G Model ARTICLE IN PRESS HORTI-6088; No. of Pages 11 Scientia Horticulturae xxx (2015) xxx–xxx Contents lists available at ScienceDirect Scientia Horticulturae journal homepage: www.elsevier.com/locate/scihorti Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis Ioannis Ganopoulos a,b,∗,1 , Theodoros Moysiadis a,1 , Aliki Xanthopoulou a,c , Maria Ganopoulou d , Evangelia Avramidou b , Filippos A. Aravanopoulos b , Eleni Tani e , Panagiotis Madesis a , Athanasios Tsaftaris a,c , Konstantinos Kazantzis f,∗∗ a Institute of Applied Biosciences, CERTH, Thermi, Thessaloniki 57001, Greece Forest Genetics & Tree Breeding, Faculty of Forest & Environmental Science, Aristotle University of Thessaloniki, Greece c Department of Genetics and Plant Breeding, School of Agriculture, Aristotle University of Thessaloniki, Thessaloniki 54124, Greece d Mathematics Department, Aristotle University of Thessaloniki, Thessaloniki 54006, Greece e Department of Plant Breeding and Biometry, Agricultural University of Athens, Iera Odos 75, 11855 Athens, Greece f Pomology Institute (Hellenic Agricultural Organisation—‘DEMETER’), P.O. Box 122, Naoussa, 59200, Greece b a r t i c l e i n f o Article history: Received 23 June 2015 Received in revised form 25 August 2015 Accepted 30 September 2015 Available online xxx Keywords: Phenotypic diversity Prunus avium Germplasm collection Hierarchical cluster analysis a b s t r a c t This study presents an assessment of morpho-physiological diversity for one hundred and forty-six sweet cherry (Prunus avium) cultivars, originating from different countries and maintained in an ex situ GeneBank collection. Data of thirty-five traits, describing phenology, plant morphology, yield and fruit quality were recorded over three years and analyzed using principal component analysis (PCA). An unsupervised hierarchical cluster analysis was performed the different cultivars using the Euclidean distance metric and the Ward’s agglomeration method. Significant positive and negative correlations were detected among the different morpho-physiological traits. The sweet cherry cultivars were classified into three main clusters, suggesting that the characterized sweet cherry collection has high potential for specific breeding goals. Correlations among the traits, which will be useful for breeding in fruit size and quality, are discussed. Sweet cherry cultivars, which were classified in diverse clusters, could be potential parents for hybridization and new genotypes could be created with a combination of desirable traits that complement one another. These new genotypes could have a high heterotic behavior and thus could substantially contribute to existent sweet cherry breeding programs. © 2015 Elsevier B.V. All rights reserved. 1. Introduction Cherries (Prunus avium L.) include sweet cherry trees cultivated for human consumption and wild cherry trees, also called mazzards, cultivated for their wood. Cherries are thought to have originated in the Caucasus area, whereas at present they are found across mainland Europe and western Asia (Webster, 1996). Undoubtedly, cherries were an early food source for primitive inhabitants of Europe, as pits have been recovered from cave dwellings that date back to 4000–5000 B.C. Sweet cherries were probably first cultivated in Greece (Hedrick, 1915; Marshall, 1954). Today, cherry cultivation is one of the most popular fruit tree crop ∗ Corresponding author at: Institute of Applied Biosciences, CERTH, Thermi, Thessaloniki 570 01, Greece. ∗∗ Corresponding author. E-mail addresses: giannis.ganopoulos@gmail.com (I. Ganopoulos), nagrefpi@otenet.gr (K. Kazantzis). 