Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Genetic Diversity and Population Structure in a Legacy Collection of Spring Barley Landraces Adapted to a Wide Range of Climates

Genetic Diversity and Population Structure in a Legacy Collection of Spring Barley Landraces Adapted to a Wide Range of Climates

  • Raj K. Pasam, 
  • Rajiv Sharma, 
  • Alexander Walther, 
  • Hakan Özkan, 
  • Andreas Graner, 
  • Benjamin Kilian
PLOS
x

Abstract

Global environmental change and increasing human population emphasize the urgent need for higher yielding and better adapted crop plants. One strategy to achieve this aim is to exploit the wealth of so called landraces of crop species, representing diverse traditional domesticated populations of locally adapted genotypes. In this study, we investigated a comprehensive set of 1485 spring barley landraces (Lrc1485) adapted to a wide range of climates, which were selected from one of the largest genebanks worldwide. The landraces originated from 5° to 62.5° N and 16° to 71° E. The whole collection was genotyped using 42 SSR markers to assess the genetic diversity and population structure. With an average allelic richness of 5.74 and 372 alleles, Lrc1485 harbours considerably more genetic diversity than the most polymorphic current GWAS panel for barley. Ten major clusters defined most of the population structure based on geographical origin, row type of the ear and caryopsis type – and were assigned to specific climate zones. The legacy core reference set Lrc648 established in this study will provide a long-lasting resource and a very valuable tool for the scientific community. Lrc648 is best suited for multi-environmental field testing to identify candidate genes underlying quantitative traits but also for allele mining approaches.

Introduction

Cultivated barley, Hordeum vulgare L. ssp. vulgare, the domesticated form of wild barley H. vulgare L. ssp. spontaneum (C. Koch) Thell., is one of the oldest cereal crops [1]. Barley withstands hot and dry climates, marginal soils, to some extent salinity and a broad range of soil pH conditions [2], [3]. Today, barley is grown from 70° N in Norway to 46° S in Chile. The morphological, physiological and functional variation in barley reflects the underlying genetic diversity, which eases the environmental adaptation of this species [4], [5].

In barley, as in most crops, genetic bottlenecks occurred during domestication and crop improvement. For most loci current elite varieties harbour less genetic diversity than their wild relatives or early domesticates [6][9]. Landraces are traditional domesticated populations of locally adapted genotypes maintained by local farmers over generations. Early barley cultivars were direct selections among landraces or descended from genetic recombination of different landraces. Since then, new barley varieties are mainly developed through the reshuffling of alleles resulting in a more or less constant repertoire of alleles within the elite gene pool. Overall, the genetic basis in present elite barley breeding materials is rather limited.

Since the beginning of the 20th-century landraces were largely replaced by modern cultivars [10], [11], which are higher yielding under optimal conditions but which can completely fail under harsh environments [12]. Today, most barley landraces have disappeared from practical farming. Many of them are still maintained in ex situ repositories. Globally, landraces represent the largest part of barley germplasm conserved in genebanks (44%, 128.870 accessions) [13], [14].

Examples demonstrating the utilization of landrace genetic diversity include the introgression of i) plant height dwarfing alleles (Rht1 and Rht2) derived from the Japanese wheat landrace “Shiro Daruma” [15], ii) several disease and insect resistance genes in wheat [16], [17], iii) submergence tolerance (Sub1) in rice [18], iv) broad spectrum powdery mildew resistance allele mlo11 (from an Ethiopian barley landrace) [19], the rym4 virus resistance [20] or v) the boron-toxicity tolerance in barley, which was obtained from the Algerian landrace “Sahara” [21].

Genome-wide association analysis (GWAS), a population based method to identify marker-trait associations based on linkage disequilibrium (LD) is being recently used extensively in crop plants [22]. Genetic diversity, population size and stratification, extent of genome-wide LD, allele frequency distribution, marker type and coverage as well as other parameters determine the accuracy, resolution and power of GWAS [22][24]. As a consequence of genetic bottlenecks during domestication and crop improvement, allele frequency changes resulted in different levels of LD and genetic diversity. Thus, the extent of LD increases from wild barley to landrace and to elite cultivars, whereas the reverse trend was observed for genetic diversity [25]. The trade-off of higher LD in current cultivars is lower resolution in GWAS [26]. To fine-map QTL, the varying extent of LD observed in different genepools of barley (e.g. wild, landraces, and cultivars) could be exploited and provides a great opportunity for high-resolution association mapping [27].

Several studies were performed in barley to investigate genetic diversity in different germplasm collections using molecular markers, and few of these collections (panels) were established for GWAS. However, most studies were based on either cultivar collections [28][30], or mixtures of cultivars and landraces [31][33] or landrace panels from distinct geographical regions only [34][37]. Very few collections of barley landraces collected from a wider geographic range were established but none of them was designed for higher resolution GWAS [11], [38][39].Here we describe the establishment of a comprehensive and diverse collection of spring barley landraces adapted to a wide range of climates (Landrace Collection 1485, “Lrc1485”). We studied genetic diversity and population structure using microsatellite (Simple Sequence Repeat, SSR) markers, which provided the basis for targeted research activities. The legacy core reference set (CRS) Lrc648 established here is available for the scientific community to integrate data and to improve our elite barley varieties under changing environmental conditions.

Material and Methods

Plant material

In total, 1491 spring barley landrace accessions adapted to diverse climate conditions were originally considered for this study. This representative collection was carefully selected among 22,093 Hordeum accessions available in 2008 at the “Federal ex-situ Genebank for Agricultural and Horticultural Crop Plants” maintained at the Leibniz Institute of Plant Genetics and Crop Plant Research (IPK) in Gatersleben, Germany. The landraces originated from 41 countries in Europe, West and Central Asia, North and East Africa and covering 5.63°–62.47° N and 16.62°–71.5° E. The selection was based on the following criteria: i) Mansfeld's taxonomical classification system considering growth habit, row type of the ear, kernel coverage, spike density, and seed colour [40], ii) the collection site had to be well documented (S1 Table), and iii) passport data (Characterization and Evaluation data since 1946) - obtained from IPK's Genebank Information System, GBIS, http://gbis.ipk-gatersleben.de). Barley landraces collected during targeted expeditions before 1992 were considered whenever possible to work with trustable materials. From Syria and Jordan only a few landraces (six accessions) were included - as barley landraces from this region were comprehensively investigated by [34], [41]. Apart from other considerations, the proportion of landraces selected from each country should represent the overall composition of IPK's spring barley landrace collection containing 6,800 accessions meaning the number of accessions per country differed (Fig. 1, Table 1, S1 Table).

thumbnail
Figure 1. Geographic information system (GIS)-based topographic maps indicate the natural distribution of 1485 barley landrace accessions.

(a) Structure (K = 4); (b) Structure (K = 10) - inferred clusters. Every individual accession is represented by a coloured circle indicating the membership to a cluster. Admixed accessions are indicated by black stars. See Fig. 2, Fig. 3, S2 Fig., S3 Fig., S4 Fig., Table 1, Table 3, Table 4, S3 Table and S1 Table for more information.

https://doi.org/10.1371/journal.pone.0116164.g001

thumbnail
Table 1. Distribution of landraces considered in this study according to their countries of origin, caryopsis type and row type.

https://doi.org/10.1371/journal.pone.0116164.t001

The IPK genebank practices the splitting of original landrace accessions and maintains them as morphologically distinct accessions to counteract the possible loss of rare alleles in the original population of genotypes [42], [43]. For this reason, a current landrace accession might correspond to a single representative genotype of the original landrace population collected. Whenever possible, latitude and longitude coordinates were inferred using the original collection site descriptions. A broader source location (nearby city or province or country capital) was considered to infer the geographic coordinates, when the exact collection location was not documented. Searches were performed using Google maps (http://maps.google.com/maps) and global Gazetteer version 2.2 (http://www.fallingrain.com/world/index.html).

