Abstract
The increasing importance of pumpkin (Cucurbita pepo L.) cultivars for seed production has led to considerable breeding efforts for novel high-yielding and disease-resistant cultivars lacking seedcoats. Because it is very important to use genetically diverse genotypes for the development of cultivars with a broad genetic and phenotypic base, this study focused on phenotypic and genetic diversity within and among available pumpkin accessions with mutated seedcoat phenotypes. Fifty-one accessions were collected from various sources and countries, which showed a wide variety of seedcoat types. Genetic analysis with 18 simple sequence repeat (SSR) markers revealed that 37.59% of the total genetic diversity was attributable to interpopulation differentiation and 62.41% to individual differentiation within populations. The average genetic differentiation between accessions (FST) was from 0.030 to 0.760, whereas expected heterozygosity (He) was between 0.048 and 0.491 and observed heterozygosity (Ho) between 0.056 and 0.522. Based on unweighted pair group method with arithmetic mean (UPGMA) analysis, the genetic relationship among accessions reflects the primary geographical origin of accessions. Marker amplification yielded a total of 109 alleles with an average number of alleles per locus of 6.06. Gene diversity per locus varied between 0.027 and 0.879, whereas the polymorphism information content (PIC) varied between 0.027 and 0.867. This is the first report about intra-accession phenotypic and genotypic variability of pumpkins with mutated seedcoats cultivated for their seeds, which are today used in the baking industry, seed oil production, and in traditional and modern medicine.
The genus Cucurbita (Cucurbitaceae) includes a vast variety of fruits, including gourds, squashes, and pumpkins. One of the economically most important species of the genus is Cucurbita pepo L., which produces plants with a wide range of growth habits, fruit sizes, shapes, and colors. The edible round-fruited types of C. pepo are called “pumpkins,” the edible non-round types are called “squash,” whereas the non-edible sorts are called “gourds” (Paris et al., 2003). Based on fruit shape, the species is divided into eight edible-fruited cultivar groups and two cultivar groups of ornamental gourds: Pumpkin (round), Cocozelle (long, bulbous cylindrical), Vegetable Marrow (short, tapered cylindrical), Zucchini (uniformly cylindrical), Orange Gourd (small, round), Acorn (turbinate, furrowed), Scallop (flat, scalloped), Crookneck (long, narrow neck), Straightneck (short, thick neck), and Ovifera Gourd (small, various shapes) (Paris, 1986, 2000).
At the end of the 19th century, a mutation in the gene (genes) responsible for pumpkin seedcoat development occurred in Styria, and phenotypes with collapsed testa layers segregated from normal field pumpkins (Teppner, 2000). The absence of four lignified seedcoat layers revealed the dark green color of the innermost layer—the chlorenchyma—composed of cells with high levels of protochlorophyll pigment (Kreft et al., 2009). Since then, these genotypes have been cultivated and bred for the production of pumpkin seeds that are used as snacks and in the baking industry or further processed to extract the oil. The oil from roasted pumpkin seeds is particularly valued in central European cuisine as a result of its strong nut-like taste and dark green color. The seeds are also used in traditional medicine and are present in several medicinal products because of their demonstrated therapeutic properties against benign prostatic hyperplasia and other diseases (Gossell-Williams et al., 2006; Shirvan et al., 2014; Yadav et al., 2010). Despite the increasing economic significance of pumpkin cultivars for seed production, there is no information available about population structure within and among accessions with mutated seedcoats and the range of inbreeding present in these accessions. Previous genetic studies have been mainly focused on interspecific diversity within the genus Cucurbita or genetic diversity among C. pepo subspecies with an emphasis on other cultivar groups (Ferriol et al., 2003; Formisano et al., 2012; Gong et al., 2008a, 2012, 2013; Mady et al., 2013; Paris et al., 2003).
SSR (microsatellites) are very popular molecular markers because they are highly informative, multiallelic, polymerase chain reaction (PCR)-based, and codominant. In this study, SSRs were used to evaluate genetic variability within and among 51 C. pepo accessions with wild-type or mutated seedcoats, the latter commonly referred to as hull-less seeds, thin-coated seeds, or naked seeds. These accessions are valuable resources for breeding purposes and the obtained genetic information will complement phenotypic characterization during selection.
