Abstract
The USDA-ARS National Plant Germplasm System Malus collection is maintained by the Plant Genetic Resources Unit (PGRU) in Geneva, NY. In the 1990s, a core subset of 258 trees was hand-selected to be representative of the grafted Malus collection. We used a combination of genotypic and phenotypic data to compare the diversity of the 198 diploid trees in the original core subset with that of 2114 diploid trees in the grafted field collection for which data were available. The 198 trees capture 192 of the 232 total microsatellite alleles and have 78 of the 95 phenotypic characters. An addition of 67 specific individuals increases the coverage to 100% of the allelic and phenotypic character states. Several de novo core sets that capture all the allelic and phenotypic character states in 100 individuals are also provided. Use of these proposed sets of individuals will help ensure that a broad range of Malus diversity is included in evaluations that use the core subset of grafted trees in the PGRU collection.
Large field collections of plant genetic resources, although critical for many clonal and perennial crops, are expensive to maintain and can be unwieldy for research purposes. In the 1980s, Brown and colleagues described a need to identify representative subsets of larger collections. These collections could be given higher priority for evaluations and could be made widely available at multiple locations, thus increasing access to associated data for the gene bank user community (Brown, 1989; Rubenstein et al., 2006). Core collections of crops have since been developed using a number of different strategies, including selection of materials from geographic or ecological groups or selection of highly differentiated materials. These stratified sampling techniques seek to maximize allelic richness and capture the available genetic and phenotypic diversity (Schoen and Brown, 1995). Various combinations of morphology, climatic, geographic, genetic distances, and genetic diversity assessments have been implemented to develop core collections for diverse crops (Balakrishnan et al., 2000; Balfourier et al., 2007; Bisht et al., 1998; Chavarriaga-Aguirre et al., 1999; Franco et al., 2001, 2006; Jansen and van Hintum, 2007; Lázaro and Aguinagalde, 2006; McKhann et al., 2004; Ronfort et al., 2006; Wang et al., 2006a, 2006b, 2007). Many researchers have turned to optimization algorithms such as MSTRAT, which seek to maximize allelic and/or phenotypic richness by ensuring that the maximum number of character states is represented in the core set of individuals (Gouesnard et al., 2001). Further studies have examined the value of relying on neutral genetic markers to estimate collection diversity (Bataillon et al., 1996; Reeves et al., 2012).
Apple (Malus ×domestica) is a globally produced clonal crop for which field collections play a crucial role in preserving genetic diversity and providing breeding material. In particular, apple production is currently threatened by disease, pest susceptibility, suboptimal cold and heat tolerance, minimal resistance to drought and wet soils, undesirable storage and transport characteristics, and expensive production methods (Yue et al., 2012). Simultaneously, there is rising consumer demand for organic produce and heritage apple cultivars. The USDA-ARS National Plant Germplasm System (NPGS) apple collection in Geneva, NY, has key genetic resources that can be used in breeding and research programs to address the threats to apple crop production and to provide disease- and climate-hardy heritage cultivars to interested growers. The grafted portion of this collection is a set of 3442 individuals maintained for their unique genotypes. These clonally maintained individuals are primarily cultivars but also include wild Malus species representatives. The trees were in an orchard that was planted on seedling rootstock when the phenotypic data for this research were collected. Since that time, the entire collection has been repropagated onto EMLA 7 rootstock. The grafted collection is complimented by a “species” collection consisting of over 3000 seedling trees grown from wild-collected seeds. The current study focuses on the grafted “clonally propagated” part of the NPGS apple collection.