1 These two authors contribute equally to this work. in Greece (Koukourojiannis, 1996). Commercial demand for sweet cherry resulted in the contemporary increased agricultural production. Greek growers, in order to follow current market demands, have changed their sweet cherry production from local cultivars, mainly from ‘Van’ and ‘Burlat’ groups (Ganopoulos et al., 2011), to the non-native sweet cherry cultivars. The replacement of local, well adapted to the local climate conditions, by commercial sweet cherry cultivars led to a dramatic loss of genetic diversity in terms of adaptability, tolerance to diseases and fruit quality. Therefore, it is of utmost importance the need for preservation of endangered fruit germplasm and establishment of programs that target the conservation of genetic resources in Greece (Ganopoulos et al., 2011). Keulemans (1993) underlined the contribution of local cultivars to variability and better adaptability. The Experimental Station of Institute of Pomology in Greece, which was established in 1970s, created the first Greek germplasm collection with local and international sweet cherry cultivars (Kazantzis, 2013). http://dx.doi.org/10.1016/j.scienta.2015.09.061 0304-4238/© 2015 Elsevier B.V. All rights reserved. Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model HORTI-6088; No. of Pages 11 2 ARTICLE IN PRESS I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx One of the most important factors influencing plant breeding is the existence of genetic diversity. Therefore, the identification and estimation of the genetic diversity as well as its nature are of paramount importance for a successful breeding program. Furthermore, it is crucial to have not only the information on the genetic diversity but also the information on plant germplasm and the physical presence of the germplasm in order to preserve it, exploit it and sustainable use it in plant breeding and in agriculture or even for other uses (Khadivi-Khub et al., 2014). To the best of our knowledge, this is the first study describing the complete process of morpho-physiological characterization of a worldwide sweet cherry germplasm collection. Furthermore, similar research in fruit crops has been conducted in a worldwide olive GeneBank collection by using morphological markers (Díez et al., 2012; Trujillo et al., 2014). In order to identify and analyze the genetic diversity of the various sweet cherry cultivars used, one can rely solely on the phenotypic traits (IPGRI, 1985; UPOV, 1976). Morphological analysis is a quick and commonly used method to identify and characterize the germplasm through phenotyping. To discover phenotypic traits that mostly contribute to the total diversity in a germplasm collection and characterize the levels of similarity/dissimilarity among the cultivars, a characterization of phenotypic diversity and structure is needed (de Oliveira et al., 2012; Furones-Pérez and Fernández-López, 2009; Mehmood et al., 2014). Multivariate data analysis includes powerful statistical techniques for analyzing data with many variables simultaneously to identify patterns and relationships. Since information obtained from morphological characterization is derived from a large data set consisting of qualitative and quantitative traits, the use of multivariate analysis is particularly preferred (de Oliveira et al., 2012; Furones-Pérez and Fernández-López, 2009; Mehmood et al., 2014). Furthermore, multivariate analysis has been used for genetic variability estimation. The most suitable multivariate techniques for morphological characterization of genotypes are principal component analysis (PCA) and cluster analysis (Mohammadi and Prasanna, 2003; Peeters and Martinelli, 1989). The combination of these statistical methods could provide comprehensive information of characteristics that crucially contribute to genetic diversity in plants (Khodadadi et al., 2011). The aims of this study were (i) to evaluate the phenotypic diversity in 146 worldwide sweet cherry cultivars preserved in a Greek GeneBank collection, (ii) to identify specific traits, and (iii) to detect relationships among the studied cultivars. 