Introduction
The evolution of crop plants can be described as a two-step process, where the first step is the domestication of wild plants and the second step the improvement of primitive landraces into cultivars by plant breeding. Both these two steps are associated with losses of genetic diversity, known as the domestication and improvement bottleneck (Tanksley and McCouch, Reference Tanksley and McCouch1997). Recent research, involving selection screens, quantitative trait loci (QTL) mapping and allele mining, has begun to reveal the molecular genetics behind the evolution of crop plants (reviewed by Doebley et al., Reference Doebley, Gaut and Smith2006). Cloning the superior alleles that contributed to the domestication and improvement process has been of interest not only to the research community but also to plant breeding companies. Obviously, one key to future crops lies in understanding the evolutionary history of crops from the past.
One obstacle for conducting this type of survey can be the lack of wild and unimproved plant material. Since the introduction of improved cultivars around 1900, landraces and obsolete cultivars have rapidly become extinct (FAO, 1997). In northern Europe only a small fraction of the landraces that were the basis for agriculture for thousands of years has been preserved in gene banks. Furthermore, landraces and obsolete cultivars maintained by breeding companies and gene banks risk losing genetic integrity during regenerations (Steiner et al., Reference Steiner, Ruckenbauer and Goecke1997; Börner et al., Reference Börner, Chebotar and Korzun2000; Chebotar et al., Reference Chebotar, Röder, Korzun and Börner2002).
As a supplement to living material, specimens maintained in museums, for example, can be used for molecular genetic analyses. Although research on aged and ancient DNA has focused very much on humans and other mammals, attention to plants has increased lately (reviewed by Gugerli et al., Reference Gugerli, Parducci and Petit2005). Seeds are a common plant material for the analysis of aged and ancient DNA, and DNA has been isolated and analysed successfully from seeds of up to 4000 years of age, from different species (Blatter et al., Reference Blatter, Jacomet and Schlumbaum2002; Freitas et al., Reference Freitas, Bendel, Allaby and Brown2003; Jaenicke-Després et al., Reference Jaenicke-Després, Buckler, Smith, Gilbert, Cooper, Doebley and Pääbo2003; Manen et al., Reference Manen, Bouby, Dalnoki, Marinval, Turgay and Schlumbaum2003; Walters et al., Reference Walters, Reilley, Reeves, Baszczak and Richards2006; Lia et al., Reference Lia, Confalonieri, Ratto, Hernández, Alzogaray, Poggio and Brown2007; Lister et al., Reference Lister, Bower, Howe and Jones2008). These seeds have been found in archaeological excavations or, for the more recent samples, in buildings or even herbaria. A severe constraint to the use of non-viable historical material is that information attached to the material, such as dating, cultivar name and growth location, is often vague or lacking (Jones et al., Reference Jones, Lister, Bower, Leigh, Smith and Jones2008). Another problem lies in the number of individuals available to sample, which is often too small to gain an understanding of population genetic aspects of the crop. In contrast to archaeological and herbaria samples, seed collections are more likely to contain plant material from many individuals, as well as having associated information that increases the strength of conclusions that can be drawn. An additional benefit of seed collections in comparison to herbaria is that DNA seems to degenerate more slowly in seeds than in vegetative plant parts (Walters et al., Reference Walters, Reilley, Reeves, Baszczak and Richards2006; Lister et al., Reference Lister, Bower, Howe and Jones2008).
In Europe several herbaria and seed collections of crop plants from the 19th and early 20th century have been preserved (Jones et al., Reference Jones, Lister, Bower, Leigh, Smith and Jones2008). Perhaps, best recognized is the Percival Collection of wheats from the 1920s stored at the Natural History Museum in London, UK (Morrison, Reference Morrison, Caligari and Brandham2001). Seed material from this collection has been utilized successfully to assess DNA preservation (Lister et al., Reference Lister, Bower, Howe and Jones2008). In the Museum of Vänersborg, Sweden, 1800 seed samples from an agricultural exhibition in 1880 are stored (Johansson et al., Reference Johansson, Aarsrud and Öberg2003). Seeds from this collection have been used for amplified fragment length polymorphism (AFLP) analyses of a Beta vulgaris cultivar, where the aged seeds were compared with homonymous plants preserved ‘on-farm’, and it was concluded that the cultivated accession had changed dramatically (Poulsen et al., Reference Poulsen, Holten and von Bothmer2007).