Material and Methods
Fifty-one accessions originating from five genebanks (U.S. Department of Agriculture, Agricultural Research Service, North Central Regional Plant Introduction Station, Iowa State University, Regional Plant Introduction Station, Ames, IA; Institute of Special Crops, Agricultural Research Center Styria, Austria; Arche Noah Association, Austria; Crop Research Institute, Czech Republic; CRA Consiglio per la ricerca e la sperimentazione in agricoltura, Italy), three seed companies (Semenarna Ljubljana, Slovenia; Saatzucht Gleisdorf, Austria; H.S.C. New Zealand), a Slovenian plant breeder (Prof. Dr. Anton Ivančič) and from a commercial package of snack seeds imported from China were included in the study (Table 1).
Results obtained for 51 C. pepo accessions collected and analyzed for seedcoat characteristics and genetic variability.
The type of seedcoat was determined for each accession based on the classification of Murovec et al. (2012). Five seeds per accession were germinated and total genomic DNA was extracted from the leaf tissue of individual plants by a modified CTAB method (Kump and Javornik, 1996). PCR was performed as described by Gong et al. (2008a) and the amplified products were separated by capillary electrophoresis using an ABI PRISM® 3100 Genetic Analyzer (Thermo Fisher Scientific, Waltham, MA). Electropherograms were analyzed with GeneMapper software Version 3.5 (Applied Biosystems, Foster City, CA) and GeneScanTM 600 LIZ® Size Standard was used as an internal standard. Results were recorded as allele lengths and converted for further analyses to appropriate matrixes with the software CONVERT Version 1.31 (Glaubitz, 2004).
Genetic diversity among and within the 51 accessions of C. pepo was determined through allele analysis of 18 SSR markers (Table 2). The total number of amplified alleles, the effective number of alleles, the He and Ho, and PIC for each SSR marker were determined with PowerMarker Version 3.25 software (Liu and Muse, 2005) and POPGENE Version 1.31 (Yeh et al., 1999). For each accession, the number of polymorphic loci, number of alleles per accession and locus, effective number of alleles, mean observed and expected heterozygosity, the fixation index (F) and the FST between accessions were calculated by GenAlEx 6.5 (Peakall and Smouse, 2006, 2012). The analysis of molecular variance (AMOVA) was performed using Arlequin 3.5 (Excoffier et al., 2005). The pairwise genetic differences among accessions were estimated according to the Nei (1973) coefficient as implemented in PowerMarker Version 3.25. The resulting distance matrix was used for hierarchical clustering using the UPGMA analysis algorithm as implemented in PowerMarker Version 3.25 and visualized by MEGA5 software (Tamura et al., 2011).
Amplification results for 18 simple sequence repeat (SSR) markers on 51 accessions of C. pepo.
Results
Seed types.
Seeds were collected from several sources to have a comprehensive collection of seeds with mutated seedcoats. The collected seeds originated from 14 different countries with the highest number of accessions from Austria followed by the United States and Turkey (Table 1). Visual evaluation based on seedcoat characteristics revealed that a wide variety of seed types was present in our collection. They were categorized into 12 different seed types (Fig. 1), as described in Murovec et al. (2012), and labeled from A to L, where the letters indicate: A = wild type, B = central sclerenchymal, C = marginal sclerenchymal 1, D = marginal sclerenchymal 2, E = distorted sclerenchymal, F = aerenchymal, G = undulated, H = collapsed 1, I = collapsed 2, J = thin-coated, K = partial hull-less, and L = white hull-less. The results of the seed types observed in each accession are presented in Table 1. All the different seed types were further analyzed with scanning electron microscopy and the results presented in Murovec et al. (2012).
Twelve different seed types observed in a collection of 51 C. pepo accessions as described: (A) wild type, (B) central sclerenchymal, (C) marginal sclerenchymal 1, (D) marginal sclerenchymal 2, (E) distorted sclerenchymal, (F) aerenchymal, (G) undulated, (H) collapsed 1, (I) collapsed 2, (J) thin-coated, (K) partial hull-less, and (L) white hull-less.
Citation: HortScience 50, 2; 10.21273/HORTSCI.50.2.211
Genetic diversity based on microsatellite data.