Core collections have been developed for both the entire Malus grafted collection and for subsets of wild species collections using either phenotypic and/or genetic data. For example, over 1000 M. sieversii seedlings from the “species” collection that represent eight collection sites in Kazakhstan have been genotyped using microsatellite markers and evaluated for key phenotypic traits. In 2005, Volk et al. proposed two sets of 35 individuals that represent 174 and 278 trees of M. sieversii from Kazakhstan collection sites 6 and 9, respectively (Volk et al., 2005). In 2009, Richards et al. developed a third core set of 35 individuals that captures the measured genotypic and phenotypic diversity from the remaining Kazakhstan collection locations of M. sieversii (Richards et al., 2009a, 2009b). These smaller sets of individuals were designed to capture the diversity of the larger plantings so that a subset of individuals could be targeted for field maintenance as grafted clones and long-term back-up using cryopreservation technologies (Volk et al., 2005). Similarly, a core set of 27 M. orientalis individuals was proposed as a method to capture the genetic diversity represented by 776 seedling trees of M. orientalis in the “species” collection as measured using seven microsatellite markers (Volk et al., 2009).
In the 1990s, a set of 258 individuals was hand-selected to be a representative “core collection” of the grafted collection of apples at the Geneva, NY, repository (Supplemental Table 1). Phenotypic, genotypic, and image data are now available for most of the individuals in the core collection. These individuals have also been distributed to multiple locations, including Washington State University, the University of Illinois, and the University of Minnesota for inclusion in breeding programs as well as for horticultural trait, biotic, and abiotic resistance evaluations (Hokanson et al., 1998). Previous reports have focused on assessing the diversity and genetic relationships among accessions in the core collection (or a portion thereof) using microsatellite markers (Hokanson et al., 1998, 2001; Potts et al., 2012). In this project, we consider the phenotypic and allelic diversity of the diploid trees in the apple core collection in relation to the diversity in the grafted diploid apple collection. We assess the current diversity captured and propose additional individuals from the grafted collection that either complement or replace individuals in the current core collection so that a broader spectrum of diversity is represented.
Materials and Methods
Plant material.
Most of the diploid trees in the grafted PGRU orchards were sampled for potential inclusion in the genetic analyses. Trees were genotyped at nine simple sequence repeat (SSR) loci (see below), and only diploid trees with microsatellite signatures that could be scored in a comparative manner were included in the final data set. The final data set comprised 2114 samples, also referred to as individuals or accessions, each identified with a unique PI number. Of these samples, 198 are part of the current core collection that was proposed in the 1990s (Table 1).
Taxonomic diversity in the grafted apple collection of the USDA-ARS Plant Genetic Resources Unit (Geneva, NY).z
Phenotypic data.
Phenotypic data were collected in orchard evaluations performed in Geneva, NY. Available data were downloaded from the Genetic Resources Information Network database [GRIN (U.S. Department of Agriculture, 2012)]. The data for each accession were collected in a single growing season after the trees had reached maturity (not necessarily in the same year for all of the accessions). Continuous descriptors were placed into categories to facilitate core collection assessments. Phenotypic data were primarily related to fruit trait characteristics. Data were collected on fruit that were mature, defined as when the overcolor on the fruit was apparent. “Fruit weight” was measured as the mean weight of 10 fruit at maturity and data were classified into 50-g increments. “Fruit shape” data were classified as globose, short, flat, conical, ellipsoid, or oblong. “Fruit ground color” at maturity data used the first color descriptor of the GRIN classification: light green, green, light yellow, yellow, orange, brown, pink, red, or purple. Similarly, “overcolor pattern” was classified as blush, striped, splashed, or none. “Fruit overcolor” was measured at maturity and GRIN data were classified into the following categories: none, green, yellow, orange, brown, pink, red, dark red, and purple. “Overcolor intensity” was measured as the percent of overcolor on fruit and placed into categories. “Fruit russet” was measured as the percent of russeting on fruit and placed into categories in 10% increments between 0% and 100%.