A very large 19th-century seed collection has been stored at the Swedish Museum of Cultural History since it was donated by the Royal Swedish Academy of Agriculture and Forestry (KSLA) in 1963. The collection contains several sub-collections of seeds from exhibitions, collection expeditions, seed testing and breeding institutes, as well as from test cultivations. The collection is used mostly for display and has not been analysed for its research potential. An inventory of the KSLA seed collection was made and sub-samples representing different species and ages were analysed for viability and DNA quality. Here, we show the results of this screen of the seed collection. Furthermore, we demonstrate that molecular markers, such as microsatellites, can be used to determine genetic relationships between the accessions found in the seed collection.
Materials and methods
Seed material
Newly harvested seeds had been stored in glass containers sealed with either glass or wooden lids or, most commonly, cork plugs (Fig. 1). The seed containers were stored at room temperature, except for brief periods at lower temperatures. The approximate water content of seeds was 7–10%. Sub-samples from 100 accessions representing ten species differing in age and origin were taken from the containers (see Table 2). Fresh seeds of each species, provided by SW Seed, Svalöv, Sweden or Runåberg Seeds, Spekeröd, Sweden (Beta vulgaris cv. Boltardy) were used as positive controls.
Figure 1 Examples of seed containers in the KSLA seed collection.
Germination tests
Germination tests were performed using two different methods. In a standard germination test, 25 (Pisum sativum, Vica sativa and Beta vulgaris) or 50 seeds (all other species) were planted in or on top of sterile sand (ISTA, 1999). Seeds were also placed on agar using modifications to a method especially developed for aged seed (Poulsen et al., Reference Poulsen, Holten and von Bothmer2006). The agar method consisted of surface sterilizing 50 (Pisum sativum, Vica sativa and Beta vulgaris) or 100 seeds (all other species) in 70% ethanol for 30 s followed by 8 min in 1.5% sodium hypochlorite containing a few drops of Tween, then placing seeds on ¼ strength Murashige and Skoog (MS) medium (Duchefa, Haarlem, The Netherlands) solidified with 0.5% plant agar (Duchefa, Haarlem, The Netherlands). Germination assays were performed in a growth chamber at 20°C and 16 h light. Plates were inspected every second day for 50 d.
DNA extraction and analysis
DNA was isolated from 30–70 mg of dry seeds of each accession using the FastDNA® Spin Kit and the FastPrep® Instrument (Qbiogene Inc., California, USA). This procedure, where the seeds are ground in sealed individual tubes, allows rapid and efficient homogenization without sample-to-sample contamination risk. DNA was extracted from single seeds of cereals or one-quarter to one-half seeds for Pisum sativum and Vicia sativa. Unfortunately, insufficient DNA to be visible on gels was extracted from single individuals from the other species. Thus, one glomerule of Beta vulgaris, 10 seeds of Brassica napus, 20 seeds of Trifolium pratense and 50 seeds of Phleum pratense were used in each extraction. DNA was eluted in 100 μl of the supplied buffer. Negative controls were performed in parallel with each extraction series.
Fragment sizes of extracted DNA were estimated by banding patterns of 20 μl of extract run on 1% agarose 0.5 × TBE (Tris-borate-EDTA) gels. The size markers GeneRuler™ LowRange (Fermentas, Burlington, Canada) and Quick-Load™ 1 kb DNA ladder (New England Biolabs, Ipswich, Massachusetts, USA) were included as references. Gels were stained with ethidium bromide (EtBr) and examined with ultraviolet (UV) light. Gels were inspected by eye to identify the most abundant fragment sizes (with maximum fluorescence).
Polymerase chain reaction amplification
Amplification of fragments of different sizes was performed using different primer pairs: internal transcribed spacer (ITS) region of nuclear ribosomal DNA ITS1+ITS4 resulting in a ~700 bp fragment, ITS1+ITS2 resulting in a ~350 bp fragment (White et al., Reference White, Brims, Lee, Taylor, Innis, Gelfand, Sninsky and White1990), and nuclear ribosomal large subunit DNA LR6+LR17R resulting in a 109 bp fragment (http://www.biology.duke.edu/fungi/mycolab/primers.htm). Polymerase chain reaction (PCR) amplifications were performed in a 20 μl reaction volume comprised of 1 U Taq polymerase (Invitrogen, Carlsbad, California, USA), 1 × Invitrogen buffer, 1.5 mM MgCl2, 0.2 μM each primer, 0.2 mM each dNTP and 0.5 μl of DNA extract. Reaction runs consisted of 3 min at 94°C (initial denaturation): 35 cycles each of 94°C for 30 s, 48–60°C for 30 s and 72°C for 30 s; and a final elongation step of 72°C for 5 min. Annealing temperatures were 56, 60 and 48°C, for primers ITS1+ITS4, ITS1+ITS2 and LR6+LR17R, respectively. Amplification products were electrophoresed on 1.5–2.0% agarose 0.5 × TBE gels and visualized by EtBr staining and UV light. Two ~700 bp PCR products from aged samples of each species were sequenced (MWG, Germany). In each PCR run the extraction blank was used as a negative control.
