Germplasm sampling
More actions
Germplasm sampling is the choice of how many plants a collector takes seed or other material from, which plants, how much from each and how many populations to visit when gathering germplasm for a gene bank or other ex situ collection. Its aim is a sample that carries the genetic variation of the population into storage, within the time and money a collecting mission has.[1]: 1–2, 22 The gene bank standards of the Food and Agriculture Organization (FAO) ask for seed from at least 30 to 60 plants in a sample, depending on how the species breeds.[2]: 18
In 2022 and 2023 Ademola Aina and colleagues took seed from 1,815 female plants in 88 feral populations of Cannabis across 12 states of the United States, with the aim of stocking the hemp collection of the National Plant Germplasm System.[3]
Allele targets and sample size
Many of the standard sampling rules come from simple probability equations first worked out in the 1970s.[4] D. R. Marshall and A. H. D. Brown proposed in 1975 that a sample should hold at least one copy of 95 per cent of the alleles that occur in the population at a frequency above 0.05. A random sample of 59 unrelated gametes meets that target. That is 30 plants in a species that mates entirely at random and 60 in one that only self-pollinates, hence the FAO's range of 30 to 60.[2]: 20 Austin Koontz and colleagues call the 95 per cent threshold "typical (if somewhat arbitrary)".[5]
Rarer alleles and more genes raise the number. Probability models published by José Crossa and colleagues in 1993 put the sample at 160 to 250 plants of a random-mating population to catch alleles of frequency 0.05 or more at each of 150 loci with a probability of 90 to 95 per cent. To keep alleles of frequency 0.001 with a probability of 0.9999, the same models need 533 to 1,708 plants. The frequency of the rarest allele sought moves the figure more than the number of alleles or loci does.[1]: 7 The models treat loci as independent, an assumption Crossa and Roland Vencovsky call unrealistic. They also leave out how the population is structured and how it mates.[1]: 3
A second measure counts the genetic drift a sample carries forward. The effective population size of a seed sample measures how well the sample stands for its population. Collecting, storage, regrowing and the drawing of seed for each new generation can each lower it.[1]: 2 When seed is gathered from a large population, the effective size of the sample depends above all on the number of mother plants. In Crossa and Vencovsky's example, 1,000 seeds taken in equal numbers from 100, 200 or 500 mother plants give effective sizes of 308, 500 and 800. The same 1,000 seeds drawn at random from a bulk of the harvest give 286, 447 and 667.[1]: 10
Mother plants, seeds and sites
A collector sets three numbers: seeds per mother plant, mother plants per population and populations.[1]: 22 For a single population Crossa and Vencovsky advise against taking much seed from a few plants. They recommend seed from at least 25 mother plants in a cross-pollinating species and 50 in a self-pollinating one, with an equal or nearly equal number of seeds from each plant in every species.[1]: 21 The mother plants should be unrelated, so they need to stand some distance apart. Studies of how far pollen travels can set that distance.[1]: 10 Collecting forms should note the number of plants sampled. A second visit to the same site can add variation present at another time.[2]: 20
Collectors in practice tend to visit a few populations and take dozens to hundreds of seeds from each plant. Fewer seeds from more mother plants capture more diversity. Where genetic work is not possible, Sean Hoban and colleagues recommend "few seeds per plant but many unique maternal plants" from most or all of the wild populations.[6]
Where a species is split into partly separate populations, the effective size of a sample depends strongly on how many of them are sampled. This holds even when the populations differ by only 2 per cent of the total diversity, an FST of 0.02. With an FST of 0.05, an effective size of 1,000 needs about 100 populations. Neighbouring populations tend to resemble each other. Crossa and Vencovsky advise collecting from as many sites as possible when nothing is known of the population structure. They also advise choosing the populations least alike.[1]: 19–21
The National Tree Seed Project of the Millennium Seed Bank Partnership had collected European ash seed from 659 trees at 59 sites in the United Kingdom. Hoban, Simon Kallow and Clare Trivedi modelled those collections. For a goal of one copy of each allele, the gain from further sampling levelled off at about 35 populations, 10 to 30 trees per population and 30 seeds per tree. A goal of 50 copies of each allele needed more than 150 populations.[4]
How large a collection must be
Hoban and colleagues genotyped wild plants and garden collections of 11 threatened trees and shrubs in five genera whose seed cannot be kept in a seed bank. Leaving out alleles found only once or twice, the collections held 40 to 95 per cent of the wild alleles. Resamples of the wild data showed that 24 to 207 plants would capture 95 per cent, depending on the species and on whether the rarest alleles were counted. Species of the same genus needed different numbers: one Zamia needed almost twice as many plants as the other. For a species about which nothing is known, the authors put the minimum at about 30 to 200 plants sampled at random.[6]