Genotyping using SSR markers

Four seeds per landrace accession were randomly selected and sown under greenhouse conditions at IPK in 2008. Leaf material from one representative plant per accession was harvested three weeks after sowing. DNA was isolated from freeze-dried leaves using a BioRobot 9600 Work Station and the MagAttract 96 DNA plant core kit (Qiagen, Germany). Forty-five fluorescence-labelled SSR markers were selected based on their mapping position in the barley genome, covering all seven chromosomes [44], [45] (Table 2, S1 Fig.). Primers were labelled with HEX, FAM and TAMRA dyes allowing multiplexing of primers pairs into 15 multiplexes (M1 to M15) with three primer pairs per amplification. PCR reactions were performed following the protocols described by [46]. Amplification products were separated on a MegaBACE 1000 capillary sequencer (Amersham Biosciences). Fragment sizes were recorded and analyzed using MegaBACE fragment profiler software version 1.2 and inspected manually. Allele sizes and peak intensities were recorded. Low intensity bands were assigned missing values. Two markers that were monomorphic (GBM1043 & GBM1036) and one marker that amplified multiple fragments (GBM1326) were excluded from further analysis (Table 2, S1 Fig.). Six accessions were excluded from diversity and population differentiation analysis due to pure DNA quality leaving 1485 accessions for analysis (S1 Table).

Inferring the population structure

The population structure of the 1485 landraces (Lrc1485) considered was inferred using Structure 2.2 [47], [48] based on 42 SSR markers. This approach uses a Bayesian clustering analysis to assign individuals to clusters (K) without prior knowledge of their population affinities. Structure simulations were performed with the number of presumed clusters from K = 1 to K = 20 and five runs per K value. For each run, the initial burn-in period was set to 50,000 followed by 100,000 Markov Chain Monte Carlo (MCMC) iterations. The most probable number of clusters was determined by plotting the estimated likelihood values [LnP(D)] as a function of K [48]. Furthermore, delta (K) values were also calculated as proposed by [49]. A cut-off limit of 60% membership coefficient (Q-matrix) was considered to assign the individuals to a particular group as suggested by [50]. Accessions that did not meet this criterion were considered as admixed.

Principal Component Analyses (PCA) were performed using Past 2.12 [51]. Neighbor-Joining tree and Neighbor-Net planar graph based on Hamming distances (uncorrected p-distance) between 1485 landraces were constructed using SplitsTree 4.13.1 [52].

Genetic diversity and population differentiation

The number and frequency of alleles, gene diversity and heterozygosity (He), were determined for all loci across the total population using Powermarker 3.25 [53]. Polymorphism Information Content (PIC) values were determined according to [54]. Genetic variation within and among populations was estimated by Analysis of Molecular Variance (AMOVA) using Arlequin 3.1 [55]. AMOVA was conducted between morphological, geographical and Structure inferred groups. Genetic differentiation among groups was calculated based on unbiased Fst estimators [56]. Pair wise population comparisons using Fixation statistics (Fst) were determined among all groups as well as allelic richness, gene diversity (GD) and the number of alleles for each group were calculated using Fstat 2.9.3 [57]. Allelic richness values for subpopulations were calculated based on rarefaction to account for varying sample size [58]. In all analyses, statistical significance was determined by performing 1000 permutations.

Spatial distribution of groups in relation to geography and climate

Geographic ground distances in kilometres between accessions were calculated based on latitude and longitude coordinates. The genetic diversity index [59] was calculated, and the genetic distance matrix was calculated using the shared allele distance approach of [60] based on allele frequencies at 42 loci. Based on this, Mantel tests were conducted between genetic distance and all other distance matrices. Mantel correlograms were generated analogous to the autocorrelation function [61] - and allowed to assess the overall relationship between matrices and to determine the significance level of correlation for each distance class. Mantel correlograms were constructed using Passage 2.1 [62].

The spatial distribution of accessions, categorized into population-based groups, was visualized in relation to climate zones based on the Köppen-Geiger classification [63]. These climate zones are obtained by classifying the mean climate conditions on land areas around the globe using climate variables such as annual and seasonal mean temperatures and precipitation. The five main zones range from tropical climate through arid and temperate climates to polar climates. Subgroups can be determined by, for example, hot summers, dry and cold winters, year-round precipitation or monsoonal conditions.

Establishing a Core Reference Set (CRS) for high resolution LD-mapping

To build a legacy CRS [64] comprising c. 600–700 genotypes, we first genetically purified the entire collection using two rounds of single seed descent (SSD) [65] under greenhouse conditions in 2008–2009 (started with the same plants from which we isolated DNA, see above). Afterwards, the materials were multiplied and survey-phenotyped under field conditions at IPK using local standard agricultural practices in two subsequent years: i) 2010 (single row per landrace genotype, isolated by one row of a spring wheat genotype, 8 plants per row, 20 cm distance between the plants), and ii) in 2011 (micro plots, 1.2×1 m; all available seeds from the 2010 harvest were sown and equally distributed over six rows). Plants were harvested and threshed manually to avoid mixtures of seeds. In total, 1014 genotypes produced sufficient seeds after two rounds SSD and the two subsequent multiplication cycles (S1 Table). Subsequently, the M strategy as implemented in Mstrat [66] was used to establish the legacy CRS. All data sets available were considered for this: i) geographical origin; ii) 42 SSR markers, and iii) morphological traits as quantitative parameters, which were obtained during seed multiplication in 2011, such as row type, caryopsis type, heading date, plant height and spike length (S1 Table). The best CRS that maximizes the number of observed alleles was then established based on five replications using Mstrat. Diversity scores calculated based on allelic richness [67] were compared and validated with a random sampling approach as well as by comparing phenotypic diversity.

Results

We considered a collection of 1485 landraces originating from 41 countries (Fig. 1). This collection comprised of 708 two-rowed (47.7%) and 777 six-rowed (52.3%) barley genotypes. In total, 284 (19.1%) naked genotypes were considered (Table 1).

High levels of polymorphism and many rare alleles in the collection

Data quality was high, and the level of missing information across SSR markers was very low (1.528%). Based on the 42 SSR markers considered, 372 alleles were obtained with fragment sizes ranging from 90 to 360 bp (Table 2, S2 Table). The number of alleles per locus ranged from 3 (GBM1363) to 22 (GBM1007, GBM1015) with an average of 8.86 alleles per locus. Major allele frequencies per locus ranged between 0.213 (GBM1256) and 0.974 (GBM1404) (mean value of 0.512) and PIC values ranged from 0.050 (GBM1404) to 0.839 (GBM1256) (average of 0.548). The majority of markers (78%) showed PIC values between 0.4 and 0.8. The average allelic richness was 5.7412, ranging from 2.1950 (GBM1363) to 15.1820 (GBM1015) (Table 2). An average gene diversity (GD) value of 0.6036 was obtained, indicating a high level of genetic variation among the accessions. Heterozygosity levels (He) were very low ranging 0–0.05, with an average value of 0.0119 per locus (Table 2). A total of 157 rare alleles (allelic frequency <1% in the total collection) were identified amounting to 42% of the total number of alleles discovered (Table 2, S2 Table).

Population structure within the panel

Structure runs were performed for K = 1 to K = 20 based on the distribution of 372 alleles at 42 SSR loci among 1485 accessions. LnP(D) values increased slowly starting from K = 10, thus probably representing the most appropriate number of major clusters in this collection. However, the maximum delta (K) value was reached at K = 4 (Fig. 2).

thumbnail
Figure 2. Structure analysis for 1485 barley landraces.

a) mean Log probability values (LnP(D)) plotted as function of K (number of clusters); b) Delta K vs. K plotted as proposed by [44]. The graph indicates the maximum change at K = 4.

https://doi.org/10.1371/journal.pone.0116164.g002

The primary division at K = 2 was observed mainly between Ethiopian (Group 1) and non-Ethiopian landraces (Group 2) (S1 Table). Here 92.59% of the landraces were assigned to either of two groups (G) considering the 60% membership coefficient (Q-matrix). The least number of admixtures (110 accessions; 7.4%) was observed at this K value.