Analysis with microsatellite markers
Microsatellite markers from six loci were amplified in Pisum sativum and Hordeum vulgare samples (see Table 3). The same DNA extracts as used previously (from one individual seed from each accession) were analysed. Procedures for Pisum sativum used semi-nested PCR (Loridon et al., Reference Loridon, McPhee, Morin, Dubreuil, Pilet-Nayel, Aubert, Rameau, Baranger, Coyne, Lejeune-Hènaut and Burstin2005). Each 20 μl reaction contained 1 U Taq polymerase (Invitrogen), 1 × Invitrogen buffer, 1.5 mM MgCl2, 0.07 μM forward primer with M13-tail at the 5′ end, 0.2 μM fluorescently labelled M13 primer, 0.27 μM reverse primer, 0.2 mM each dNTP and 0.5 μl of DNA extract. Either FAM or HEX were used as fluorescent dyes.
Microsatellite amplification of Hordeum vulgare DNA extracts followed Ramsay et al. (Reference Ramsay, Macaulay, Degli Ivanissevich, MacLean, Cardle, Fuller, Edwards, Tuvesson, Morgante, Massari, Maestri, Marmiroli, Sjakste, Ganal, Powell and Waugh2000). The forward primers were again modified with M13-tails. PCR amplification was run in two separate rounds, the first with modified forward primers and reverse primers, the second with a fluorescently labelled M13 forward primer (FAM or HEX) and the marker's reverse primer. PCRs were run in 20 μl reactions containing 0.5 U Taq polymerase (New England BioLabs), 1 × New England BioLabs ThermoPol buffer, 0.8 mM MgCl2, 0.1 μM each of forward and reverse primers and 0.25 μM of each dNTP. For the first round of PCR 1 μl of total DNA template was used, and 1 μl of the first PCR reaction was used as the template for the second round of PCR.
Amplification products were analysed by capillary gel electrophoresis and confocal laser scanning on a MegaBACE 500 DNA Analysis System (Amersham Biosciences, Uppsala, Sweden) using a 400 bp ROX-labelled internal size marker for Pisum sativum samples and on a MegaBACE 1000 DNA Analysis System using a 350 bp ROX-labelled internal size marker for Hordeum vulgare samples. Sizing of fragments was performed using the software MegaBACE Fragment Profiler 1.2 (Amersham Biosciences). Population subdivision between two-row and six-row Hordeum vulgare accessions (#192, #288, #717, #762, #787, #1818 and #659, #737, #1584, #1822 and cv. Rolfi, respectively) was estimated using Wright's FST (Wright, Reference Wright1951).
Results
The KSLA seed collection consists of 3393 accessions, mostly containing several hundred seeds each, representing 582 agronomically important species. About 96.5% of the accessions displayed excellent morphological preservation. The remaining samples, where the container or seal was broken, were completely or partly destroyed by mould or insects. The collection is comprised of several sub-collections (A–J): A, samples from test cultivations in the experimental field in Stockholm 1865–1894; B, a country-wide collection of samples from farmers' harvests in Sweden 1867; C, like B but from 1896; D, like B but from Finland 1882; E, seed from an exhibition in Italy 1873; F, seed from an exhibition in Philadelphia, USA 1876; G, seed collected by the first Swedish seed testing institute in Nydala, province of Halland and donated to the academy in 1877; H, seed for pharmaceutical use collected by the Karolinska Institutet in Africa and Asia from the 1850s and onwards, and donated to the academy in 1899; I, samples donated from the breeding company Vilmorin et Andrieux, France 1908; J, samples donated from the breeding company Svalöf, Sweden 1918. In addition to these major sub-collections several minor sub-collections and single samples are found. The distribution of the most frequently represented species by sub-collection is displayed in Table 1. The collection is comprised mostly of cereal grains (56%), but also has a smaller number of forage crops, vegetables, forest trees, other useful plants and important weeds. Most accessions are labelled with harvest year, name of cultivar (if applicable) and growth location.