2. Materials and methods 2.1. Plant material Morpho-physiological genetic diversity was assessed in 146 worldwide sweet cherry (P. avium L.) cultivars. These cultivars are part of the ex situ Greek Fruit GeneBank collection in Naoussa (Table 1) and represent the total diversity of Greek sweet cherry cultivars and a large part of international cultivars. Thirty-five variables were examined on the basis of the cherry descriptors at the experimental collection for three consecutive years (2010–2012). Different horticultural practices including fertilizer, application spraying and irrigation and other cultural practices were made on regular intervals each year. At the beginning of the study (2010), the trees were mature (8 years old) and also healthy, and in cropping condition. 2.2. Analysis of morpho-physiological traits Mature leaves were collected, approximately at the end of July. From each of the four trees studied per cultivar, seven leaves were sampled per year, and the following parameters were measured using a digital caliper with a sensitivity of ±0.01 mm. The flowers were collected at full bloom; ten flowers were taken from each of the four trees studied per cultivar and year. Fruits were collected at maturity. Cherry fruits maturity was determined based on the color characteristics of each cultivar, taking into account information provided by growers and from personal experience and observation. A sample of a total of 106 cherry fruits was taken from each of the four trees studied per cultivar and year. One hundred of them were used to determine the mean fruit weight. The remaining six cherries were used to study a series of quantitative and qualitative descriptors. The 21 quantitative traits evaluated, included Stone volume [StV], Yield [YL], Ratio volume stone/volume fruit [VSt/VFr], Ratio weight stone/weight fruit [WeSt/WeFr], Petiole length [PL], Petiole width [PeWi], Ratio petiole length/width [PL/Wi], Blade length [BlLe], Blade width [BlWi], Ratio blade length/petiole length [BlLe/PL], Stone length [StLe], Stone width [StWi], Stone thickness [StTh], Soluble solids [SoSo], Titratable acidity [TiAc], ratio titratable acidity/soluble solids [TiAc/ SoSo], Pedicel length [PeLe], Fruit polar diameter [PoDi], Fruit equatorial diameter [EqDi], Fruit width [FrWi] and Fruit weight [FrWe]. The 14 qualitative traits investigated included tree vigour [TrVi], tree habit [TrHa], tree branching [TrBr], lenticels size [LeSi], lamella shape [LaSh], number of nectaries [NuNe], petal shape [PeSh], fruit shape [FrSh], fruit size [FrSi], fruit skin colour [FrSCo], firmness of flesh [FiFl], stone shape [StSh] and stone size [StSi]. These traits were selected from the International Union for the Protection of New Cultivars of Plants descriptors proposed for sweet cherry (UPOV, 1976; IPGRI, 1985). The scoring system is the same as used by UPOV and IPGRI. When possible, all measurements of a trait were made on the same date to avoid differences in the environment or developmental stages of the tree. 2.3. Data scoring and analysis Data for the 146 sweet cherry cultivars, involving 35 traits, were analyzed via XLSTAT software (version 2014.1). The Principal Components Analysis (PCA) was applied separately for the quantitative and qualitative traits. The missing data were estimated by the mean in the first case and by the mode in the latter. In both cases the correlation matrix was used, the main reason being that standardization was necessary since the variables were measured in different units. Within PCA, factor loadings >0.55 were regarded as significant since the number of observations was 146 (see also Mehmood et al., 2014). In each case, 3D plots were constructed, with regard to the three most important principal components, to facilitate the visualization of the results. In the correlation analysis, the Pearson coefficient (parametric) was used to measure the correlation among quantitative characteristics and the Spearman coefficient (non-parametric) was used to measure the correlation among qualitative characteristics. The combined data from both the quantitative and qualitative traits were used for dendrogram construction. The chosen distance to estimate the genetic dissimilarity component was Euclidean and Ward’s method was used for the agglomerative hierarchical clustering (AHC). 