Gene bank

Part of a series on

Conservation

Seed in glass jars in the cold store of the gene bank at Gatersleben, Germany, which registers 56 cannabis accessions (2010).

Cannabis conservation·Grow-out priorities·Populations

Threat Categories

Extinction

Genetic Threats

Genetic erosion·Genetic contamination·Introgression·Gene flow·Hybridisation·Population bottleneck·Genetic drift·Inbreeding depression

Habitat & Environmental Threats

Habitat loss

Eradication Tracking & Analysis

Eradication·Herbicide spraying of cannabis·Herbicide spraying in Mexico·Herbicide spraying in the United States·Herbicide spraying in Colombia·Herbicide spraying in Belize·Herbicide spraying in South Africa·US eradication abroad·Yukthiya·Cannabis eradication in Kirivong·Cannabis eradication in Fiji·Campaigns by country

International Enforcement

Cannabis prohibition·US eradication abroad·INCB

Legal & Policy

Farmers' rights·Access and benefit-sharing·Intellectual property·Single Convention on Narcotic Drugs·1988 Convention·Cannabis legalisation movement

Population Assessment

Minimum viable population·Population viability analysis·Effective population size

Conservation Monitoring

Population monitoring·Genetic monitoring

Conservation Methods - Ex Situ

Ex situ conservation·Gene bank·Cryopreservation·In vitro conservation·Core collection·Seed storage

Conservation Methods - In Situ

In situ conservation·On-farm conservation·Genetic reserve·Community seed bank·Feral and ruderal populations

Genetic Conservation

Germplasm·Genetic resources·Crop wild relative·Conservation genetics

Sampling & Collection

Germplasm collecting·Germplasm sampling

Preservation Strategies

Conservation prioritisation·Complementary conservation

Recovery & Restoration

Reintroduction·Genetic rescue

Conservation Science

Landrace conservation

Portal·Category

A gene bank is an institution that keeps collections of plant material alive, as seed, live plants or tissue, for long-term conservation and for use by breeders and researchers.^: 2 The Food and Agriculture Organization (FAO) spells the word "genebank" in its standards, which set out how a sample should be received, stored, tested, regrown, documented, supplied and duplicated. The standards are voluntary.^: 3

Cannabis accessions sit in public gene banks in Europe, Asia, North Africa and North America. On 6 November 2026 the hemp collection at Geneva, New York, held 696 and the Vavilov Institute in Saint Petersburg held 491. In the cannabis trade a "seed bank" is a business that sells seed. Such a business is a seed vendor, not a gene bank.

Three meanings of "seed bank"

In the FAO standards a seed bank is one of three kinds of gene bank, beside the field gene bank and the in vitro or cryopreservation gene bank.^: 3 John McPartland and Ernest Small used the words in this sense in 2020, when they called preservation "in seed banks" an immediate priority for cannabis.

In the cannabis trade the words name a seller. The Seed Bank, the business of Nevil Martin Schoenmakers in the Netherlands, sold seed by mail to growers in the United States in the 1980s. Feminised seed is popular among growers and "commercial seed banks". Breeding is common practice among both. A vendor's lots are stock for sale. A gene bank's accessions are kept for conservation and supplied for research and breeding, with standards for their identity, viability and genetic integrity.^{: 2, 8–10} Seed can pass from the first to the second: 190 accessions in the National Plant Germplasm System of the United States came from Angus Pretty, founder of the seed vendor The Real Seed Company.

In plant ecology a soil seed bank is the viable seed lying in the soil. Its part in feral stands of cannabis is treated under Feral and ruderal Cannabis. Johannes Engels and Andreas Ebert use "community gene banks" for local community seed banks, which they suggest could strengthen local conservation and tie it to the national system.

From one institute to a global network

The first gene banks date from 1920, with the Vavilov Institute. A potato collection at Cambridge followed before the Second World War and a long-term seed store at Fort Collins, Colorado, in 1958. National centres appeared on every continent in the 1950s and 1960s. In the 1970s the International Board for Plant Genetic Resources began a register of gene banks prepared to hold long-term base collections.

