the method accessible beyond her own team, El Graoui packaged the entire workflow into an open-source R library. Users need only supply occurrence coordinates and climate variables, which are readily downloadable through the package, to run a series of functions that delineate the adjusted range, calculate geographic and environmental distances, and generate a final conservation score for each species. “You don’t need to be an R expert,” El Graoui assures. Anyone can inject their occurrence data and climate layers, run a few functions, and get a final score that prioritizes which species need urgent seed collection.
This commitment to open science and practical impact reflects the broader vision of UM6P, where digital tools, African data, and actionable frameworks converge to address continental and global challenges.
Applying this tool to African Vigna species revealed several surprising insights. Some taxa assumed to be secure turned out to be under-represented in seed banks, while others concentrated in narrowly defined hotspots offered opportunities to capture a wide sweep of genetic diversity in a single field expedition. These data-driven results enable botanists and policymakers to allocate limited resources more effectively, targeting those collections that will yield the greatest genetic return on investment.
Despite its advances, El Graoui acknowledges that the framework has room for growth. Incorporating actual genetic distance data, where available, would allow calibration of the geographic and environmental proxies. Factoring in species-specific traits, such as seed dispersal mechanisms or pollination modes, could refine distance estimates further. “For example, if seeds are wind-pollinated and very light, the genetic distance per kilometer will be smaller than for heavier, insect-pollinated seeds,” she observes.
Moreover, she and her colleagues from UM6P, CIRAD, and UC Davis urge the community to standardize data labels across repositories, so that ex situ and in situ records are never confused, and to develop shared platforms for depositing trait and provenance information.
UM6P’s research culture—rooted in African field contexts but globally networked—offers fertile ground for this kind of innovation. From the university’s biodiversity stations to its computational resources, the infrastructure enables early-career scientists like El Graoui to build tools that are not just publishable, but usable.
El Graoui is enthusiastic about the open nature of the project and what the future will bring. There is already a lot of work being done in this field, but each person can build on the other, she says, hoping others will expand the method and apply it to different genera and regions.”
As the global community moves toward ambitious conservation goals, such as the Kunming-Montreal Biodiversity Framework’s aim to protect 30% of all land and water for nature by 2030, equally robust targets for ex situ genetic conservation must follow.