At K = 4, 91.1% of the landraces were assigned to one of the following groups – G1 (314): two-rowed and six-rowed barleys mainly from Southwest Asia; G2 (277): two-rowed and six-rowed landraces mostly from Ethiopia; G3 (439): mainly six-rowed hulled barleys from a wide geographical range including northern Africa; and G4 (323): mainly two-rowed hulled forms from Europe and Southwest Asia. The population structure inferred using PCA provided congruent results (Fig. 1, Fig. 3, S2 Fig., S3 Fig., Table 3, Table 4, S1 Table).

thumbnail
Figure 3. Population structure of 1485 landraces using 42 SSR inferred by Structure.

a) for K = 4; and b) for K = 10. Genotypes were ordered according to their membership coefficient (Q) values to one group.

https://doi.org/10.1371/journal.pone.0116164.g003

thumbnail
Table 4. Diversity and summary statistics for Structure inferred groups a) K = 4; and b) K = 10.

https://doi.org/10.1371/journal.pone.0116164.t004

Known key determinants of the population structure for barley like i) row type of the ear, ii) kernel coverage, and iii) geographical origin divided the collection at K = 10 (Fig. 1, Fig. 3, S2 Fig., Table 3, Table 4, S1 Table). The clusters were also relatively well associated to distinct climate zones (S4 Fig.). G1 (194) consisted of naked barleys mostly from Ethiopia. G2 comprised of 66, mostly six-rowed landraces from a wide geographical range were hot desert climate is prevailing. G3 was the second largest group (226; 15.2%) and harboured the highest average number of alleles (252) and 78 group specific rare alleles. Allelic richness based on rarefaction provided the highest value for this group (Table 4b). Mainly six-rowed hulled barleys from the Mediterranean were included here: Libya (61), Morocco (48), Spain (30), Italy (21), Greece (28) and Turkey (10). All 56 landraces assigned to G4 originated from or close to Georgia. Interestingly, both two-rowed and six-rowed hulled types were present in this group. All but one of the 83 accessions assigned to G5 were hulled barleys from Ethiopia but with different row types: including all assigned deficiens (9); intermedium (4); and labile (7) genotypes. Gene diversity was lowest for G5 (0.258) (Table 4). Altogether, 80 landraces clustered in G6, comprising mainly hulled barley from Afghanistan (61) and Iran (7). Among nine naked barleys considered here, five were collected in Afghanistan. Ninety-seven, mainly two-rowed landraces from Minor Asia and the eastern part of the Fertile Crescent were assigned to G7 including lines from Turkey (29), Iraq (11), Iran (32), Afghanistan (5) and Georgia (7). Highest gene diversity value was found for this group (0.492) (Table 4). G8, the largest group at K = 10, consisted of 295 accessions of mainly two-rowed hulled types. Among them, 276 accessions originated from Europe. G9 was the smallest group harbouring 54 naked landraces, mainly of non-Ethiopian origin (52), and only 129 alleles (Table 4). Finally, G10 consisted of 138 landraces, majorly six-rowed hulled barleys from a broad geographical range north of the Mediterranean Sea (Fig. 1, S4 Fig.). At K = 10, principal components (PC) 1 and 2 explained 9.90% and 8.58% of the variation, respectively, and separated most Structure inferred groups (S2 Fig.). Relationships among accessions within each Structure inferred group are shown in S5 Fig. Variation explained by PC1 and PC2 together, per group, ranged from 11% to 51% (G1, 23.95%; G2, 39.67%; G3, 11.17%; G4, 25.32%; G5, 21.15%; G6, 18.036%; G7, 26.04%; G8, 19.99%; G9, 51.05%; G10, 16.84%) indicating especially within G1, G2, G4, G5; G7 and G9 further substructure based on e.g. row type or geographical origin.

Pairwise Fst values among Structure inferred groups at K = 10 ranged from 0.196 (between G4 - from Georgia and G7) to 0.556 (between G1 – naked Ethiopian, and G4). Similarly pairwise Fst comparisons for inferred groups at K = 4 ranged from 0.187 to 0.388 (S3 Table). Highly significant levels of genetic differentiation between and within populations were found using AMOVA. The percentage of genetic variation among populations ranged from 6.38% (all two-rowed vs. all six-rowed Ethiopian) to 37.6% (Structure inferred groups(K = 10)) and the percentage within the populations ranged from 62.40% (Structure inferred groups(K = 10)) to 93.62% (all two-rowed vs. all six-rowed Ethiopian). The values for the comparison of all Ethiopian naked vs. all non-Ethiopian naked barleys are interesting and possibly indicating two evolutionary lineages (S4 Table).

Association between eco-geographical factors and genetic diversity

Significant relationships were found between genetic distances of accessions and eco-geographical parameters at the site of origin (S5 Table). Most significantly correlated with genetic distances was geographical distance, followed by latitudinal distance (S5 Table). The Mantel test between the genetic distance matrix (shared allele distance matrix) and the geographical distance matrix revealed a significant Mantel correlation of 0.357 (P <0.0001). The correlation between genetic distance and longitudinal distance (r = 0.305, P <0.0001) was high. The correlation between genetic distance and latitudinal differences (r = 0.193, P <0.0001) was 2-fold lower than the correlation between genetic and geographic distances. In the Mantel correlogram for genetic distance vs. geographic distance, the matrix was subdivided into 20 discrete distance classes. Within the distance class of 0-300 km between the collection sites of accessions, the correlation was highest (r = 0.523). The r - values (Mantel correlation) and their significances declined with increasing distances. Spatial Mantel correlograms provided similar trends (S6 Fig., S5 Table).

Structure inferred clusters at K = 10 were assigned to distinct Köppen climates, although value should not be given to the few suspect lines (see below) (Fig. 1, S3 Fig., S4 Fig.). S6 Table shows the major Köppen climates for each of the Structure inferred group. Climate zones range from winter dry tropical climates (Aw) to fully humid boreal climates (Dfb). Most samples are located in warm temperate climates, either summer dry (Csa) or fully humid (Cfb). In more detail one can observe, that all accessions from G1 and G5 are located in winter dry climates in Ethiopia; G4 lines were only sampled from Georgia, where fully humid boreal climates (Dfb) prevail. G2 is mainly located in hot and dry desert climates (BWh) and G6 (six-rowed) prevails in wintercold steppe climates (BSk). G7 occurs along a narrow latitudinal range (30–43° N) in Minor Asia and the Fertile Crescent in warm temperate summer dry climates (Csa). G8 (two-rowed) and G10 (six-rowed) were sampled further north, majorly between 40–50°N (Cfb, temperate fully humid) and 36–47°N (Csa, temperate summer dry), respectively.

Establishment of a core reference set

To determine the theoretical optimum size of a core group representing most of the genetic diversity within the whole Lrc1485 collection, different sizes of core groups from n = 1 to n = 1485, with a step size increase of 50 accessions were computed using Mstrat and plotted against their diversity score calculated based on allelic richness. The M strategy performed better in efficiently capturing maximum allelic diversity than the random sampling approach (S7 Fig., Table 5). The theoretical optimum size of the core group was determined as in the range between c. 600 to 750 samples. More specifically, 745 individuals must be selected to harbour the maximum allelic diversity (S7 Fig.). Subsequently, based on the 1014 genotypes, which were available after two rounds of SSD and multiplication, the core reference set Lrc648 was established using Mstrat consisting of 308 two-rowed (242 hulled and 66 naked types) and 340 six-rowed (285 hulled and 55 naked types) genotypes (S8 Fig., S1 Table, S7 Table). We validated the superiority of the M strategy by comparing phenotypic diversity between the core set Lrc648 and a random set of landraces comprising the same number of accessions (Lrc648r) (S8 Table).

Discussion

The demand for higher yielding and better-adapted crop varieties has raised the need to exploit the large ex situ genebank collections [68]. So far, in case of barley, only very few collections of landraces were investigated, mostly sampled from particular geographical regions [34], [35], [69], [70]. The primary aim of this study was to establish and to characterize a diverse legacy collection of barley landraces adapted to a wide range of climates.

Lessons learned from genebank materials

Most probably due to careful selection and accurate description of the material (conservation management of IPK genebank certified according to - DIN EN ISO 9001:2008; line splitting practice for heterogeneous accessions), very few geographical outliers were observed. Another reason for the good fit of accessions into geographical clusters could be that 1035 (69.7%) accessions were sampled during IPK collection missions (or from collections which are hosted at IPK e.g. AUTMAYR-22-32, HIND-35/36, IRNFAOKU-52-54, S1 Table) and then maintained at IPK – thus significantly reducing confusions arising from seed exchange with other genebanks. For most accessions considered here, the collection site information was available (which is usually rarely the case for landraces from other genebanks), thus significantly improving spatial-evolutionary studies, too.

Few accessions appeared to be suspect and do not seem to represent their original collection site (e.g. at K = 10 for G1: the seven lines sampled in Europe and Morocco). These accessions were probably not collected originally in these areas. We assume they were sampled elsewhere and then grown ex situ in these countries, or they may have been incorrectly recorded or mixed up during seed exchange and propagation. As recently shown by [50] consequent elimination of any doubtful line identified would provide the best resolution.