Amplification of 18 microsatellite loci from the 51 accessions listed in Table 1 yielded a total of 109 alleles, which ranged in size from 108 bp (markers CMTm115 and CMTm131) to 264 bp (marker CMTp177; Table 2). The number of alleles per locus varied between two (marker CMTm115) and 16 (marker CMTp79) with an average value of 6.06. Ho of markers ranged from 0.004 (marker CMTm239) to 0.888 (marker CMTp235), whereas He ranged from 0.028 (marker CMTm239) to 0.879 (marker CMTp79). PIC of markers varied between 0.027 (marker CMTm239) and 0.867 (marker CMTp79) (Table 2).
The intra-accession genetic variability was estimated by the number of polymorphic loci, number of alleles per accession and locus (Na), effective number of alleles (Ne), and average heterozygosity (Ho, He) per accession. The results showed large differences between accessions (Table 1) with the lowest intra-accession variability observed in accessions “12-ARCHE-KU070-Tschermak Kuerbis” (three polymorphic loci, 21 alleles per accession, Na 1.167, Ne 1.080, Ho 0.078, He 0.048) and “32-PI 364240-Gribovskaja 14” (three polymorphic loci, 21 alleles per accession, Na 1.167, Ne 1.093, Ho 0.056, He 0.059), whereas the highest intra-accession variability was observed in accessions “46-PI 615104-Prostate” (16 polymorphic loci, 50 alleles per accession, Na 2.778, Ne 2.318, Ho 0.300, He 0.491) and “44-PI 490278-Butterball” (16 polymorphic loci, 44 alleles per accession, Na 2.444, Ne 1.961, Ho 0.511, He 0.422). The observed heterozygosities were not significantly different from expected heterozygosities (P = 0.0545) but in 10 accessions (of 51), they were lower than expected. The fixation index F ranged from –0.886 (“29-Beppo HSC151”) to 0.368 (“46-PI 615104-Prostate”) (Table 1). Average values calculated for each type of accession were 0.604, –0.068, and –0.146 for hybrids, cultivars, and landraces, respectively. Statistically significant differences were observed between the average fixation index of hybrids and the fixation indexes of cultivars and landraces (P = 0.0004). The difference in average fixation indices of cultivars and landraces was not statistically significant.
Allele frequencies varied among studied accessions with pairwise comparisons of population genetic differentiation (FST) between 0.030 (accessions “07-Gleisdorfer Opal” and “08-Gleisdorfer Diamant”) and 0.760 (accessions “32-PI 364240-Gribovskaja 14” and “48-PI 267664-Yellow Long”). AMOVA revealed that 37.59% of the total molecular variation was attributable to among population differentiation and 62.41% to individual differentiation within populations. UPGMA analysis clustered the analyzed accessions into several subclusters, as displayed in Figure 2, which shows high genetic differentiation between the North American and European accessions. For most European accessions, the primary country of origin is reflected in tree clusters because accessions originating from the same geographic origin clustered together like, for example, accessions from Turkey and the Republic of Macedonia, accessions from Poland, and accessions from Central Europe countries. It was even more pronounced for accessions originating from breeding programs such as three accessions from a Slovenian plant breeder Prof. Dr. Anton Ivančič (accessions “34-Rumena oljna buca,” “35-Rumena buca,” “36-Siva golica”) and hybrids from Saatzucht Gleisdorf, Austria (accessions “07-Gleisdorfer Opal” and “08-Gleisdorfer Diamant”).
Dendrogram of genetic relationships among 51 C. pepo accessions based on Nei’s coefficient (Nei, 1973) and unweighted pair group method with arithmetic mean (UPGMA) cluster analysis. Origins of accessions: (AUT) Austria, (CHI) People's Republic of China, (CZE) Czech Republic, (Geo, SO) Georgia, South Ossetia, (GER) Germany, (HUN) Hungary, (ITA) Italy, (MAC) Republic of Macedonia, (NZ) New Zealand, (POL) Poland, (RUS) Russia, (SLO) Slovenia, (TUR) Turkey, (UN) Unknown, (USA) United States of America. Seed types: (A) wild type, (B) central sclerenchymal, (C) marginal sclerenchymal 1, (D) marginal sclerenchymal 2, (E) distorted sclerenchymal, (F) aerenchymal, (G) undulated, (H) collapsed 1, (I) collapsed 2, (J) thin-coated, (K) partial hull-less, and (L) white hull-less.