3. Results 3.1. Descriptive statistics and correlations for the quantitative variables The 21 quantitative traits were measured and the descriptive statistics of minima, maxima, means, standard deviations and coefficients of variation (CV) are shown in Table 2. The results revealed extensive morphological variability. Some traits displayed high CV. These included YL (43.97%), StV (30.75%), TiAc/SoSo (30.49%), TiAc (28.64%) and VSt/VFr (28.12%). Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model HORTI-6088; No. of Pages 11 ARTICLE IN PRESS 3 I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx Table 1 Origin and the main fruit characteristics sweet cherries cultivars analyzed. Number (No) Cultivar Origin Fruit skin colour Flesh firmness Fruit size 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 2e 48-28 Adriana Angela B. Producta Delbard Bargioni I-137 Bargioni I-37 Bargioni I-38 Bargioni I-62 Bargioni I-63 Bargioni M-47 Belle Magnifique Bianca Di Verona Bigarreau Burlat Bigarreau Burlat S-370 Bigarreau Geant D’ Hedelfingen Bigarreau Goeur De Pigeon Bigarreau Marmotte Bigarreau Moreau Bigarreau Napoleon Bigarreau Reverchon Bigarreau Stark Hardy Giant Bigarreau Tigre Bing Black Russian Black Tartarian Blanka Kukleitska Brooks Burlat e1 Canada Giant Chinook Ciliegio If Roma BB2 Ciliegio If Roma T-57 Compact Stella Corniola Cristalina Cuglyeva Acacra Della Marka Modenese Di Mauria Droganova Zuta Durona Di Cesena Durone Di Vignola Durone Di Vignola II Early Rivers Empereur Francis Fercer (Arcina) Ferrovia Ferrovia spur Germersdorfer Giorgia Glorius Stark Gold Grossa Di Pistoia Grossa Rossa Guillaume Hative De Bale Hative De Berny Hebros Hedelfingen V 18775 × 20 Hudson Jaboulay Jubilee Kordia Kustendilska Hrustjalka Lambert Òim␩ Lapins Larian Merton Bigarreau Monnembegi Napoleon S-787 Negre Di Bistrita Nera Di Piemonde Nero II clone 52 P3 Nero II clone 78 P2 Nero II e1 8 New Star 26-3-7. Canada Italy Canada France Italy Italy Italy Italy Italy Italy France Italy France France Germany France France France Germany Italy USA France USA USA Europe Bulgary USA Italy Canada USA Italy Italy New Ziland Italy Canada Bulgary Italy unknown Bulgary Italy Italy Italy United Kingdom Germany France Italy Italy Hungary Italy USA Italy Italy France Switcherland France Bulgary Germany USA France USA Czech Republic Bulgary USA Canada USA United Kingdom Romania France Romania Romania Italy Italy Italy Canada Dark red Dark red Dark red Orange red Orange red Dark red Dark red Dark red Dark red Dark red Light red Orange red Dark red Dark red Darkish Darkish Dark red Dark red Orange red Dark red Dark red Darkish Darkish Darkish Darkish Dark red Dark red Dark red Dark red Red Brown red Red Brown red Dark red Dark red Dark red Orange red Yellow Yellow Dark red Darkish Darkish Dark red Orange red Dark red Dark red Dark red Light red Dark red Yellow Darkish Orange red Dark red Darkish Darkish Dark red Darkish Light red Dark red Light red Dark red Dark red Darkish Red Brown red Light red Dark red Orange red Brown red Darkish Dark red Dark red Dark red Dark red Strong Strong Weak Medium Medium Strong Strong Strong Strong Strong Strong Strong Medium Medium Medium Strong Strong Strong Strong Strong Strong Strong Strong Medium Weak Strong Medium Strong Strong Strong Medium Strong Strong Strong Strong Strong Weak Weak Strong Strong Medium Strong Weak Strong Strong Strong Strong Strong Strong Weak Strong Strong Strong Weak Weak Strong Medium Strong Strong Medium Strong Medium Medium Strong Strong