The collections held less than random samples of the same size would have. Plants of Zamia lucayana in gardens held 95 per cent of the alleles in 244 plants. Random sampling would have reached that level with 44. The authors call the difference the "genetic conservation gap". Among its causes they name gardens exchanging material among themselves without new collecting from the wild, the difficulty of reaching every wild population, collecting in one year only and selection during seed cleaning and cultivation.[6]
These minimums assume that every plant survives. If 82 plants are needed and a tenth of the seeds grow to adulthood, 820 seeds have to be collected. Hoban and colleagues put most collections at several times the minimum, of the order of 150 to 1,000 plants.[6] The estimates also depend on the markers used. In simulations by Koontz, Gavin Salas and Hoban, estimates from single-nucleotide polymorphisms came out twice as large as those from microsatellites. The uncertainty around each estimate was wide. The three authors argued that the 95 per cent standard should be revisited.[5]
Where to collect
Collecting of crop wild relatives is usually planned with no published data on the genetic diversity of the target species or with data that fall short. Markers in common use mostly measure neutral variation. Its levels often correlate poorly with variation in adaptive genes, the kind breeders most want.[7]: 4 Collecting for a breeding programme follows a different logic. The focused identification of germplasm strategy picks the populations exposed to environmental stresses of interest to breeders, to raise the chance of finding a trait.[7]: 5 A sample gathered for a trait may not represent its population.[2]: 10
Taking seed without harming the population
The FAO standards ask collectors to be sensitive to the depletion of the population they collect from. They cite the seed collecting manual of the European Native Seed Conservation Network, under which a collection may not take more than 20 per cent of the seed available in a population.[2]: 20 For wild and rare species, where seed may be scarce, the standards accept smaller samples.[2]: 22
The Center for Plant Conservation in the United States sets a lower limit for rare wild plants: no more than 10 per cent of a plant's or a population's seed output in a season and collection in no more than 10 years out of 90. The limit rests on models of 25 populations of 22 species by Menges and colleagues, published in 2004. In those models, frequent light harvests carried less risk of extinction than rare heavy ones. The centre names short-lived plants without a persistent soil seed bank as the most sensitive and considers a population of fewer than 50 plants at higher risk to begin with. Its ideal is 3,000 seeds from 50 mother plants, kept as separate maternal lines. Where a population cannot yield that, the 10 per cent limit applies.[8]
Separate sexes
In a dioecious species, seed comes from female plants. The share of males that fathered it depends on how far pollen travels and on the structure and size of the population. Crossa and Vencovsky give an example of 200 seeds from 100 females crossed at random with 100 males, in a population of 10,000 of each sex. With equal numbers of seeds from each female the effective size is 115, below the number of seeds. It falls further as the sex ratio moves away from equal. The sex of seeds and seedlings is generally hard or impossible to tell, so Crossa and Vencovsky suggest estimating the ratio among adult plants and taking equal numbers of seeds from each female. Whether 5 per cent or 0.1 per cent of the adults are sampled makes little difference to the effective size.[1]: 17
Cannabis
Cannabis is dioecious and wind-pollinated.[9] The hemp collection of the National Plant Germplasm System at Geneva, New York, dates from 2021.[3] The collection held more than 500 accessions in August 2023, 128 of them feral.[10]
Aina and colleagues found feral stands through the iNaturalist application and by word of mouth. They collected in late summer and early autumn under a federal hemp producer licence, taking seed and floral tissue plant by plant. A further population came from the Geneva collection and two from Colorado from Colorado State University. Collecting was at times hard because of the stigma attached to cannabis and because stands grew along interstate highways, on private land or in dense vegetation. The authors asked for wider collecting in the West, South and Northeast of the country. Of the plants, 760 were genotyped. They fell into five genetic groups that largely followed their state of origin.[3]
Tori Ford and colleagues built ecological niche models of Midwestern feral cannabis from digitised occurrence records, to guide its collection. They gave the highest priority to Indiana, Illinois, Kansas, Missouri and Nebraska, where the models projected a loss of suitable habitat under climate change. Michigan, Minnesota, Ohio and Iowa came lower.[10]
From 1993 the Vavilov Institute set 1,000 plants as the standard for each regrowing of a stored cannabis accession, to keep nearly all the alleles of its gene pool.[11]
See also
References
- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 Crossa, J.; Vencovsky, R. (2011). "Basic sampling strategies: theory and practice". In Guarino, L.; Ramanatha Rao, V.; Goldberg, E. (eds.). Collecting Plant Genetic Diversity: Technical Guidelines. 2011 Update. Rome: Bioversity International. ISBN 978-92-9043-922-6.