“Fruit flesh color” data were reclassified into categories using only the first digit of the numerical code representing flesh colors white, cream, green, yellow, orange, pink, red, or rose red. “Fruit flesh firmness” at maturity was classified as soft, semifirm, firm, or hard according to standards described in GRIN. “Fruit flesh oxidation” was visually measured in cut fruit after 10 min at room temperature. Descriptor values ranged from non-oxidizing (0% to 1%), slightly oxidizing (1% to 4%), oxidizing (5% to 10%), and very oxidizing (greater than 10%). “Fruit juiciness” was measured as a rating of fruit flesh juiciness based on the apple weight to apple volume (specific gravity) and recorded as a mean of five apples at maturity. Code values were very dry (less than 0.75), dry (0.76 to 0.80), medium (0.81 to 0.85), moderately juicy (0.86 to 0.90), and very juicy (greater than 0.90) (Young, 1914). Percent “soluble solids” were measured as the average refractometer readings from juice sampled from three fruit at full maturity. Quantitative values were placed into categories as described on GRIN. “Fruit flesh flavor” was classified qualitatively as aromatic, sweet, subacid, acid, and astringent.
Molecular data.
Genomic DNA was extracted from leaf tissue using DNeasy 96 plant kits (Qiagen, Valencia, CA). Nine previously published SSRs [GD12, GD15, GD96, GD142, GD147, GD162, CH01h01, CH01f02, and CH02d08 (Hokanson et al., 1998; Liebhard et al., 2002)] were amplified in all samples. Primer sequences, amplicon size ranges, and annealing temperatures are listed in Gross et al. (2012a). Amplification and scoring were carried out according to Gross et al. (2012a). Briefly, polymerase chain reaction products were scored in one of two ways. Some were visualized on a slab sequencer (LI-4200; LI-COR, Lincoln, NE) and digital images were manually interpreted using Saga Generation 2 software (LI-COR). Others were visualized on a capillary sequencer (ABI 3730; Applied Biosystems, Foster City, CA) and chromatograms were scored automatically and then corrected manually using GeneMarker software (SoftGenetics, State College, PA). When a single SSR locus was scored using both systems, control individuals of known genotypes were run on both systems to control for allele length differences resulting from instrumentation.
Evaluation and development of core collections.
De novo core collections and amended cores were constructed using the maximization algorithm (M) of Schoen and Brown (1995) as implemented in the program MSTRAT (Gouesnard et al., 2001) with all character states of molecular and phenotypic data weighted equally. Marker data were also analyzed using a modified maximization procedure [M+ (Reeves et al., 2012)] with all character states of molecular data weighted equally. The M+ procedure improves on MSTRAT in several ways, including the ability to discard missing data (rather than counting it as a character state), the ability to consider two alleles at a single locus jointly rather than separately, and the ability to weight loci equally rather than giving a higher weight to more diverse loci (Gouesnard et al., 2001; Reeves et al., 2012).
MSTRAT was used to construct amended cores; these cores included the current core samples with additional samples to maximize diversity. We allowed the program to add samples until all molecular and phenotypic characters were captured and ran 10 independent replicates of the amended core. Results for the amended core replicates were compared; samples that were present in 10 of 10 runs were considered as potential additions to the current core (Table 2; Supplemental Table 2). The modified MSTRAT program, M+, was used to generate de novo cores from the Malus main collection. As a result of computational constraints associated with the large size of the main collection, the M+ program was run, increasing the core size by increments of 20, until the core contained all the possible molecular and phenotypic characters; this was replicated 10 times. Full results of these runs are presented in the supplementary data (Supplemental Table 3).
Potential additional apple accessions that could be amended to the original apple core collection to increase its genetic and phenotypic diversity.z
General genetic diversity statistics were calculated for the 198 diploid samples of the current core, the diploid grafted collection, potential amended cores, and a representative de novo core using GenoDive (Meirmans and Van Tienderen, 2004). The number of alleles and phenotypic character states in the diploid existing core and diploid grafted collection were compared with the proposed cores to evaluate the percentage of variation captured (Table 3).
Diversity retained in diploid core collections for the grafted Geneva, NY, Plant Genetic Resources Unit Malus collection using several core assembly strategies.z
Results
Core collection representation.
In 2011, the PGRU grafted apple collection included 3442 individuals representing 51 species or named hybrids. The original core collection is represented by 258 individuals from 46 species. Of these, 67 individuals are M. ×domestica. The constraints resulting from ploidy levels and missing data left 2114 diploid individuals from the grafted collection included in the analysis. All but 169 of the 2114 diploid grafted collection individuals had phenotypic data available, and all but eight of the198 current core diploid individuals had phenotypic data available.