Strong Strong Strong Strong Weak Strong Strong Strong Strong Very small Large Small Very large Very large Small Medium Small Very large Medium Very large Large Very large Very large Medium Medium Large Medium Small Large Very large Large Very large Medium Small Medium Large Very large Very large Very large Very large Small Small Very small Medium Large Small Small Very large Large Very large Large Very small Very large Very large Very large Very large Very large Very large Small Large Large Large Very small Very small Very large Large Small Small Large Large Small Very large Very large Very large Very small Very small Small Very small Small Very large Medium Large Large Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model HORTI-6088; No. of Pages 11 4 ARTICLE IN PRESS I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx Table 1 (Continued) Number (No) Cultivar Origin Fruit skin colour Flesh firmness Fruit size 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 New York 1143 (NY 1143) New York T-27 Noble Northstar Pobeda Precoce Bernard Precoce Della Marca Primavera Prime Giant (Giant Red) R3 Daritska Beluide Rainier Rana Cherne Edra Regina Rosii Di Bistrita Sam Schmidt’s Bigarreau Seneca Solymári Gömbölyü Staccato (Splendid) Stark Gold Bigarreau Starkrimson Stella Sue Summit Sunburst Ulster V-1927 Valera Valerij Tschkalov Van Vega Verdel Ferbolus Victor Vittoria Vogue Windsor Ziraat Agiorgitika Lilantiou Athinaika Arkadias Basiliadi I.O.P. 1 I.O.P. 2 I.O.P. 3 Karamela Lilantiou Kapsiotika Kifisias Kifisias Proimotero Kokkina Anastasias Koromilokeraso Vitalou Lemonidi Mavra Anastasias Mavro Proimo Achaias Mavro Proimo Vitalou Mavro Tripoleos Mesoproimo Tragano Evias Moschato Tragano Opsimo Evias Bakirtzeika Napoleon Karamela Opsimi Karamela Tripoleos Opsimo Tragano Komotinis Petrokeraso Tragano Achaias Proimo Kolindrou Proimo Tragano Komotinis Samou Tourkika Tragana Edessis Tragana Edessis-Naoussis Tragana Edessis-Sarakinon Tragano Komotinis Fraoula Volou Chalkidos Unknown USA USA United Kingdom Canada Bulgary France Italy Germany USA Bulgary USA Bulgary Germany Romania Canada Europe USA Hungary Canada USA USA Canada Canada Canada Canada USA unknown Canada Hungary Canada Canada France Canada Italy Canada Canada Turkey Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Greece Orange red Orange red Dark red Dark red Dark red Brown red Red Dark red Dark red Orange red Orange red Orange red Dark red Dark red Dark red Darkish Dark red Dark red Red Yellow Dark red Light red Orange red Light red Light red Dark red Dark red Brown red Dark red Dark red Orange red Red Orange red Dark red Dark red Darkish Dark red Dark red Dark red Darkish Dark red Dark red Dark red Dark red Orange red Orange red Dark red Orange red Dark red Dark red Dark red Dark red Dark red Dark red Dark red Dark red Orange red Light red Orange red Orange red Dark red Orange red Dark red Dark red Light red Dark red Dark red Dark red Dark red Dark red Yellow with blush Dark red Strong Strong Strong Medium Medium Medium Weak Strong Strong Weak Strong Medium Strong Weak Medium Medium Weak Strong Strong Weak Medium Medium Medium Medium Medium Strong Strong Medium Strong Strong Strong Strong Weak Strong Strong Medium Medium Weak Medium Strong Medium Strong Strong Strong Strong Weak Weak Medium Strong Strong Medium Strong Strong Medium Medium Strong Weak Strong Medium Medium Strong Strong Weak Strong Strong Strong Strong Strong Strong Medium Medium Strong Large Medium Very large Very large Small Small Small Very small Very large Very small Very large Small Large Large Large Large Very small Very small Medium Small Very large Very large Very large Very large Very large Medium Medium Small Very large Medium Very large Large Large Medium Large Medium Very large Very small Small Medium Very large Very small Very small Very small Small Small Very small Large Medium Small Very large Large Very large