Higher values of genetic diversity within the Lrc1485 and the legacy collection Lrc648 compared to the world-wide Genobar GWAS panel

In the collection of 1485 barley landraces, 372 alleles were detected using 42 SSR markers, with an observed average allele number (AN) of 8.95 alleles per locus, which is larger than in most other studies, including [71] using 45 SSRs in 223 cultivars of worldwide origin but also including some wild barleys (AN = 7.7), [69] using 39 SSRs for Eritrean landraces (AN = 7.6), [35] with 44 SSRs for Himalayan landraces (AN = 5.54) or [72] with 12 SSRs for Sardinian landrace populations (AN = 5.6). On the other hand, [73] reported AN = 16.7 for a worldwide collection of 953 barley accessions including wild barleys based on 48 genomic SSRs – which are much more polymorphic than the cDNA derived SSRs in our study. Certainly, the number of alleles per locus depends on the genotypes considered, the loci investigated and the marker type.

A total of 157 rare alleles were detected in the whole collection. Rare alleles were mostly detected at SSR loci, which displayed a high number of polymorphic alleles. The presence of 42% rare alleles highlights the potential of this collection for subsequent allele mining studies but could potentially limit GWAS.

To evaluate the potential of Lrc1485 for GWAS in more detail, genetic diversity values were directly compared to the potentially most diverse GWAS panel for barley comprising of 224 spring barley cultivars and landraces of worldwide origin (“Genobar” panel) [32], [46], [74], [75] - using the same set of 42 SSR markers (Table 5). Overall, the Lrc1485, the CRS Lrc648 but also the subpanels of Lrc648 based on the row type of the ear - are all more diverse than the Genobar panel as indicated by e.g. the total number of alleles, number of unique alleles or average gene diversity. Unique alleles observed in the Genobar panel came from genotypes collected from East Asia as well as from the Americas. Such regions were not considered in Lrc1485 (Table 5).

LD estimates between SSR markers were computed (S9 Fig.). All pair-wise comparisons showed very low LD (r2 <30), which is not surprising due to the relatively low marker coverage (S1 Fig.) and the large population size. Inferring genome wide LD dynamics in this population from few SSR markers would be a vague interpretation. Therefore, high-density marker coverage across the genome is required (a least a few hundred of mapped and equally spaced SNPs) to estimate more precisely the LD extent and the pattern at the population level and across the genome.

Ten major clusters define most of the population structure within the spring Lrc1485 panel

We assessed population structure by different approaches. Based on Bayesian clustering and PCA analyses we considered K = 10 as the most appropriate number of major clusters in the Lrc1485 collection. Although the maximum delta (K) value was reached at K = 4, clusters at K = 10 were better defined based on geographical origin, row type of the ear and caryopsis type – and also associated to distinct Köppen climate zones.

Clusters G1, G3, G4, G5 G8 and G10 were distinct and well supported. An intrinsic genetic substructure was visible for groups, which included accessions from a larger geographical range (G2, G6, G7, G9). PCA provided mostly congruent results for these clusters. However, PCA does not classify accessions into discrete clusters in all cases, especially not when admixed accessions and accessions of various geographical origins with a constant gene flow are included [76]. Neighbor-Joining and Neighbor-Net analysis supported these findings. All clusters obtained by Structure analysis were distinguishable, although a high amount of reticulation was evident, owing to the fact that common alleles per locus were shared among geographic regions (S10 Fig., S11 Fig.).

To explore the genetic diversity and relationships among and within Structure inferred groups, various diversity statistics were assessed (Table 4). Gene diversity (GD) over 42 loci was highest for G7, G6, G2 and G3, which is comparable with other local collections [69], [77]. Highest values for allelic richness and group specific rare alleles were found in G3, which might be due to i) materials sampled from a wide range of eco-geographical conditions around the Mediterranean Basin (local climates and niches at similar latitudes) (Fig. 1, S4 Fig.); or ii) gene flow among barley genepools [31].

As expected for barley, the row type of the ear was an important determinant of the population structure [28], [78] and six groups comprised majorly either two-rowed (G7, G8) or six-rowed (G2, G3, G6, G10) types, respectively (Table 3). Karyopsis type defined G1 and G9 as they included only naked types. Adaptation to local climates is defined e.g. through G4 (from Georgia) and G5 (from Ethiopia) containing both row types. However, within groups G1, G4, G5 and G9 subclusters can largely be explained by the row type of the ear.

Eco-geographical factors and spatial genetics

Detailed knowledge of environmental parameters at a certain collection site can help determining the role of relevant climatic factors influencing the genetic differentiation and adaptation of genotypes to their environments. Furthermore, this knowledge might help selecting most suitable parental lines for population development and breeding programs.

The distribution of accessions was found rather along the latitudes, meaning a wider W-E than N-S window. Just Ethiopia is the southern-most sampling area, i.e. stretching the latitudinal direction the most. In general, climatologically, climate zones are determined by i) incoming solar radiation (more at equator, less at poles); ii) b) altitude (higher elevation will yield similar climate zones as usually on higher latitudes); and iii) proximity to the sea yielding maritime vs. continental climates. Thus the general pattern of climate zones (as the word “zonal” means in this context) is band along the latitudes.

The distribution pattern of Structure inferred groups probably indicate a preferred distribution path rather zonal (W-E) than meridional (N-S), i.e. higher correlation with longitude and indeed this was observed for Lrc1485 groups. As domesticated barley spread meridionally from the Fertile Crescent to north-western Europe, the crop encountered considerable ecological and environmental change. Natural mutation, selection and enrichment of favorable alleles at key loci such as Ppd-H1 [79], HvCEN [80] or vernalization-related genes VRN-H1 [81], [82], VRN-H2 [83] and VRN-H3 (also known as HvFT1, [84]) contributed to successful environmental adaptation and range extension in barley.

However, based on the IPK genebank information system, we selected only spring barleys, which flowered at IPK without the need for any cold period to promote flowering. In Central Europe (such as at IPK), spring types are sown between early March and end of April depending on the weather conditions every year [68]. Thus, vernalization-or frost tolerance-related loci should be less relevant for the adaptation of spring barleys to their environments. However, as shown by [85], the Genobar spring barley collection, harbored 8 haplotypes at VRN-H3. Genotyping the same set of 224 accessions at VRN-H2 suggested the presence of recessive alleles at VRN-H2 (Kilian et al. unpublished), observed as deletions of a cluster of up to three ZCCT family genes, which contribute to the spring growth habit [86]. Thus, ideally, genotypes should be characterized at molecular and phenotypic levels, before assigning them to a specific growth habit (i.e. winter, facultative, spring). However, in the genebank context this has not been achieved for most of the collections yet. Therefore, we expect some facultative types within Lrc1485.

As already shown partly for the Genobar collection, which harbors six haplotypes at PpdH1 [85] and six out of seven haplotypes detected in the domesticated genepool at HvCEN studied by [80], we also expect a remarkable number of haplotypes at key genes responsible for environmental adaptation within Lrc648. Thus, we suggest using the legacy CRS Lrc648 or their two subpanels in particular for allele mining and gene discovery studies.

Although altitude must be considered as an important factor influencing the genetic diversity distribution [87], this factor was not considered in our studies because precise geographic coordinates were not available for all landraces.

The initial Köppen analysis presented here does not give a very clear picture due to outliers and thus a rather high within-group-variability (S6 Table). However for some groups a particular climate applies. Regarding climate change one could investigate Köppen maps based on future climate projections and check where climate zones suitable for barley cultivation will be located in the future. Modelling crop performance under changing climates will help guiding the breeding programs to the expected future needs [88][90].

New insights into barley evolution: two examples

With our genetic analysis in 1485 barley landraces, new insights into barley domestication history can be obtained. In total, 299 accessions from Ethiopia were genotyped (Table 1), thus providing probably the largest SSR data set generated for Ethiopian landraces so far (S1 Table, S2 Table). Overall, Ethiopian barleys were apparently found to be distinct from all other groups (S3 Table, S4 Table), which is in line with previous studies [87], [91][93]. Different evolutionary forces (environment) and domestication histories (e.g. agricultural practices, cultural preferences of human tribes) in the Ethiopian highlands compared to the Fertile Crescent might be reasons for this distinctness [5], [94]. At K = 10, 277 Ethiopian landraces were assigned to either of two groups comprising naked (G1, 194 lines) and hulled types (G5, 83 lines). Both groups were further sub-structured according to their row type of the ear (S5 Fig.).