We used principal component analyses to visualize how well the core sets capture the diversity in multidimensional space represented by the grafted collection. The principal component analyses included samples for which complete genotypes were available: 103 core collection individuals and a total of 1415 individuals from the grafted collection. Although this analysis only visualized 50% of the core and 66% of the total collection, we feel that this is an unbiased sample and allows us to evaluate how well the core collection represents the larger set of genotypes. Regressing the first two principal components (accounting for 38% of the variance) shows a broad overlap (Fig. 1). When visualized as a graph, the three-dimensional plot with the first three principal components is a cloud of points with no obvious divisions and the core selections appear to be randomly distributed within the cloud. The first 11 principal components account for 80% of the variation (data not shown).
Principal component analysis visualizing the first two principal components (accounting for 38% of the variance) of 103 core collection individuals (gray) and 1415 grafted Malus individuals (black) for which complete genotypes were available.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.375
Principal component analysis visualizing the first two principal components (accounting for 38% of the variance) of 103 core collection individuals (gray) and 1415 grafted Malus individuals (black) for which complete genotypes were available.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.375
Principal component analysis visualizing the first two principal components (accounting for 38% of the variance) of 103 core collection individuals (gray) and 1415 grafted Malus individuals (black) for which complete genotypes were available.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.375
Core collections.
Our results from both the MSTRAT and M+ programs indicate that a smaller core collection may be sufficient to capture the molecular diversity documented using the nine SSR loci and the phenotypic diversity described by the 13 selected fruit character traits. Entirely new core sets were identified that capture the complete molecular and phenotypic diversity with 100 samples (Table 3; Supplemental Table 3). If this approach were to be taken, several of the individuals that are currently in the core collection would be retained. An average of 15.4 (ranging from 13 to 17) samples were present in both the current core collection and the de novo core sets, and eight of these samples were present in both the current core and all 10 de novo replicates (Supplemental Table 3).
Alternatively, the current core could be amended with an additional 67 samples to capture 100% of the measured molecular and phenotypic diversity or an additional 36 individuals to capture 95% of the molecular and 89% of the phenotypic variation (Tables 2 and 3). Because this analysis does not include polyploid individuals (either in the core or main collection), additional polyploid samples with unique phenotypic characteristics could be selected from the Geneva collections to complement the core sets identified in these analyses.
Discussion
The genetic composition of the Malus core collection, based on SSR analyses, has been previously published. Studies revealed that eight or nine SSR markers differentiate many of the collection accessions with potential “duplicates” or “sports” being the exceptions (Gross et al., 2012b; Hokanson et al., 1998). High levels of allelic diversity were present within the original core collection materials (Hokanson et al., 2001; Potts et al., 2012). To our knowledge, however, none of the previous research considered the extent to which the diversity in the grafted PGRU apple collection was captured in the original core collection. We used both molecular and phenotypic traits to assess the diversity of the core collection in relation to a larger portion of the grafted collection. Realizing that the original core collection has been used in multiple plantings and research projects, we thought it prudent to identify sets of individuals that could be added to existing projects or locations that wish to enhance the diversity of their research collections (Table 2). Future projects, however, may choose to make use of a smaller core collection that captures the diversity of the diploid accessions in the main collection with 100 individuals. The identity of the members of the new core sets is provided (Supplemental Table 3).
The phenotypic data available for the MSTRAT analyses are limited by the fact that they are collected from a single individual in a single season. The data are primarily categorical and have been broadly classified. These data were used as additional information to aid in the selection of core collections. Because the categories were broadly defined, it is likely that many of the traits would be classified similarly in subsequent years.