Citation: HortScience 50, 2; 10.21273/HORTSCI.50.2.211
Discussion
Seed characterization revealed that a wide variety of different seed types was present in our collection. Although most literature data describe only the most common three types of C. pepo seeds, the presented results show that at least 12 different seed types are present (Fig. 1). Ten of them show mutant seedcoat phenotypes with the outer four testa layers collapsed in different ways revealing the innermost green chlorenchyma layer, as described in more detail in Murovec et al. (2012). One of the mutant phenotypes even lacks all the five seedcoat layers present in wild-type seeds (epidermis, hypodermis, sclerenchyma, aerenchyma, and chlorenchyma), thus exposing the cotyledons, and this type of seed is the only one that should be called hull-less or naked seed. More than half of the analyzed accessions (27 accessions) contained more than one seed type, probably as a result of pooling of seeds from different plants and as a result of segregation of the trait. In accessions “24-PI 379308-Furazna,” “26-PI 379310-Zeljanka,” “37-PI 164997-Tergomlek,” and “38-PI 420328-Turkey #1,” only wild-type (Fig. 1A) seeds were present, although genebank descriptions indicated that the accessions should contain mutant seeds. For example, in accession “26-PI 379310-Zeljanka,” 60% of seeds should be naked seeds, whereas seeds accession “37-PI 164997-Tergomlek” should be nearly hull-less. Packages of other accessions contained mixtures of wild and mutant seed types or only mutant seed types as shown in Table 1 for each accession.
Microsatellite markers are valuable tools for genetic diversity studies and are today the most frequently used molecular markers as a result of their high abundance and random distribution in eukaryotic genomes, high polymorphism, and codominant inheritance. In the genus Cucurbita, they have been already used for construction of genetic linkage maps of C. pepo (Gong et al., 2008a; Zraidi et al., 2007) and C. moschata (Gong et al., 2008b) for evaluation of genetic diversity in Spanish landraces of C. pepo (Formisano et al., 2012) and for evaluation of genetic relationships and evolution of the genus (Gong et al., 2012, 2013; Paris et al., 2003).
In this study, a relatively high average number of alleles per locus (6.06) was detected compared with the previously reported value of 4.78 for the same loci (Gong et al., 2008a) (Table 2). This is probably the result of the larger number of accessions included in this study (51 accessions, each represented by five individuals) compared with eight genotypes of C. pepo and two related species in the study of Gong et al. (2008a). Lower average numbers of alleles per locus were also recently reported by Formisano et al. (2012) (3.8 alleles per locus), Gong et al. (2012) (3.0 alleles per locus), and Gong et al. (2013) (4.3 alleles per locus), although they analyzed a larger number of accessions belonging to both C. pepo subspecies and all edible cultivar groups in combination with eight other Cucurbita species. It could be explained by the fact that only the most polymorphic loci were used in this study, whereas other authors used a higher number of markers, some of which are obviously less polymorphic. Among amplified loci, marker CMTp79 showed the highest ability to discriminate among accessions as a result of the highest number of different alleles amplified (16), highest PIC (0.867), and highest gene diversity (0.879), whereas marker CMTm239 showed the lowest discrimination power with values of 3, 0.027, and 0.028 for the number of different alleles, PIC, and gene diversity, respectively.
A highly negative average value of the fixation index (–0.604) was calculated for hybrids, which indicates an excess of heterozygosity. It is in accordance with the mode of hybrid seed production through pollination of genetically diverse homozygous lines to obtain high heterozygosity and, consequently, high hybrid vigor. The average fixation index for cultivars was very close to zero (–0.068), indicating random mating of large populations, whereas the average fixation index for landraces was negative (–0.146).
As expected, the lowest genetic diversity (pairwise FST 0.030) was observed between two hybrid varieties (accessions “07-Gleisdorfer Opal” and “08-Gleisdorfer Diamant”), both developed in the same breeding company and also very similar in terms of their plant morphological characteristics. The results were confirmed by UPGMA analysis based on Nei’s genetic distance. Interestingly, the two hybrid varieties did not cluster together with “01-Gleisdorfer Ölkürbis” (Fig. 2), which is an old Austrian population cultivar also developed in Saatzucht Gleisdorf (Gleisdorf Seed Breeding Institute). The high genetic distance might also explain the differences in oil characteristics that are detected by some pumpkin seed oil producers (personal communications), which persuades some growers still to sow “01-Gleisdorfer Ölkürbis” despite the otherwise outperformance of hybrids. Based on genetic analysis, the cultivar 01-Gleisdorfer Ölkürbis is genetically closer to the Slovenian population cultivar 33-Slovenska golica (pairwise FST 0.043), also sown in Austria and Slovenia for pumpkin seed oil production and the Austrian “03-WIES-SK9-Ölkürbis ‘Retzer Gold’” (pairwise FST 0.054). A close relationship was also confirmed for the three breeding lines of Prof. Dr. Anton Ivančič (“34-Rumena oljna buca,” “35-Rumena buca,” “36-Siva golica”), which clustered together with an average FST of 0.137.