Although morphologically diverse, Ethiopian hulled barleys harboured a relatively low level of nucleotide diversity as indicated by the lowest gene diversity value, the second lowest level for allelic richness and only five group specific rare alleles. These results provide further evidence that Ethiopian barleys went through a major genetic bottleneck followed by adaptation to climatic (e.g. rainfall patterns, altitude) and edaphic conditions in the Abyssinian highlands (Table 4, S6 Table) [87], [95][98]. It is interesting that the early flowering haplotype IV at HvCEN (which derived from major haplotype II) predominated in genotypes assigned to G5 (89%) [80]; S1 Table). Furthermore, based on our preliminary data set at PpdH1 (Sharma et al. unpublished), most lines from G5 carry insensitive alleles, thus in combination with haplotype IV at HvCEN, providing favorable alleles for the two growing seasons of barley in Ethiopia - Meher and Belg [87].

Interestingly, Ethiopian naked barleys (G1) harbored 57 group specific alleles (second largest number found) and higher value for allelic richness compared to G9. Pairwise comparison of Fst values between Structure inferred groups showed that G1 is most closely related to G5 (0.35) supporting a common origin of Ethiopian barleys. However, G1 and G9 (the two naked groups) are genetically less related (0.49), while G9 is closer connected to G7 (Fst 0.29). Thus, our data suggest at least two evolutionary lineages of naked barleys, both of which probably originated in the eastern Fertile Crescent from a monophyletic natural mutation (17 kb deletion harboring an ethylene response factor (ERF) family transcription factor gene) at the nud locus on chromosome 7H [99][101]. However, resequencing larger germplasm sets at the nud locus are required to test this hypothesis and to shed more light on the origin of naked barley. Interestingly, outside Ethiopia, two-rowed naked landraces were mainly sampled from Iraq and Iran, while six-rowed naked types were collected further east (Afghanistan) [102], [103].

Interestingly, the second highest variation explained among groups by AMOVA was found between all Ethiopian naked vs. all non-Ethiopian naked barleys (32.02%) (and thus supporting two evolutionary lineages for naked barley), while all Ethiopian hulled vs. all non-Ethiopian hulled barleys explained only 18.50% (S4 Table). AMOVA of all hulled vs. all naked types explained 16.17% of variation, which is nearly two-fold higher compared to the variation between all two-rowed vs. all six-rowed types (8.87%). Also AMOVA of all hulled Ethiopian vs. all naked Ethiopian types (26.69) compared to all two-rowed Ethiopian vs. all six-rowed Ethiopian landraces (6.38) suggested that hulled and naked genepools are genetically more distant than the two-rowed and six-rowed clusters. Hulled and naked types probably evolved largely independent under cultivation and were domesticated for different end-use qualities [35].

Conclusions

The Lrc1485 harbors great genetic diversity. However, the collection size of 1485 genotypes is not manageable in most phenotypic studies, and smaller panels are needed. The legacy CRS Lrc648 established here is best suited for multi-environmental field testing under various climates. However, to work with even more manageable sets, we suggest dividing the Lrc648 into two-rowed and six-rowed subpanels, depending on the trait of interest (Table 5). Increased marker coverage [80] and precise phenotypic data for Lrc648 will help to identify candidate genes also for agronomic and adaptation-related traits using GWAS [27]. Re-sequencing candidate genes [104], [105] or genomic regions underlying quantitative traits using next generation sequencing (NGS) approaches [106][109] can be applied for Lrc648.

Seeds of the Lrc648 can be requested in small quantities from IPK. Seed delivery just awaits the Standard Material Transfer Agreement (SMTA) procedure.

Supporting Information

S1 Fig.

Distribution of 45 SSR markers used across the seven linkage groups of barley.

https://doi.org/10.1371/journal.pone.0116164.s001

(TIF)

S2 Fig.

Scatter plot of 1485 barley landraces based on Principal Component Analysis calculated from 42 SSR data. a) for K = 4, b) for K = 10. Colours correspond to the different Structure inferred groups.

https://doi.org/10.1371/journal.pone.0116164.s002

(TIF)

S3 Fig.

Geographical distribution of 1485 landraces over Köppen climate zones according to Structure inferred groups at K = 4. Each group (G1-G4) and admixed types were separately plotted. (a) G1; (b) G2; (c) G3; (d) G4; (e) admixed types. Climate abbreviations are explained in S6 Table.

https://doi.org/10.1371/journal.pone.0116164.s003

(TIF)

S4 Fig.

Geographical distribution of 1485 landraces over various Köppen climate zones according to Structure inferred groups at K = 10. Each group (G1-G10) and admixed types were separately plotted. (a) G1; (b) G2; (c) G3; (d) G4; (e) G5; (f) G6; (g) G7; (h) G8; (i) G9; (j) G10; (k) admixed types. Abbreviations are explained in S6 Table.

https://doi.org/10.1371/journal.pone.0116164.s004

(TIF)

S5 Fig.

Individual PCA's for each Structure inferred group at K = 10. Each plot represents a single group: (a) G1, (b) G2, (c) G3, (d) G4, (e) G5, (f) G6, (g) G7, (h) G8, (i) G9, (j) G10. Blue circles indicate two-rowed and red circles six-rowed barleys.

https://doi.org/10.1371/journal.pone.0116164.s005

(TIF)

S6 Fig.

Correlograms showing spatial genetic autocorrelation patterns among: (a) genetic distance and geographical distance; (b) genetic distance and longitude; (c) genetic distance and latitude; (d) genetic distance and annual mean temperature; (e) genetic distance and mean diurnal range; (f) genetic distance and mean temperature of warmest quarter; (g) genetic distance and annual precipitation. The x-axis represents distinct classes and the y-axis represents the mantel r values.

https://doi.org/10.1371/journal.pone.0116164.s006

(TIF)

S7 Fig.

Comparing the sampling efficiency based on Mstrat and random sampling to capture most efficiently genetic diversity to establish a core reference set. Average diversity score calculated based on allelic richness was plotted against the sample size. Red circles indicate scores of the core collection using the M strategy and blue circles indicate scores of randomly selected accessions.

https://doi.org/10.1371/journal.pone.0116164.s007

(TIF)

S8 Fig.

Scatter plot of Lrc1485 and Lrc648 based on PCA calculated from 42 SSR data. Landraces selected for Lrc648 are indicated in red colour.

https://doi.org/10.1371/journal.pone.0116164.s008

(TIF)

S9 Fig.

Linkage Disequilibrium (LD) display of 1485 barley landraces using 42 SSR markers. LD was calculated in TASSEL 2.1(www.maizegenetics.net/tassel) using 1000 permutations. Markers are arranged according to the genetic positions on barley genome (see S1 Fig.).

https://doi.org/10.1371/journal.pone.0116164.s009

(TIF)

S10 Fig.

Evolutionary relationships of 1485 barley landraces I. The evolutionary history was inferred by a) a Neighbor-Joining tree, and b) a Neighbor-Joining strict consensus tree computed in SplitsTree software.

https://doi.org/10.1371/journal.pone.0116164.s010

(TIF)

S11 Fig.

Evolutionary relationships of 1485 barley landraces II. The Neighbor-Net planar graph of uncorrected p-distances visualizes the high amount of reticulation in the collection (Taxa = 1485; Chars = 372; Fit = 90,615; Splits = 4604).

https://doi.org/10.1371/journal.pone.0116164.s011

(TIF)

S1 Table.

The Lrc1485 collection. a) Details of accession names, their taxonomical designations, row type and kernel coverage, collection sites, collection missions and all other information available from IPK genebank documentation System (GEBIS); and b) Structure assignments to groups K = 2 to K = 20, phenotypic data for Lrc648 and haplotype information for HvCEN obtained from [74], if available are given. Accessions selected for the CRS Lrc648 are indicated. Six accessions were excluded from further analysis due to large extent of missing values. NS – not selected for core group of 648 genotypes; *lost during single seed descent and multiplication.

https://doi.org/10.1371/journal.pone.0116164.s012

(XLSX)

S2 Table.

Final SSR data set for the whole collection of 1491 barley landraces investigated. Fragment sizes and 0/1 matrix are provided.

https://doi.org/10.1371/journal.pone.0116164.s013

(XLSX)

S3 Table.

Pairwise comparison of Fst values between the Structure inferred groups groups a) for K = 4 and b) for K = 10. Significance of P-values computed after 1000 permutations are represented above diagonal and the Fst values are presented below.

https://doi.org/10.1371/journal.pone.0116164.s014

(DOCX)

S4 Table.

Analysis of Molecular Variance (AMOVA). Summary of partitioning of genetic variation among and within different groups of Lrc1485.

https://doi.org/10.1371/journal.pone.0116164.s015

(DOCX)

S5 Table.