A germplasm collection can potentially be represented by multiple core collections. These core collections may be identified based on geographical, molecular, or phenotypic diversity. Given the diverse user community that is served by the PGRU apple collection, it is reasonable to have core collections that are tailored to specific portions of the collection. For example, previously proposed core sets for the close wild relatives of apple, M. sieversii and M. orientalis, are available for in-depth analyses of the diversity within these species. M. sieversii is represented by three core collections selected from seedlings obtained from unique habitats in Kazakhstan.
The core collections herein are designed to be representative of the large, multispecies, grafted collection of trees at the PGRU. As a result of the complications resulting from obtaining reliable SSR data for polyploid individuals, only diploid individuals were used in our analyses. This reduced the number of individuals in our data sets from 258 in the original core to 198 and from 3442 in the grafted collection (as of 2011) to 2114. When just these diploid accessions are considered, the original core collection captures ≈83% of the measured diversity in the grafted collection in ≈10% of the individuals. The newly proposed cores of 100 individuals capture 100% of the measured diversity in ≈5% of the individuals (Table 3). Additional polyploid individuals could be selected to complement the core collections that have been proposed for the diploid accessions.
MSTRAT has been used extensively to identify core sets of individuals that capture the diversity of plant collections (Balfourier et al., 2007; Ellwood et al., 2006; Jansen and van Hintum, 2007; McKhann et al., 2004). MSTRAT has the advantage of allowing the inclusion of both phenotypic and genotypic categorical data sets that are very large. MSTRAT also lets the user “lock-in” certain individuals that already exist in core sets and identify complimentary individuals that capture more diversity, thus producing an amended core (Gouesnard et al., 2001). However, the current program includes missing data as a character state and considers the two alleles present at a particular diploid locus independently rather than considering them jointly (i.e., with MSTRAT, an allele would need to occur at both positions in a locus rather than only occurring once in the data set). The program M+ does not currently have the “lock-in” feature nor does it accept phenotypic data but does overcome the missing data and independent loci drawbacks of MSTRAT (Reeves et al., 2012). Both programs were used in our analyses to propose improved core sets based on the available data.
Conclusion
Large field collections are crucial to maintaining diversity in clonal and perennial crops, but these same collections can be unwieldy for researchers to evaluate or use as a result of their size. Core collections can streamline research that seeks to evaluate genetic or phenotypic diversity and provide a gateway into the collection for plant breeders screening for traits of interest. Although core collections can potentially be based on any number of different characters, it is reasonable to strive for full representation of the documented genetic and phenotypic diversity found in the main collection. We have proposed both a new 100 individual core and an additional 67 individuals that can be used to bring the current core up to 100% representation of the grafted collection. These collections can be supplemented with representatives having traits that were not measured in this study such as differing ploidy levels and geographic origins.
Literature Cited
Balakrishnan, R., Nair, N.V. & Sreenivasan, T.V. 2000 A method for establishing a core collection of Saccharum officinarum L. germplasm based on quantitative-morphological data Genet. Resources Crop Evol. 47 1 9