Table 1 Distribution of accessions by species in the major sub-collection. Letters refer to the sub-collections (see Results). Only sub-collections with more than 40 accessions and species with more than ten accessions are individually presented. Other species and sub-collections are summarized
To analyse viability and DNA preservation, sub-samples were taken from 100 accessions representing ten different species and a range of ages (Table 2). No seeds from the collection germinated, whereas fresh seed samples from the same species all germinated to over 90%. Thus, seeds in the collection of the ten tested species are most likely non-viable. DNA yields were typically between 100 and 300 ng mg− 1 seed, with the highest yields from Pisum sativum and the lowest from Beta vulgaris. In fresh seed DNA yields were typically 3–5 times higher.
Table 2 Accessions from the seed collection and fresh reference samples tested for DNA fragment size and PCR amplification. Accession number (Acc. #) refers to the seed collection inventory number in the Swedish Museum of Cultural History. Nd=not detected; * indicates that the product was sequenced to confirm species identity; + and − indicate presence or not, respectively, of amplification product using ITS and LR primer sequences with the indicated fragment length and (+) or (−) indicates amplification results after DNA extracts were diluted
DNA isolated from fresh seed samples was primarily of high molecular weight, with the greatest intensity of fragments above 10 kb (Fig. 2a). In contrast, DNA isolated from the aged seed showed various degrees of degradation (Table 1, Fig. 2a). In the grass seeds (cereals and Phleum pratense) the majority of DNA fragments were very short (230 ± 70 bp). DNA isolated from the Fabaceae species was less degraded, although some exceptions of considerably more severely degraded samples were found. The best preserved DNA was found in Pisum sativum (2200 ± 1850 bp) followed by Vicia sativa (1720 ± 1050 bp) and Trifolium pratense (900 ± 830 bp). In Beta vulgaris DNA yields were small and degree of degradation variable between accessions (1070 ± 1060 bp). Finally, Brassica napus showed an intermediate degree of degradation (860 ± 460 bp). DNA was not detected in two Triticum aestivum samples (#132 and #134), one Secale cereale sample (#77) and one Brassica napus sample (#1697).
Figure 2 Examples of agarose gel electrophoresis of (a) extracted genomic DNA and (b–d) PCR amplification products from aged and fresh samples. Primer pairs are (b) ITS1+ITS4 resulting in ~700 bp products; (c) ITS1+ITS2 resulting in ~350 bp products; and (d) LR6+LR17R resulting in 109 bp products. Agarose gels were stained with ethidium bromide and visualized by UV light.
The ability to characterize degraded DNA reliably was tested by PCR amplification of fragments using universal primers and subsequent sequencing to verify species, as described by Walters et al. (Reference Walters, Reilley, Reeves, Baszczak and Richards2006). Three different sets of primers were used, resulting in amplification products of ~700 bp, ~350 bp and 109 bp in length. DNA isolated from fresh seed samples readily amplified (Fig. 2b–d, right half of gels), and PCR products were frequently obtained from aged seeds (Fig. 2b–d, left half of gels), with 76, 86 and 84 samples of the 100 tested providing amplification products for the ~700, 350 and 109 bp fragments, respectively (Table 2). PCR products were not obtained in the highly degraded grain samples for which DNA was not detected (Triticum aestivum #132 and #134, Secale cereale #77). PCR products were obtained in Trifolium pratense #1344 and Vicia sativa #1336, #1337, #1493 and 1521 after DNA extracts were diluted 1:10. Negative extraction controls did not result in any amplification product.
Amplified ~700 bp products were sequenced from two of the aged samples from each species (indicated by * in Table 2), and species identification was confirmed by a BLAST search of GenBank (May 2008) for 19 of the 20 samples. The obtained sequence from the amplification of the Phleum pratense #1354 extract had 97% identity with Aspergillus sequences. The amplification of fungal DNA instead of Phleum pratense DNA was also visible as a fragment length polymorphism (Fig. 2b) resulting in a shorter fragment.
To test whether DNA polymorphisms in single-copy nuclear genes could be detected we analysed the 11 Pisum sativum and Hordeum vulgare samples (Table 2) with microsatellite markers (Table 3). DNA extracted from Pisum sativum and Hordeum vulgare had the longest and shortest average fragment sizes, respectively. Size of amplified microsatellite products were visualized on high-density agarose gels (Fig. 3) and quantified using capillary gel electrophoresis and confocal laser scanning.