Mantel correlogram tables. a) Mantel correlogram tables between: genetic distance and geographic distance; b) genetic distance and longitude difference matrix; c) genetic distance and latitude difference matrix; d) genetic distance and annual mean temperature; e) genetic distance and mean diurnal range; f) genetic distance and temperature of warmest quarter; and g) genetic distance and annual precipitation. 1different distance classes are shown; 2lower and upper boundary values for each class; 3number of pairs for which the correlation was calculated within each distance class; 4the mantle correlation for each class; 5the significance of mantel correlation for each class.

https://doi.org/10.1371/journal.pone.0116164.s016

(DOCX)

S6 Table.

Overview of Köppen climates prevailing in Structure inferred groups. Variety - number of different climates within the group; *the few geographical outliers were not removed. (XLSX)

https://doi.org/10.1371/journal.pone.0116164.s017

(XLSX)

S7 Table.

Comparison of diversity statistics for different sample sizes of core groups generated from 1485 accessions as well as Lrc1485, Lrc1014, Lrc648 and Lrc648r using 42 SSR markers and climatic variables. N - number of accessions; AN - average allele number; GD - gene diversity; PIC - polymorphism information content; MAF - average major allele frequency.

https://doi.org/10.1371/journal.pone.0116164.s018

(DOCX)

S8 Table.

Comparison of phenotypic diversity between the core set Lrc648 (based on Mstrat) and the random set Lrc648r. Min. - minimum; Max. – maximum; SD. - standard deviation of the measured traits heading date (Hd) (in days to flowering), spike length (Sl) (in cm) and plant height (Ht) (cm).

https://doi.org/10.1371/journal.pone.0116164.s019

(DOCX)

Acknowledgments

We thank the Federal ex situ Genebank Gatersleben, Michael Grau and Helmut Knüpffer for providing seeds, passport-, and characterization & evaluation data for all accession numbers. We thank Frank Ordon, Kerstin Neumann, Joanne Russell, Brian Steffenson, David Bryant, Stephan Weise and Celestine Wabila for discussions, support and critical reading of the manuscript. We are greatly indebted to Dr. C.O. Qualset and to an anonymous reviewer, who helped us to improve an earlier version of this manuscript. We thank the following for excellent technical assistance: Kathrin Baake, Birgit Dubsky, Heike Harms, Christiane Kehler, Ute Krajewski, Jürgen Marlow, Marita Nix, Peter Schreiber, Kerstin Wolf, the “Experimental Fields and Nurseries” and “Genome Diversity” groups at IPK and all students involved. We are grateful for the support from all GENOBAR consortium partners.

Author Contributions

Conceived and designed the experiments: BK AG RKP. Performed the experiments: RKP RS AW HÖ BK. Analyzed the data: RKP RS AW BK. Contributed reagents/materials/analysis tools: RS AW HÖ. Wrote the paper: BK RKP AW RS AG.