The relationships among accessions are presented in the UPGMA tree (Fig. 2), which reflects the relationship between the primary geographical origins of the accessions and their genetic distance. At the top of the tree (from “17-PI 209783-Giessener Ölkürbis” to “19-PI 531323-Szentesi Oliva”), European accessions mostly from Germany, Austria, Slovenia, Hungary, Italy, and Poland are grouped into several clusters. Most of these accessions contained seed Types I (collapsed 2), J (thin-coated), and L (white hull-less), which are most often used for pumpkin seed production, because the lack of the outermost four lignified testa layers facilitates the extraction of pumpkin seed oil. Moreover, these types of seed are also most valuable for pumpkin seed snack production and in the baking industry because the residues of lignified layers (like in mutant seed types B = central sclerenchymal, C = marginal sclerenchymal 1, D = marginal sclerenchymal 2, E = distorted sclerenchymal, F = aerenchymal, G = undulated, and H = collapsed 1; Fig. 1) are unwanted. The only Chinese accession “14-Snack seeds” and the only New Zealand accession, “29-Beppo HSC151,” grouped with European accessions. Some genetic relatedness was expected still to exist because the mutation in gene (genes) for the lignification of the seedcoat occurred in central Europe (Styria). Accessions from Turkey and the Republic of Macedonia clustered into a well-defined cluster (from “41-PI 420331-Turkey #4” to “37-PI 164997-Tergomlek”) of 11 accessions with most seeds of Types A (wild=type) and F (aerenchymal). At the bottom of the tree, below a small cluster composed of a Russian and a Georgian–South Ossetia accession, are all the eight North American accessions with several seedcoat types.
An evaluation of a comprehensive collection of C. pepo accessions with mutated seedcoats is presented in this study. The results of morphological characterization of seeds showed that a wide variety of different seed types exists among the studied accessions (Fig. 1; Table 1). These data, in combination with the results of genetic analysis, will be useful for future breeding programs. As previously reported (Gong et al., 2012; Košmrlj et al., 2013), there is high genetic similarity within the gene pool of hull-less pumpkins cultivated today and the discovery of genetically more distant accessions could overcome this drawback.
Literature Cited
Excoffier, L., Laval, G. & Schneider, S. 2005 Arlequin (version 3.0): An integrated software package for population genetics data analysis Evol. Bioinform. 1 47 50
Ferriol, M., Pico, B. & Nuez, F. 2003 Genetic diversity of a germplasm collection of Cucurbita pepo using SRAP and AFLP markers Theor. Appl. Genet. 107 271 282
Formisano, G., Roig, C., Esteras, C., Ercolano, M.R., Nuez, F., Monforte, A.J. & Pico, M.B. 2012 Genetic diversity of Spanish Cucurbita pepo landraces: An unexploited resource for summer squash breeding Genet. Resources Crop Evol. 59 1169 1184
Glaubitz, J.C. 2004 CONVERT: A user-friendly program to reformat diploid genotypic data for commonly used population genetic software packages Mol. Ecol. Notes 4 309 310
Gong, L., Paris, H.S., Nee, M.H., Stift, G., Pachner, M., Vollmann, J. & Lelley, T. 2012 Genetic relationships and evolution in Cucurbita pepo (pumpkin, squash, gourd) as revealed by simple sequence repeat polymorphisms Theor. Appl. Genet. 124 875 891
Gong, L., Paris, H.S., Stift, G., Pachner, M., Vollmann, J. & Lelley, T. 2013 Genetic relationships and evolution in Cucurbita as viewed with simple sequence repeat polymorphisms: The centrality of C. okeechobeensis Genet. Resources Crop Evol. 60 1531 1546