Balfourier, F., Roussel, V., Strelchenko, P., Exbrayat-Vinson, F., Sourdille, P., Boutet, G., Koenig, J., Ravel, C., Mitrofanova, O., Beckert, M. & Charmet, G. 2007 A worldwide bread wheat core collection arrayed in a 384-well plate Theor. Appl. Genet. 114 1265 1275
Bataillon, T.M., David, J.L. & Schoen, D.J. 1996 Neutral genetic markers and conservation genetics: Simulated germplasm collections Genetics 144 409 417
Bisht, I.S., Mahajan, R.K., Loknathan, T.R. & Agrawal, R.C. 1998 Diversity in Indian sesame collection and stratification of germplasm accessions in different diversity groups Genet. Resources Crop Evol. 45 325 335
Brown, A.H.D. 1989 The case for core collections, p. 136–156. In: Brown, A.H.D., O.H. Frankel, D.R. Marshall, and J.T. Williams (eds.). The use of plant genetic resources. Cambridge Univ. Press, New York, NY
Chavarriaga-Aguirre, P., Maya, M.M., Tohme, J., Duque, M.C., Iglesias, C., Bonierbale, M.W., Kresovich, S. & Kochert, G. 1999 Using microsatellites, isozymes and AFLPs to evaluate genetic diversity and redundancy in the cassava core collection and to assess the usefulness of DNA-based markers to maintain germplasm collections Mol. Breed. 5 263 273
Ellwood, S.R., D’Souza, N.K., Kamphuis, L.G., Burgess, T.I., Nair, R.M. & Oliver, R.P. 2006 SSR analysis of the Medicago truncatula SARDI core collection reveals substantial diversity and unusual genotype dispersal throughout the Mediterranean basin Theor. Appl. Genet. 112 977 983
Franco, J., Crossa, J., Ribaut, J.M., Betran, J., Warburton, M.L. & Khairallah, M. 2001 A method for combining molecular markers and phenotypic attributes for classifying plant genotypes Theor. Appl. Genet. 103 944 952
Franco, J., Crossa, J., Warburton, M.L. & Taba, S. 2006 Sampling strategies for conserving maize diversity when forming core subsets using genetic markers Crop Sci. 46 854 864
Gouesnard, B., Bataillon, T.M., Decoux, G., Rozale, C., Schoen, D.J. & David, J.L. 2001 MSTRAT: An algorithm for building germplasm core collections by maximizing allelic or phenotypic richness J. Hered. 92 93 94
Gross, B.L., Henk, A.D., Forsline, P.L., Richards, C.M. & Volk, G.M. 2012a Identification of interspecific hybrids among domesticated apple and its wild relatives Tree Genet. Genomes 8 1223 1235
Gross, B.L., Volk, G.M., Richards, C.M., Forsline, P.L., Fazio, G. & Chao, C.T. 2012b Identification of ‘duplicate’ accessions within the USDA-ARS National Plant Germplasm System Malus collection J. Amer. Soc. Hort. Sci. 137 333 342
Hokanson, S.C., Lamboy, W.F., Szewc-McFadden, A.K. & McFerson, J.R. 2001 Microsatellite (SSR) variation in a collection of Malus (apple) species and hybrids Euphytica 118 281 294
Hokanson, S.C., Szewc-McFadden, A.K., Lamboy, W.F. & McFerson, J.R. 1998 Microsatellite (SSR) markers reveal genetic identities, genetic diversity and relationships in a Malus × domestica Borkh. core subset collection Theor. Appl. Genet. 97 671 683
Jansen, J. & van Hintum, T. 2007 Genetic distance sampling: A novel sampling method for obtaining core collections using genetic distances with an application to cultivated lettuce Theor. Appl. Genet. 114 421 428
Lázaro, A. & Aguinagalde, I. 2006 Genetic variation among Spanish pea landraces revealed by inter simple sequence repeat (ISSR) markers: Its application to establish a core collection J. Agr. Sci. 144 53 61
Liebhard, R., Gianfranceschi, L., Koller, B., Ryder, C.D., Tarchini, R., Van de Weg, E. & Gessler, C. 2002 Development and characterization of 140 new microsatellites in apple (Malus × domestica Borkh.) Mol. Breed. 10 217 241
McKhann, H.I., Camilleri, C., Bérard, A., Bataillon, T., David, J.L., Reboud, X., Le Corre, V., Caloustian, C., Gut, I.G. & Brunel, D. 2004 Nested core collections maximizing genetic diversity in Arabidopsis thaliana Plant J. 38 193 202