References

  1. 1. Zohary D, Hopf M, Weiss E (2012) Domestication of plants in the old world, 4th edition. Oxford: Oxford Univ Press. 264 p.
  2. 2. Nevo E, Ordentlich A, Beiles A, Raskin I (1992) Genetic divergence of heat production within and between the wild progenitors of wheat and barley: evolutionary and agronomical implications. Theor Appl Genet 84:958–962.
  3. 3. Weltzien E (1988) Evaluation of barley (Hordeum vulgare L.) landrace populations originating from different growing regions in the Near East. Plant Breed 101:95–106.
  4. 4. Graner A, Bjørnstad Å, Konishi T, Ordon F (2003) Molecular diverisity of the barley genome. In: von Bothmer R, van Hintum T, Knüpffer H, Sato Keditors. Diversity in Barley (Hordeum vulgare). Elsevier Science B. V. Amsterdam, The Netherlands. pp. 121–141.
  5. 5. Ward DJ (1962) Some evolutionary aspects of certain morphological characters in a world collection of barley. USDA Technical Bulletin 1276. Washington, D.C.
  6. 6. Doebley JF, Gaut BS, Smith BD (2006) The molecular genetics of crop domestication. Cell 127:1309–1321.
  7. 7. Kilian B, Ozkan H, Kohl J, Haseler A Von, Barale F, et al. (2006) Haplotype structure at seven barley genes: relevance to gene pool bottlenecks, phylogeny of ear type and site of barley domestication. Mol Genet Genom 276:230–241.
  8. 8. Prada D (2009) Molecular population genetics and agronomic alleles in seed banks: Searching for a needle in a haystack? J Exp Bot 60:2541–2552.
  9. 9. Saghai Maroof MA, Biyashev RM, Yang GP, Zhang Q, Allard RW (1994) Extraordinarily polymorphic microsatellite DNA in barley: species diversity, chromosomal locations, and population dynamics. Proc Natl Acad Sci USA 91:5466–5470.
  10. 10. Fischbeck G (2003) Diversification through breeding. In: von Bothmer R, van Hintum T, Knüpffer H, Sato Keditors. Diversity in Barley (Hordeum vulgare). Elsevier Science B. V. Amsterdam, The Netherlands. pp. 29–52.
  11. 11. Jones H, Civáň P, Cockram J, Leigh FJ, Smith LM, et al. (2011) Evolutionary history of barley cultivation in Europe revealed by genetic analysis of extant landraces. BMC Evol Biol 11:320.
  12. 12. Yahiaoui S, Cuesta-Marcos A, Gracia MP, Medina B, Lasa JM, et al. (2014) Spanish barley landraces outperform modern cultivars at low-productivity sites. Plant Breed 133:218–226.
  13. 13. Anonymous (2008) Global strategy for the ex-situ conservation and use of barley germplasm. pp. 1–65.
  14. 14. Knüpffer H (2009) Triticeae genetic resources in ex situ genebank collections. In: Muehlbauer G, Feuillet C, editors. Genetics and Genomics of the Triticeae. Plant Genetics and Genomics: Crops and Models 7. Springer Science + Business Media LLC New York. pp. 31–80.
  15. 15. Kihara H (1983) Origin and history of “Daruma”, a parental variety of Norin 10. In: Sakamoto Sed. Proc. 6th Int. Wheat Genetics Symp. Kyoto University Press, Kyoto, Japan, Nov. 28 - Dec 3. pp 13–19.
  16. 16. Hoisington D, Khairallah M, Reeves T, Ribaut JM, Skovmand B, et al. (1999) Plant genetic resources: what can they contribute toward increased crop productivity? Proc Natl Acad Sci U S A 96:5937–5943.
  17. 17. Newton AC, Akar T, Baresel JP, Bebeli PJ, Bettencourt E, et al. (2010) Cereal landraces for sustainable agriculture. A review. Agron Sustain Dev 30:237–269.
  18. 18. Bailey-Serres J, Fukao T, Ronald P, Ismail A, Heuer S, et al. (2010) Submergence tolerant rice: SUB1's journey from landrace to modern cultivar. Rice 3:138–147.
  19. 19. Piffanelli P, Ramsay L, Waugh R, Benabdelmouna A, D’Hont A, et al. (2004) A barley cultivation-associated polymorphism conveys resistance to powdery mildew. Nature 430:887–891.
  20. 20. Graner A, Bauer E (1993) RFLP mapping of the rym4 virus resistance gene in barley. Theor Appl Genet 86:689–693.
  21. 21. Sutton T, Baumann U, Hayes J, Collins NC, Shi B-J, et al. (2007) Boron-toxicity tolerance in barley arising from efflux transporter amplification. Science 318:1446–1449.
  22. 22. Mackay I, Powell W (2007) Methods for linkage disequilibrium mapping in crops. Trends Plant Sci 12:57–63.
  23. 23. Yu J, Pressoir G, Briggs WH, Vroh Bi I, Yamasaki M, et al. (2006) A unified mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat Genet 38:203–208.
  24. 24. Malosetti M, van der Linden CG, Vosman B, van Eeuwijk FA (2007) A mixed-model approach to association mapping using pedigree information with an illustration of resistance to Phytophthora infestans in potato. Genetics 175:879–889.
  25. 25. Tanksley SD, McCouch SR (1997) Seed banks and molecular maps: Unlocking genetic potential from the wild. Science 277:1063–1066.
  26. 26. Hamblin MT, Buckler ES, Jannink JL (2011) Population genetics of genomics-based crop improvement methods. Trends Genet 27:98–106.
  27. 27. Waugh R, Jannink JL, Muehlbauer GJ, Ramsay L (2009) The emergence of whole genome association scans in barley. Curr Opin Plant Biol 12:218–222.
  28. 28. Hamblin MT, Close TJ, Bhat PR, Chao SM, Kling JG, et al. (2010) Population structure and linkage disequilibrium in US barley germplasm: Implications for association mapping. Crop Sci 50:556–566.
  29. 29. Mezaka I, Legzdina L, Waugh R, Close T, Rostoks N (2013) Genetic Diversity in Latvian spring barley association mapping population. In: Zhang G, Li C, Liu Xeditors. Advance in Barley Sciences. Springer Netherlands. pp. 25–35.
  30. 30. Tondelli A, Xu X, Moragues M, Sharma R, Schnaithmann F, et al. (2013) Structural and temporal variation in genetic diversity of European spring two-row barley cultivars and association mapping of quantitative traits. Plant Genome 6:1–14.
  31. 31. Yahiaoui S, Igartua E, Moralejo M, Ramsay L, Molina-Cano JL, et al. (2008) Patterns of genetic and eco-geographical diversity in Spanish barleys. Theor Appl Genet 116:271–282.
  32. 32. Pasam RK, Sharma R, Malosetti M, van Eeuwijk FA, Haseneyer G, et al. (2012) Genome-wide association studies for agronomical traits in a world-wide spring barley collection. BMC Plant Biol 12:16.
  33. 33. Muñoz-Amatriaín M, Cuesta-Marcos A, Endelman JB, Comadran J, Bonman JM, et al. (2014) The USDA barley core collection: genetic diversity, population structure, and potential for genome-wide association studies. PLOS One 9:e94688.
  34. 34. Russell JR, Booth A, Fuller JD, Baum M, Ceccarelli S, et al. (2003) Patterns of polymorphism detected in the chloroplast and nuclear genomes of barley landraces sampled from Syria and Jordan. Theor Appl Genet 107:413–421.
  35. 35. Pandey M, Wagner C, Friedt W, Ordon F (2006) Genetic relatedness and population differentiation of Himalayan hulless barley (Hordeum vulgare L.) landraces inferred with SSRs. Theor Appl Genet 113:715–729.
  36. 36. Gong X, Westcott S, Li C, Yan G, Lance R, et al. (2009) Comparative analysis of genetic diversity between Qinghai-Tibetan wild and Chinese landrace barley. Genome 52:849–861.
  37. 37. Comadran J, Russell JR, Booth A, Pswarayi A, Ceccarelli S, et al. (2011) Mixed model association scans of multi-environmental trial data reveal major loci controlling yield and yield related traits in Hordeum vulgare in Mediterranean environments. Theor Appl Genet 122:1363–1373.
  38. 38. Rodriguez M, Rau D, O’Sullivan D, Brown AHD, Papa R, et al. (2012) Genetic structure and linkage disequilibrium in landrace populations of barley in Sardinia. Theor Appl Genet 125:171–184.
  39. 39. Jilal A, Grando S, Henry RJ, Lee LS, Rice N, et al. (2008) Genetic diversity of ICARDA's worldwide barley landrace collection. Genet Resour Crop Evol 55:1221–1230.
  40. 40. Mansfeld R (1950) Das morphologische System der Saatgerste, Hordeum vulgare L. s. l. Der Züchter 20:8–24.
  41. 41. Russell J, Dawson IK, Flavell AJ, Steffenson B, Weltzien E, et al. (2011) Analysis of>1000 single nucleotide polymorphisms in geographically matched samples of landrace and wild barley indicates secondary contact and chromosome-level differences in diversity around domestication genes. New Phytol 191:564–578.
  42. 42. Lehmann CO, Mansfeld R (1957) Zur Technik der Sortimentserhaltung. Die Kultpfl 5:108–138.
  43. 43. Sackville Hamilton NR, Engels JMM, van Hintum T, Koo B, Smale M (2002) Accession management: combining or splitting accessions as a tool to improve germplasm management efficiency. IPGRI Tech.
  44. 44. Thiel T, Michalek W, Varshney RK, Graner A (2003) Exploiting EST databases for the development and characterization of gene-derived SSR-markers in barley (Hordeum vulgare L.). Theor Appl Genet 106:411–422.
  45. 45. Stein N, Prasad M, Scholz U, Thiel T, Zhang H, et al. (2007) A 1,000-loci transcript map of the barley genome: new anchoring points for integrative grass genomics. Theor Appl Genet 114:823–839.
  46. 46. Haseneyer G, Stracke S, Paul C, Einfeldt C, Broda A, et al. (2010) Population structure and phenotypic variation of a spring barley world collection set up for association studies. Plant Breed 129:271–279.
  47. 47. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus Genotype Data. Genetics 155:945–959.
  48. 48. Falush D, Stephens M, Pritchard JK (2003) Inference of population structure using multilocus genotype gata: Linked loci and correlated cllele Frequencies. Genetics 164:1567–1587.
  49. 49. Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software structure: a simulation study. Mol Ecol 14:2611–2620.
  50. 50. Jakob SS, Rodder D, Engler JO, Shaaf S, Ozkan H, et al. (2014) Evolutionary history of wild barley (Hordeum vulgare subsp. spontaneum) analyzed using multilocus sequence data and paleodistribution modeling. Genome Biol Evol 6:685–702.
  51. 51. Hammer Ø, Harper DAT, Ryan PD (2001) Paleontological statistics software package for education and data analysis. Palaeontol Electron 4:9.
  52. 52. Huson DH, Bryant D (2006) Application of phylogenetic networks in evolutionary studies. Mol Biol Evol 23:254–267.