- ↑ 2.0 2.1 2.2 2.3 2.4 2.5 Food and Agriculture Organization of the United Nations (2014). Genebank Standards for Plant Genetic Resources for Food and Agriculture (PDF) (Revised ed.). Rome: FAO. ISBN 978-92-5-108262-1.
- ↑ 3.0 3.1 3.2 Aina, Ademola; Wenger, Jonathan P.; Stanton, Eliot; Majumdar, Chandrani Gon; ElSohly, Mahmoud; Weiblen, George D.; Ellison, Shelby (1 July 2025). "Genetic diversity, population structure, and cannabinoid variation in feral Cannabis sativa germplasm from the United States". Scientific Reports. 15 (1) 20423. doi:10.1038/s41598-025-07912-8. PMC 12215616.
- ↑ 4.0 4.1 Hoban, S.; Kallow, S.; Trivedi, C. (2018). "Implementing a new approach to effective conservation of genetic diversity, with ash (Fraxinus excelsior) in the UK as a case study". Biological Conservation. 225: 10–21. doi:10.1016/j.biocon.2018.06.017.
- ↑ 5.0 5.1 Koontz, A.C.; Salas, G.D.; Hoban, S.M. (2026). "Genetic marker type impacts ex situ conservation minimum sample size estimates and their variance". Conservation Genetics. 27 (1) 3. doi:10.1007/s10592-025-01730-1.
- ↑ 6.0 6.1 6.2 6.3 Hoban, S.; Callicrate, T.; Clark, J.; Deans, S.; Dosmann, M.; et al. (2020). "Taxonomic similarity does not predict necessary sample size for ex situ conservation: a comparison among five genera". Proceedings of the Royal Society B: Biological Sciences. 287 (1926) 20200102. doi:10.1098/rspb.2020.0102. PMC 7282903.
- ↑ 7.0 7.1 Parra-Quijano, M.; Iriondo, J.M.; Torres Lamas, E. (2011). "Strategies for the collecting of wild species". In Guarino, L.; Ramanatha Rao, V.; Goldberg, E. (eds.). Collecting Plant Genetic Diversity: Technical Guidelines. 2011 Update. Rome: Bioversity International. ISBN 978-92-9043-922-6.
- ↑ Maschinski, J.; Walters, C.; Murray, S.; Kunz, M.; Schneider, H.; et al. (2020). "Collecting seeds from wild rare plant populations". CPC Best Plant Conservation Practices. Center for Plant Conservation. Retrieved 9 November 2026.
- ↑ Welling, M.T.; Shapter, T.; Rose, T.J.; Liu, L.; Stanger, R.; King, G.J. (2016). "A belated green revolution for Cannabis: virtual genetic resources to fast-track cultivar development". Frontiers in Plant Science. 7 1113. doi:10.3389/fpls.2016.01113.
- ↑ 10.0 10.1 Ford, Tori; Aina, Ademola; Ellison, Shelby; Gordon, Tyler; Stansell, Zachary (July 2024). "Utilizing digitized occurrence records of Midwestern feral Cannabis sativa to develop ecological niche models". Ecology and Evolution. 14 (7) e11325. doi:10.1002/ece3.11325. PMC 11239322.
- ↑ Lemeshev, N.; Rumyantseva, L.; Clarke, R.C. (1995). "Report on the maintenance of hemp (Cannabis sativa L.) germplasm accessioned in the Vavilov Research Institute Gene Bank – 1994". Journal of the International Hemp Association. 2 (1): 10–13.