Table 3 Marker name and observed fragment sizes, chromosomal location, number of detected alleles and number of accessions with amplified product for the microsatellite markers tested in the Pisum sativum and Hordeum vulgare samples
a Markers from Loridon et al. (Reference Loridon, McPhee, Morin, Dubreuil, Pilet-Nayel, Aubert, Rameau, Baranger, Coyne, Lejeune-Hènaut and Burstin2005).
Figure 3 Examples of DNA polymorphisms detected by microsatellite markers. The figure shows markers AD83 and AA278 amplified from DNA extracted from ten Pisum sativum accessions and a modern pea cultivar. Amplification products were separated on a 3.5% MetaPhor agarose gel and stained with ethidium bromide for visualization. Fragment sizes (bp) determined by MegaBACE are indicated by the numerals below the gel.
For each species, six primers for fragments ranging in size from 156 to 300 bp gave polymorphic products for a total of 32 and 24 alleles in Pisum sativum and Hordeum vulgare, respectively (Table 3). All 11 samples within each species had some distinct alleles. In Pisum sativum all six markers could be amplified in all samples. Amplification was less successful in Hordeum vulgare samples. For example, only three of the six markers could be amplified in accession #737. For all other samples, all or nearly all markers amplified well and only the occasional marker (AF043094A) failed to amplify. The presence of non-amplifying null alleles at these loci cannot be ruled out. A high degree of homozygosity in both Pisum sativum and Hordeum vulgare is indicated by a single band for each microsatellite (Fig. 3 gives sample data for Pisum sativum) and is expected in inbreeding species. However, genotyping errors resulting from differential amplification of alleles at a locus (called allelic drop-out) are common in deteriorated DNA samples (Gagneux et al., Reference Gagneux, Boesch and Woodruff1997) and require additional experiments before complete homozygosity can be concluded.
To validate results by biological relevance we estimated FST values between the two-row and six-row Hordeum vulgare samples. The two types of Hordeum vulgare are known to be well genetically segregated (Kolodinska Brantestam et al., Reference Kolodinska Brantestam, von Bothmer, Dayteg, Rashal, Tuvesson and Weibull2007). Average FST across the markers was 0.27, suggesting moderate to high genetic differentiation. The degree of differentiation differed between markers, with one allele not present in two-row accessions being fixed in six-row accessions for one marker, and another marker being close to monomorphic in the two-row accessions. Other markers showed little differentiation.
Discussion
Aged and ancient plant material has lately gained attention for its utility in analyses of plant evolutionary and breeding history with the aid of DNA techniques. As recently reviewed by Jones et al. (Reference Jones, Lister, Bower, Leigh, Smith and Jones2008), in such studies herbaria and seed collections can serve as excellent complements to living and archaeological material. However, depending on the questions asked, the value of the materials will depend on several factors: (1) Can the origin and age of the material be correctly determined? (2) Is the quality of the seed DNA sufficient to allow for correct PCR amplification? (3) Is the amount of seed sufficient to be considered representative? This is especially important when the aim is to perform population genetic analyses. In this paper we have evaluated the hitherto unexplored KSLA seed collection for its research potential.
Most samples within the KSLA seed collection are dated with the year, in several cases the harvest day, and information on cultivar name and/or growing location is given on the majority of samples. Storage of seed in sealed glass containers protects them from decay and microbial infestation, as was evident in the KSLA collection where intact containers were free of infestation. The importance of container type for longevity of seeds in storage of viable seed has been stressed by Gómez-Campo (Reference Gómez-Campo2006) and, most likely, the same features apply for storing non-viable seeds.
Detectable amounts of DNA could be extracted from single seeds of the cereal species, Pisum sativum and Vicia sativa, although for the other species bulked samples had to be used. The utility of bulked samples for genetic analysis is limited as equal contribution of single samples to the DNA pool cannot be assumed. However, bulk genotyping may provide insight into the presence of certain alleles in a population. In contrast, the species where genotyping of single individuals is possible are more interesting for population genetic studies, both for comparisons within and between accessions. Many accessions in the KSLA collection consist of hundreds or thousands of seeds, meaning that, at least for cereals, each sample is collected from a fairly high number of individual plants. Although the exact sampling procedures are unknown, it is reasonable to assume that the seeds are representative of the original population.