By the FAO's report of 2010, about 1,750 gene banks and collections held about 7.4 million accessions. Only 25 to 30 per cent of them were estimated to be distinct. Seed made up about 96 per cent of all accessions. Of the accessions analysed, 24 per cent were landraces. These totals cover all crops. The crops under the multilateral system of the plant treaty are listed in its Annex I. Cannabis is not among them.^: 46–49

What a gene bank does with a sample

A sample should come with proof that it was legally acquired and with passport data giving its origin. Seed collected in the field should come from 30 to 60 plants, depending on how the species breeds.^: 18 The sample becomes an accession, uniquely identifiable. The gene bank must keep that identity through storage, regrowing and supply.^{: 8, 162}

Seed is dried at 5–20 °C and 10–25 per cent relative humidity, then sealed. A base collection holds the most original samples at −18 ± 3 °C and 15 ± 3 per cent relative humidity. These are not meant for distribution. An active collection at 5–10 °C is the stock that supplies users.^: 24 The Geneva collection stores its seed in a cold room at −20 °C. How long cannabis seed lasts under such conditions is treated under ex situ conservation.

The first germination test should come within 12 months of receipt. Most cultivated crops should start above 85 per cent. Later tests fall due at one-third of the time predicted for viability to drop to 85 per cent of the starting value, at most after 40 years.^: 30 An accession is regrown when viability falls below that mark or when the seed left would not cover three sowings of a representative population.^: 36 Geneva planned to regrow at 70 per cent viability or fewer than 2,000 seeds. Yong-Bi Fu expects genetic change in any accession kept over many cycles of regrowing, through genetic drift and selection.

Every accession should have passport data in a database. About 60 per cent should be characterised within five to seven years of acquisition. Evaluation trials should run in at least three environments over at least three years.^{: 41, 45, 50} A gene bank can also offer a core collection, a subset of about 10 per cent of its accessions chosen to hold most of its diversity. Seed goes out under national law and the relevant treaties, usually as 30 to 50 viable seeds.^: 53

The entrance to the Svalbard Global Seed Vault, which holds safety duplicates from gene banks elsewhere (2015).

A safety duplicate of every original accession should sit in a distant place under the same or better conditions. Each gene bank should plan for power cuts, fire, floods and earthquakes.^{: 57, 61} War drove the ICARDA gene bank from Aleppo to Lebanon and Morocco in 2016. Part of its collection was restored from duplicates. Floods reached the national gene banks of Thailand in 2011 and the Philippines in 2006. Fire reached the Philippine bank in 2012. The Svalbard Global Seed Vault in Norway held more than 1.4 million seed samples from 134 institutions in 2026. Depositors keep ownership of their boxes and only the depositor can withdraw them. The Geneva collection planned its backups for Fort Collins.

Living collections

Field gene banks keep live plants. They are the usual choice for seed that cannot be dried and chilled, for plants that set little seed and for those propagated vegetatively. The term also covers pots and trays in greenhouses or shade houses.^: 4 A field site should resemble the climate and soil where the plants came from and should have secure land tenure. A map shows the place of each accession. Each accession needs a duplicate at a second site or in another form of storage where possible.^{: 66, 76, 109} Such collections are costly to keep, so the FAO advises building complementary collections in vitro or in cryopreservation.^: 13

Slow growth in vitro serves the short and medium term, while cryopreservation in liquid nitrogen serves the long term.^: 4 The University of Mississippi keeps indoor-grown collections of living cannabis plants and has been working to freeze shoot tips of all the varieties it holds. Growers keep clonal lines as mother plants. Their decline over time is treated under ex situ conservation.

Gene banks that hold cannabis

The west building of the Vavilov Institute in Saint Petersburg (2010).