Meirmans, P.G. & Van Tienderen, P.H. 2004 GENOTYPE and GENODIVE: Two programs for the analysis of genetic diversity of asexual organisms Mol. Ecol. Notes 4 792 794
Potts, S.M., Han, Y., Khan, M.A., Kushad, M.M., Rayburn, A.L. & Korban, S.S. 2012 Genetic diversity and characterization of a core collection of Malus germplasm using simple sequence repeats (SSRs) Plant Mol. Biol. Rpt. 30 827 837
Reeves, P.A., Panella, L.W. & Richards, C.M. 2012 Retention of agronomically important variation in germplasm core collections: Implications for allele mining Theor. Appl. Genet. 124 1155 1171
Richards, C.M., Volk, G.M., Reeves, P.A., Reilley, A.A., Henk, A.D., Forsline, P.L. & Aldwinckle, H.S. 2009a Selection of stratified core sets representing wild apple (Malus sieversii) J. Amer. Soc. Hort. Sci. 134 229 235
Richards, C.M., Volk, G.M., Reilley, A.A., Henk, A.D., Lockwood, D.R., Reeves, P.A. & Forsline, P.L. 2009b Genetic diversity and population structure in Malus sieversii, a wild progenitor species of domesticated apple Tree Genet. Genomes 5 339 347
Ronfort, J., Bataillon, T., Santoni, S., Delalande, M., David, J.L. & Prosperi, J.-M. 2006 Microsatellite diversity and broad scale geographic structure in a model legume: Building a set of nested core collection for studying naturally occurring variation in Medicago truncatula BMC Plant Biol. 6 28
Rubenstein, K.D., Smale, M. & Widrlechner, M.P. 2006 Demand for genetic resources and the U.S. National Plant Germplasm System Crop Sci. 46 1021 1031
Schoen, D.J. & Brown, A.H.D. 1995 Maximising genetic diversity in core collections of wild relatives of crop species, p. 55–76. In: Hodgkin, T., A.H.D. Brown, T.J.L van Hintum, and E.A.V. Morales (eds.). Core collections of plant genetic resources. Wiley-Sayce, Rome, Italy
U.S. Department of Agriculture 2012 National genetic resources program. Germplasm resources information network (GRIN). 3 Dec. 2012. <http://www.ars-grin.gov/cgi-bin/npgs/html/>
Volk, G.M., Richards, C.M., Henk, A.D., Reilley, A.A., Reeves, P.A., Forsline, P.L. & Aldwinckle, H.S. 2009 Capturing the diversity of wild Malus orientalis from Georgia, Armenia, Russia, and Turkey J. Amer. Soc. Hort. Sci. 134 453 459
Volk, G.M., Richards, C.M., Reilley, A.A., Henk, A.D., Forsline, P.L. & Aldwinckle, H.S. 2005 Ex situ conservation of vegetatively propagated species: Development of a seed-based core collection for Malus sieversii J. Amer. Soc. Hort. Sci. 130 203 210
Wang, J.-C., Hu, J., Liu, N.-N., Xu, H.-M. & Zhang, S. 2006a Investigation of combining plant genotypic values and molecular marker information for constructing core subsets J. Integr. Plant Biol. 48 1371 1378
Wang, J.C., Hu, J., Xu, H.M. & Zhang, S. 2007 A strategy on constructing core collections by least distance stepwise sampling Theor. Appl. Genet. 115 1 8
Wang, L., Guan, Y., Guan, R., Li, Y., Ma, Y., Dong, Z., Liu, X., Zhang, H., Zhang, Y., Liu, Z., Chang, R., Xu, H., Li, L., Lin, F., Luan, W., Yan, Z., Ning, X., Zhu, L., Cui, Y., Piao, R., Liu, Y., Chen, P. & Qiu, L. 2006b Establishment of Chinese soybean (Glycine max) core collections with agronomic traits and SSR markers Euphytica 151 215 223
Young, W.J. 1914 A study of variation in the apple Amer. Nat. 48 595 634
Yue, C., Gallardo, R.K., Luby, J., McFerson, J., Liu, L. & Iezzoni, A. 2012 Technical and socio-economic challenges to setting and implementing priorities in North American rosaceous fruit breeding programs HortScience 47 1320 1327
Apple accessions in the original Malus core collection, which was designed to be representative of the Plant Genetic Resources Unit grafted apple collection in the 1990s.z
Ten independently identified 265-sample amended cores generated by the MSTRAT program (Gouesnard et al., 2001).z
Ten independently identified 100-sample core collections were generated by the M+ program (Reeves et al., 2012) based on genotypic data.z