Gong, L., Stift, G., Kofler, R., Pachner, M. & Lelley, T. 2008a Microsatellites for the genus Cucurbita and an SSR-based genetic linkage map of Cucurbita pepo L Theor. Appl. Genet. 117 37 48
Gong, L., Pachner, M., Kalai, K. & Lelley, T. 2008b SSR-based genetic linkage map of Cucurbita moschata and its synteny with Cucurbita pepo Genome 51 878 887
Gossell-Williams, M., Davis, A. & O'Connor, N. 2006 Inhibition of testosterone-induced hyperplasia of the prostate of Sprague-Dawley rats by pumpkin seed oil J. Med. Food 9 284 286
Košmrlj, K., Murovec, J. & Bohanec, B. 2013 Haploid induction in hull-less seed pumpkin through parthenogenesis induced by X-ray-irradiated pollen J. Amer. Soc. Hort. Sci. 138 310 316
Kreft, M., Zorec, R., Janes, D. & Kreft, S. 2009 Histolocalisation of the oil and pigments in the pumpkin seed Ann. Appl. Biol. 154 413 418
Kump, B. & Javornik, B. 1996 Evaluation of genetic variability among common buckwheat (Fagopyrum esculentum Moench) populations by RAPD markers Plant Sci. 114 149 158
Liu, K.J. & Muse, S.V. 2005 PowerMarker: An integrated analysis environment for genetic marker analysis Bioinforma. 21 2128 2129
Mady, E.A., Helaly, A.A.D., Abu El-Hamd, A.N., Abdou, A., Shanan, S.A. & Craker, L.E. 2013 Genetic diversity assessment of summer squash landraces using molecular markers Mol. Biol. Rep. 40 4269 4274
Murovec, J., Draslar, K. & Bohanec, B. 2012 Detailed analysis of Cucurbita pepo seed coat types and structures with scanning electron microscopy Botany-Botanique 90 1161 1169
Nei, M. 1973 The theory and estimation of genetic distance, p. 45–54. In: Morton, N.E. (ed.). Genetic structure of populations. University Press of Hawaii, Honolulu, HI
Paris, H.S. 1986 A proposed subspecific classification for Cucurbita pepo Phytologia 61 133 138
Paris, H.S. 2000 Segregation distortion in Cucurbita pepo, p. 199–202. In: Katzir, N. and H.S. Paris (eds.). Acta Horticulturae, Proceedings of cucurbitaceae 2000, vol. 510. International Society for Horticultural Science, Ma'ale Ha Hamisha, Israel
Paris, H.S., Yonash, N., Portnoy, V., Mozees-Daube, N., Tzuri, G. & Katzir, N. 2003 Assessment of genetic relationships in Cucurbita pepo (Cucurbitaceae) using DNA markers Theor. Appl. Genet. 106 971 978
Peakall, R. & Smouse, P.E. 2006 GenAlEx 6: Genetic analysis in Excel. Population genetic software for teaching and research Mol. Ecol. Notes 6 288 295
Peakall, R. & Smouse, P.E. 2012 GenAlEx 6.5: Genetic analysis in Excel. Population genetic software for teaching and research-an update Bioinforma. 28 2537 2539
Shirvan, M.K., Mahboob, M.R.D., Masuminia, M. & Mohammadi, S. 2014 Pumpkin seed oil (Prostafit) or Prazosin? Which one is better in the treatment of symptomatic benign prostatic hyperplasia J. Pak. Med. Assoc. 64 683 685
Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S. 2011 MEGA5: Molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods Mol. Biol. Evol. 28 2731 2739
Teppner, H. 2000 Cucurbita pepo (Cucurbitaceae)—History, seed coat types, thin coated seeds and their genetics Phyton-Ann. Rei Bot. 40 1 42
Yadav, M., Jain, S., Tomar, R., Prasad, G.B.K.S. & Yadav, H. 2010 Medicinal and biological potential of pumpkin: An updated review Nutr. Res. Rev. 23 184 190
Yeh, F.C., Yang, R.C. & Boyle, T. 1999 POPGENE Version 1.31: Microsoft Windows-based freeware for population genetic analysis. University of Alberta and Center for International Forestry Research, Alberta, Canada
Zraidi, A., Stift, G., Pachner, M., Shojaeiyan, A., Gong, L. & Lelley, T. 2007 A consensus map for Cucurbita pepo Mol. Breed. 20 375 388