Cannabis accessions by holder, with the date of each count

Holder	Country	Accessions	Of which landrace or traditional cultivar	Date of count	Source
Plant Genetic Resources Unit, National Plant Germplasm System, Geneva, New York	United States	696	215	6 November 2026
Vavilov Institute, Saint Petersburg	Russia	491	302 (4 October 2026)	6 November 2026
INRA Maroc, Rabat	Morocco	106	106	6 November 2026
Suceava Genebank	Romania	97	30	6 November 2026
Leibniz Institute of Plant Genetics and Crop Plant Research, Gatersleben	Germany	56	19	6 November 2026
CREA Research Centre for Cereals and Industrial Crops, Bologna	Italy	9	0	6 November 2026
Nordic Genetic Resource Center	Sweden	3	0	6 November 2026
China National GeneBank; Plant Genetic Resources Program, Pakistan	China; Pakistan	2 each	not stated	6 November 2026
Lebanese Agricultural Research Institute	Lebanon	1	1	6 November 2026
Institute of Natural Fibres and Medicinal Plants, Poznań	Poland	139 to about 150	not stated	2015
Yunnan Academy of Agricultural Sciences, Yunnan	China	about 350	not stated	reported by 2016

In October 2026 the Vavilov Institute's accessions came chiefly from Ukraine (99), Russia (74), the former East Germany (72), China (42), the former Yugoslavia (35) and the European part of the former Soviet Union (35), with 11 from Kazakhstan and two from Afghanistan. All were stored short term. In 1998 the collection held nearly 500 accessions of known origin.

The Geneva collection dates from 2021. Of its 696 accessions, 215 have the status of landrace, 133 of wild material and 109 of cultivar. Another 182 have an uncertain status. On 6 November 2026, 91 could be requested and 605 were marked not available. Hundreds of wild lots arrived as small samples of varying quality, for which "seed increases must be done to produce seeds available for distribution". Seed that is available goes free of charge to research projects. None goes to home gardens. A further 159 entries in the American database are historic, among them PI 5021, which "no longer exists and can not be requested". Another 18 are vouchers in its Plant Variety Protection collection.

All 106 accessions at INRA Maroc have the status of traditional cultivar or landrace. All but one come from the Rif Mountains. Suceava's 97 are mostly Romanian (90) and include 29 inbred lines, 19 improved cultivars and 17 wild samples. The Polish collection held about 150 accessions by the institute's own paper of 2015 and 139, mostly European, by the 2016 review of Matthew Welling and colleagues.

Other cannabis collections in 2016 included a National Cannabis Collection at Tápiószele in Hungary, one in Prague and the former collection of the Centre for Plant Breeding and Reproduction Research at Wageningen, which held 156 accessions. Ecofibre in Australia and HortaPharm in Amsterdam held company collections. The availability of germplasm from any of them may vary over time. Losses from North American collections in the later 20th century are treated under ex situ conservation. Access under drug-control law is treated under germplasm.

Reading the counts

An accession count is a count of numbered samples. It does not measure diversity. The counts in the table cannot be added, because one lot can sit in two gene banks: the Gatersleben institute's CAN 22, from Georgia, is also G 33598 in the American system. The Global Information System of the FAO holds what each holder registers. On 6 November 2026 it held 276 cannabis accessions, none of them from the American or Russian collections, with no country of origin given for 132.

The status "landrace" is a label given by the holder or donor. McPartland and Small held that germplasm of the Central and South Asian landraces kept outside their centres of diversity had been contaminated by crossbreeding. Coverage is uneven. Of 808 accessions reported from collections worldwide in the 2016 review, 58 came from China. Its authors named East Asia a priority for collecting. An absence from these collections is no evidence that the plant has gone from a place.

See also

• Community seed bank
• Core collection
• Cryopreservation
• Ex situ conservation
• Germplasm
• Grow-out priorities
• Seed lots
• In situ conservation
• Seed storage