Large differences were observed in the quality of the DNA extracted from aged seeds (Table 2). Stability of DNA seems to be correlated roughly with seed longevity (Walters et al., Reference Walters, Wheeler and Grotenhuis2005), with longer fragments found in Pisum sativum and Vicia sativa accessions, intermediate stability in Beta vulgaris, Brassica napus and Trifolium pratense accessions, and most deterioration found in accessions of the grasses. Seed morphology may play a role here, with large, hard-shelled seeds better protecting DNA from degradation.
PCR amplification was usually successful when the size of the amplified product was less than the average size of the extracted DNA fragments. An exception is Vicia sativa with proportionally long DNA fragments preserved but amplification by PCR being problematic. Inhibitory substances were most likely present in Vicia sativa and PCR amplification was more successful after a reduction of the amount of template.
In seeds, as with all biological materials, there is a relationship between age and DNA quality (e.g. Walters et al., Reference Walters, Reilley, Reeves, Baszczak and Richards2006). However, within the short time-span tested here (90–145 years) we could not clearly correlate DNA detoriation with age. Several of the oldest samples (from 1865) permitted PCR amplification of the ~700 bp product and the only samples yielding DNA neither detectable on gels nor possible to amplify by PCR was Triticum aestivum #132 and #134 and Secale cereale #77. These samples all come from the same sub-collection – a donation from the French breeding company Vilmorin et Andrieux 1908. Why these samples have more severely damaged DNA than the others is unclear and further analyses are required to conclude whether this specific sub-collection has DNA of inferior quality.
We have shown that single-copy DNA such as microsatellite markers can be amplified readily in DNA from single seeds. The use of such markers is desirable for several reasons. The use of species-specific primers is preferable to general primers to avoid amplification of, for example, fungal DNA. Furthermore, microsatellite markers are commonly used for analyses of population genetics, historical isolation and genetic differentiation within a species (Selkoe and Toonen, Reference Selkoe and Toonen2007). Here we show that microsatellite markers can be used to detect DNA polymorphisms in individual seeds of both the relatively preserved DNA of Pisum sativum and the more degraded DNA of Hordeum vulgare from the 19th century.
Due to great availability in crop species, species specificity and low requirement for long DNA fragments, microsatellites have also been used in studies of ancient and aged plant material. For example, in 2600- to 1700-year-old Vitis vinifera seeds, Manen et al. (Reference Manen, Bouby, Dalnoki, Marinval, Turgay and Schlumbaum2003) could detect polymorphism with three primer pairs. From Zea maize seeds and cobs, 400–1300 years old, Lia et al. (Reference Lia, Confalonieri, Ratto, Hernández, Alzogaray, Poggio and Brown2007) could amplify products with three different primer pairs, although no polymorphism was detected. Kobayashi et al. (Reference Kobayashi, Ebana, Fukuoka and Nagamine2006) amplified 19 different microsatellite markers in Oryza sativa seeds that were almost 100 years old. However, such results have to be evaluated carefully, as microsatellites are prone to slippage during amplification, not the least when the template DNA is deteriorated (Gugerli et al., Reference Gugerli, Parducci and Petit2005). One way to validate results is to evaluate them in the context of breeding history. In this study, we could distinguish clearly between the six-row and two-row barley accessions. While both the number of markers and individuals are too small to draw any major conclusions regarding population structure, it is nevertheless clear that the KSLA material will be useful for future population genetic studies.
In conclusion, we have shown that material from an extensive 19th-century seed collection permits DNA extraction and analysis. The well-documented material with high biological integrity opens up possibilities for exploring several issues concerning crop evolution, agricultural history in northern Europe and preservation of genetic resources. Genetic variation among landraces and obsolete cultivars can be studied, as well as genetic shifts through selection during the early decades of modern plant breeding. Comparisons between original populations of collected seed and those regenerated to maintain supply or viability may show the potential effects of genetic drift and contamination during gene-bank maintenance. Thus, DNA technology has increased the information value of unviable seeds stored in historical collections to levels unimagined by those who originally deposited the seeds in the museum.
Acknowledgements
We thank Annette Molbaek and Åsa Schippert at IBK, Linköping University for skilful assistance with MegaBACE. Curator Johan Åkerlund at the Swedish Museum of Cultural History is acknowledged for providing the historical documentation of the seed collection. This work was funded by the Lagersberg Foundation, the Swedish Board of Agriculture, the Royal Swedish Academy of Agriculture and Forestry, the Nilsson-Ehle foundation and Carl XVI Gustafs 50-year anniversary fund for science, technology and environment.