  53. 53. Liu K, Muse S V (2005) PowerMarker: An integrated analysis environment for genetic marker analysis. Bioinformatics 21:2128–2129.
  54. 54. Botstein D, White RL, Skolnick M, Davis RW (1980) Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am J Hum Genet 32:314–331.
  55. 55. Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: A new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour 10:564–567.
  56. 56. Weir BS, Cockerham CC (1984) Estimating F-Statistics for the analysis of population structure. Evolution (NY) 38:1358–1370.
  57. 57. Goudet J (1995) FSTAT (Version 1.2): A Computer Program to Calculate F-Statistics. J Hered 86:485–486.
  58. 58. Leberg PL (2002) Estimating allelic richness: Effects of sample size and bottlenecks. Mol Ecol 11:2445–2449.
  59. 59. Weir BS (1996) Genetic data analysis II: Methods for discrete population genetic data. 376.
  60. 60. Chakraborty R, Jin L (1993) Determination of relatedness between individuals using DNA fingerprinting. Hum Biol an Int Rec Res 65:875–895.
  61. 61. Escudero A (2003) Spatial analysis of genetic diversity as a tool for plant conservation. Biol Conserv 113:351–365.
  62. 62. Rosenberg MS, Anderson CD (2011) PASSaGE: Pattern Analysis, Spatial Statistics and Geographic Exegesis. Version 2. Methods Ecol Evol 2:229–232.
  63. 63. Kottek M, Grieser J, Beck C, Rudolf B, Rubel F (2006) World map of Köppen−Geiger climate classification. Meteorol Zeitschrift 15:259–263.
  64. 64. Glaszmann JC, Kilian B, Upadhyaya HD, Varshney RK (2010) Accessing genetic diversity for crop improvement. Curr Opin Plant Biol 13:167–173.
  65. 65. Acquaah G (2012) Principles of plant genetics and breeding, 2nd Edition. Wiley-Blackwell. 758 p.
  66. 66. Gouesnard B (2001) MSTRAT: An algorithm for building germplasm core collections by maximizing allelic or phenotypic richness. J Hered 92:93–94.
  67. 67. Schoen DJ, Brown AH (1993) Conservation of allelic richness in wild crop relatives is aided by assessment of genetic markers. Proc Natl Acad Sci U S A 90:10623–10627.
  68. 68. Keilwagen J, Kilian B, Ozkan H, Babben S, Perovic D, et al. (2014) Separating the wheat from the chaff - a strategy to utilize plant genetic resources from ex-situ genebanks. Sci Rep 4:5231.
  69. 69. Backes G, Orabi J, Wolday A, Yahyaoui A, Jahoor A (2009) High genetic diversity revealed in barley (Hordeum vulgare) collected from small-scale farmer's fields in Eritrea. Genet Resour Crop Evol 56:85–97.
  70. 70. Leino MW, Hagenblad J (2010) Nineteenth century seeds reveal the population genetics of landrace barley (Hordeum vulgare). Mol Biol Evol 27:964–973.
  71. 71. Varshney RK, Baum M, Guo P, Grando S, Ceccarelli S, et al. (2010) Features of SNP and SSR diversity in a set of ICARDA barley germplasm collection. Mol Breed 26:229–242.
  72. 72. Bellucci E, Bitocchi E, Rau D, Nanni L, Ferradini N, et al. (2013) Population structure of barley landrace populations and gene-flow with modern varieties. PLOS ONE 8:e83891.
  73. 73. Malysheva-Otto LV, Ganal MW, Roder MS (2006) Analysis of molecular diversity, population structure and linkage disequilibrium in a worldwide survey of cultivated barley germplasm (Hordeum vulgare L.). BMC Genet 7:6.
  74. 74. Long NV, Dolstra O, Malosetti M, Kilian B, Graner A, et al. (2013) Association mapping of salt tolerance in barley (Hordeum vulgare L.). Theor Appl Genet 126:2335–2351.
  75. 75. Abdel-Ghani AH, Neumann K, Wabila C, Sharma R, Dhanagond S, et al. (2014) Diversity of germination and seedling traits in a spring barley (Hordeum vulgare L.) collection under drought simulated conditions. Genet Resour Crop Evol. DOI 10.1007/s10722-014-0152-z
  76. 76. Patterson N, Price AL, Reich D (2006) Population structure and eigenanalysis. PLOS Genet 2:2074–2093.
  77. 77. Castillo A, Dorado G, Feuillet C, Sourdille P, Hernandez P (2010) Genetic structure and ecogeographical adaptation in wild barley (Hordeum chilense Roemer et Schultes) as revealed by microsatellite markers. BMC Plant Biol 10:266.
  78. 78. Cuesta-Marcos A, Szucs P, Close TJ, Filichkin T, Muehlbauer GJ, et al. (2010) Genome-wide SNPs and re-sequencing of growth habit and inflorescence genes in barley: implications for association mapping in germplasm arrays varying in size and structure. BMC Genomics 11:707 doi:10.1186/1471-2164-11-707.
  79. 79. Turner A, Beales J, Faure S, Dunford RP, Laurie DA (2005) The Pseudo-Response Regulator Ppd-H1 provides adaptation to photoperiod in barley. Science 310:1031–1034.
  80. 80. Comadran J, Kilian B, Russell J, Ramsay L, Stein N, et al. (2012) Natural variation in a homolog of Antirrhinum CENTRORADIALIS contributed to spring growth habit and environmental adaptation in cultivated barley. Nat Genet 44:1388–1392.
  81. 81. Yan L, Loukoianov A, Tranquilli G, Helguera M, Fahima T, et al. (2003) Positional cloning of the wheat vernalization gene VRN1. Proc Natl Acad Sci USA 100:6263–6268.
  82. 82. Szucs P, Skinner JS, Karsai I, Cuesta-Marcos A, Haggard KG, et al. (2007) Validation of the VRN-H2/VRN-H1 epistatic model in barley reveals that intron length variation in VRN-H1 may account for a continuum of vernalization sensitivity. Mol Genet Genomics 277:249–261.
  83. 83. Yan L, Loukoianov A, Blechl A, Tranquilli G, Ramakrishna W, et al. (2004) The wheat VRN2 gene is a flowering repressor down-regulated by vernalization. Science 303:1640–1644.
  84. 84. Yan L, Fu D, Li C, Blechl A, Tranquilli G, et al. (2006) The wheat and barley vernalization gene VRN3 is an orthologue of FT. Proc Natl Acad Sci USA 103:19581–19586.
  85. 85. Stracke S, Haseneyer G, Veyrieras JB, Geiger HH, Sauer S, et al. (2009) Association mapping reveals gene action and interactions in the determination of flowering time in barley. Theor Appl Genet 118:259–273.
  86. 86. Dubcovsky J, Chen C, Yan L (2005) Molecular characterization of the allelic variation at the VRN-H2 vernalization locus in barley. Mol Breed 15:395–407.
  87. 87. Tanto Hadado T, Rau D, Bitocchi E, Papa R (2010) Adaptation and diversity along an altitudinal gradient in Ethiopian barley (Hordeum vulgare L.) landraces revealed by molecular analysis. BMC Plant Biol 10:121.
  88. 88. Alderman PD, Quilligan E, Asseng S, Ewert F, Reynolds MPeditors (2013) Proceedings of the workshop Modeling Wheat Response to High Temperature. CIMMYT, Mexico June 19–21.
  89. 89. Semenov MA, Stratonovitch P (2013) Designing high-yielding wheat ideotypes for a changing climate. Food Energy Secur 2:185–196.
  90. 90. Allard RW, Zhang Q, Saghai Maroof MA, Muona OM (1992) Evolution of multilocus genetic structure in an experimental barley population. Genetics 131:957–969.
  91. 91. Bjørnstad A, Demissie A, Kilian A, Kleinhofs A (1997) The distinctness and diversity of Ethiopian barleys. Theor Appl Genet 94:514–521.
  92. 92. Orabi J, Backes G, Wolday A, Yahyaoui A, Jahoor A (2007) The Horn of Africa as a centre of barley diversification and a potential domestication site. Theor Appl Genet 114:1117–1127.
  93. 93. Igartua E, Moralejo M, Casas AM, Torres L, Molina-Cano JL (2013) Whole-genome analysis with SNPs from BOPA1 shows clearly defined groupings of Western Mediterranean, Ethiopian, and Fertile Crescent barleys. Genet Resour Crop Evol 60:251–264.
  94. 94. Allard RW, Kahler AL, Weir BS (1972) The effect of selection on esterase allozymes in a barley population. Genetics 72:489–503.
  95. 95. Schiemann E (1939) Gedanken zur Genzentrentheorie Vavilovs. Naturwissenschaften 27:394–401.
  96. 96. Schiemann E (1951) New results on the history of cultivated cereals. J Hered 5:305–320.
  97. 97. Bjørnstad A, Abay F (2010) Multivariate patterns of diversity in Ethiopian barleys. Crop Sci 50:1579.
  98. 98. Tsehaye Y, Bjørnstad Å, Abay F (2012) Phenotypic and genotypic variation in flowering time in Ethiopian barleys. Euphytica 188:309–323.
  99. 99. Taketa S, Kikuchi S, Awayama T, Yamamoto S, Ichii M, et al. (2004) Monophyletic origin of naked barley inferred from molecular analyses of a marker closely linked to the naked caryopsis gene (nud). Theor Appl Genet 108:1236–1242.
  100. 100. Pourkheirandish M, Komatsuda T (2007) The importance of barley genetics and domestication in a global perspective. Ann Bot 100:999–1008.
  101. 101. Taketa S, Amano S, Tsujino Y, Sato T, Saisho D, et al. (2008) Barley grain with adhering hulls is controlled by an ERF family transcription factor gene regulating a lipid biosynthesis pathway. Proc Natl Acad Sci USA 105:4062–4067.
  102. 102. Schiemann E (1943) Entstehung der Kulturpflanzen. Ergebn Biol 19:409–452.
  103. 103. Kilian B, Knüpffer H, Hammer K (2014) Elisabeth Schiemann (1881–1972): a pioneer of crop plant research, with special reference to cereal phylogeny. Genet Resour Crop Evol 61:89–106.
  104. 104. Bhullar NK, Street K, Mackay M, Yahiaoui N, Keller B (2009) Unlocking wheat genetic resources for the molecular identification of previously undescribed functional alleles at the Pm3 resistance locus. Proc Natl Acad Sci USA 106:9519–9524.
  105. 105. Bhullar NK, Zhang Z, Wicker T, Keller B (2010) Wheat gene bank accessions as a source of new alleles of the powdery mildew resistance gene Pm3: a large scale allele mining project. BMC Plant Biol 10:88.
  106. 106. B, Graner A (2012) NGS technologies for analyzing germplasm diversity in genebanks. Brief Funct Genomics 11:38–50.
  107. 107. Mayer KFX, Waugh R, Brown JWS, Schulman A, Langridge P, et al. (2012) A physical, genetic and functional sequence assembly of the barley genome. Nature 491:711–716.
  108. 108. Mascher M, Richmond TA, Gerhardt DJ, Himmelbach A, Clissold L, et al. (2013) Barley whole exome capture: a tool for genomic research in the genus Hordeum and beyond. Plant J 76:494–505.
  109. 109. Varshney RK, Terauchi R, McCouch SR (2014) Harvesting the promising fruits of genomics: applying genome sequencing technologies to crop breeding. PLOS Biol 12:e1001883.