Small Medium Very small Very small Very large Small Very small Small Large Very small Very small Medium Very large Large Very large Very large Very small Large Very large Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model ARTICLE IN PRESS HORTI-6088; No. of Pages 11 5 I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx Table 2 Descriptive statistics for 21 quantitative traits in 146 sweet cherry cultivars. Variable Minimum Maximum Mean Std. deviation CV (%) Stone volume Yield Ratio volume stone/volume fruit Ratio weight stone/weight fruit Petiole length Petiole width Ratio petiole length/width Blade length Blade width Ratio blade length/petiole length Stone length Stone width Stone thickness Soluble solids Titratable acidity Ratio titratable acidity/soluble solids Pedicel length Fruit polar diameter Fruit equatorial diameter Fruit width Fruit weight 0.190 25.000 7.700 1.700 9.900 4.540 0.400 2.720 0.110 2.060 8.400 7.000 5.400 9.600 3.000 0.900 2.700 11.500 18.100 15.500 3.100 0.750 190.000 35.300 26.600 16.000 8.080 0.600 6.220 0.220 5.250 18.000 15.700 11.700 22.300 17.400 7.100 7.300 26.600 28.500 24.000 12.100 0.439 71.096 16.931 14.714 12.437 6.058 0.487 4.195 0.158 3.102 11.120 8.936 7.102 16.644 8.539 2.112 4.515 21.387 23.586 19.838 6.896 0.135 31.264 4.762 3.320 1.266 0.583 0.040 0.562 0.019 0.593 1.099 0.896 0.720 2.470 2.446 0.644 0.719 1.947 2.016 1.486 1.502 30.75 43.97 28.12 22.56 10.17 9.62 8.21 13.39 12.02 19.11 9.88 10.02 10.13 14.84 28.64 30.49 15.92 9.11 8.54 7.49 21.78 Strong positive, linear correlations were observed among all the 21 quantitative traits (Table 3). The highest significant, positive correlation was between FrWe and EqDi (0.858). Significant positive correlations were also observed between EqDi and FrWi (0.844), PL and PeWi (0.709), StLe and StWi (0.686), VSt/VFr and WeSt/WeFr (0.646), PeWi and BlWi (0.565), EqDi and WeSt/WeFr (0.541), FrWe and WeSt/WeFr (0.495), and FrWi and WeSt/WeFr (0.458). On the other hand, there were also high, significant, negative correlations between some quantitative traits (Table 3). These included for instance, the TiAc/ SoSo and TiAc (-0.714) and VSt/VFr and StVo (−0.712) (Table 3). We reported significant correlations between traits contributing to fruit yield and quality, which is helpful for plant breeding. For example, to attain high yield and superior quality cultivars, cross combinations could be performed between cultivars with large fruit size (‘Lemonidi’, ‘Mpakirtzeika’, ‘Starkrismson’ and ‘Ziraat’), low stone volume (‘Brooks’, ‘Hative De Berny’, ‘Hative De Bale’, ‘Seneca’, ‘Merton Bigarreau’ and ‘Monnembegi’) and high total soluble solids (‘2e 48-28′ , ‘Monnembegi’, ‘Napoleon S-787′ , ‘R3 Daritska Beluide’, ‘Rainier’, ‘Rosii Di Bistrita’, ‘Sue’, ‘Arkadias’, ‘Kapsiotika’ and ‘Tragana Edessis-Naoussis’). 3.2. Principal component analysis of quantitative variables The distribution of cultivars, based on the PC-1, PC-2 and PC3, shows the phenotypic variation among the cultivars and how widely dispersed they are along both axes (Fig. 1). Using Kaiser’s criterion (“Eigenvalue” >1) (Kaiser, 1958), we reduce the dimension implied by the 21 quantitative traits to six significant components that explained 73.59% of the total variation (Table 4; Fig. 1). The first component, which accounted for 22.90% of the total variation, included stone volume, petiole length, blade width, fruit polar diameter, fruit equatorial diameter, fruit width and fruit weight. The second component, which explained 16.75% of the total variation, was mainly correlated to characters related to ratio of volume stone/volume fruit, ratio weight stone/weight fruit, stone length, stone width and stone thickness. The third component that explained 10.15% of the total variation included petiole length, petiole width and stone length. The fourth component, accounted for 9.55% of the total variation, was determined by titratable acidity and ratio of titratable acidity/soluble solids. The fifth component explained 7.71% of the total variation, and included blade length Fig. 1. ‘Scree plot’ of eigenvalues obtained from the PCA process for (A) quantitative and (B) qualitative traits. Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model HORTI-6088; No. of Pages 11 6 StV YL VSt/VFr StV 1 YL 0.199 1 −0.712 0.008 1 VSt/VFr WeSt/WeFr −0.227 −0.027 0.646 PeLe 0.130 0.115 0.050 PeWi 0.158 0.095 −0.028 PL/Wi 0.059 −0.041 −0.118 BlLe −0.225 −0.244 0.200 BlWi 0.262 0.277 −0.025 0.085 0.282 −0.055 BlLe/ PeLe 0.353 0.091 −0.280 StLe 0.427 0.142 −0.284 StWi 0.352 0.214 −0.243 StTh SoSo 0.024 0.049 −0.013 0.282 −0.034 −0.186 TiAc TiAc/ SoSo −0.271 0.150 0.190 PeLe 0.212 0.106 −0.216 0.206 0.081 0.302 PoDi 0.389 0.070 0.231 EqDi FrWi 0.460 0.245 0.169 FrWe 0.334 0.142 0.330 WeSt/WeFr 1 0.100 0.012 −0.094 0.155 0.065 −0.131 −0.471 −0.434 −0.487 0.066 0.096 −0.048 −0.272 0.457 0.541 0.458 0.495 PeLe PeWi 1 0.709 1 −0.389 0.316 0.167 0.077 0.671 0.565 0.431 0.333 0.068 0.087 0.126 0.219 0.181 0.223 0.007 0.058 0.134 0.150 −0.104 −0.142 −0.027 0.058 0.287 0.195 0.267 0.221 0.310 0.258 0.328 0.257 PL/Wi BlLe BlWi 1 −0.118 −0.171 −0.259 0.063 0.123 0.054 0.060 0.020 −0.052 0.096 -0.114 -0.025 -0.038 -0.070 1 0.141 1 −0.614 0.277 −0.156 −0.005 −0.176 0.182 −0.216 0.228 −0.003 0.000 −0.042 0.123 0.050 −0.123 0.150 0.096 −0.038 0.216 0.010 0.312 −0.050 0.360 0.041 0.330 BlLe/PeLe StLe StWi StTh SoSo TiAc TiAc/SoSo PeLe PoDi EqDi FrWi FrWe 1 0.032 0.094 0.263 0.027 0.088 −0.067 −0.059 0.050 −0.030 0.085 0.020 1 0.686 0.542 −0.123 −0.080 0.049 0.060 0.335 0.118 0.129 0.161 1 0.790 −0.078 −0.002 −0.053 0.091 0.192 0.298 0.320 0.259 1 −0.108 0.016 −0.099 0.126 0.103 0.191 0.313 0.208 1 0.460 0.135 0.214 90.021 −0.005 0.090 0.021 1 −0.714 0.062 0.004 0.137 0.256 0.142 1 0.058 0.034 −0.133 −0.210 −0.109 1 −0.172 −0.104 −0.025 −0.046 1 0.595 0.660 0.701 1 0.844 0.858 1 0.843 1 Values in bold are different from 0 with a significance level alpha = 0.05. Trait abbreviation: Stone volume [StV], Yield [YL], Ratio volume stone/volume fruit [VSt/VFr], Ratio weight stone/weight fruit [WeSt/WeFr], Petiole length [PL], Petiole width [PeWi], Ratio petiole length/width [PL/Wi], Blade length [BlLe], Blade width [BlWi], Ratio blade length/petiole length [BlLe/ PL], Stone length [StLe], Stone width [StWi], Stone thickness [StTh], Soluble solids [SoSo], Titratable acidity [TiAc], ratio titratable acidity/soluble solids [TiAc/SoSo], Pedicel length [PeLe], Fruit polar diameter [PoDi], Fruit equatorial diameter [EqDi], Fruit width [FrWi] and Fruit weight [FrWe]. ARTICLE IN PRESS Variables I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 Table 3 Correlation coefficients (Pearson) among 21 quantitative traits in 146 sweet cherry cultivars. G Model ARTICLE IN PRESS HORTI-6088; No. of Pages 11 7 I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx Table 4 First 6 components from the PCA analysis of 21 quantitative traits in 146 sweet cherry cultivars. Traits F1 F2 F3 F4 F5 F6 StV YL VSt/Vfr WeSt/WeFr PL PeWi PL/Wi BlLe BlWi BlLe/ PeLe StLe StWi StTh SoSo TiAc TiAc/ SoSo PeLe PoDi EqDi FrWi FrWe Cumulative% Variability (%) 0257 0.132 0.005 0.102 0.266 0.240 −0.033 −0.050 0.270 0.133 0.165 0.242 0.228 0.013 0.125 −0.114 0.000 0.305 0.360 0.394 0.373 22.906 22.906 −0.256 −0.078 0.414 0.483 0.043 −0.041 −0.105 0.189 0.003 −0.125 −0.300 −0.320 −0.328 0.031 −0.015 0.043 −0.161 0.185 0.189 0.136 0.200 39.661 16.755 0.014 0.016 −0.128 −0.004 0.340 0.300 −0.102 0.038 0.298 0.275 −0.318 −0.243 −0.144 0.225 0.395 −0.296 0.100 −0.232 −0.167 −0.096 −0.158 49.815 10.155 0.266 −0.214 −0.202 0.086 −0.329 −0.167 0.266 0.059 −0.225 −0.357 −0.043 −0.014 −0.096 0.200 0.438 −0.377 0.101 −0.029 0.145 0.145 0.083 59.374 9.558 −0.037 −0.196 −0.019 −0.112 0.188 0.261 0.118 0.653 0.225 −0.440 0.079 0.098 0.029 −0.022 −0.140 0.125 0.314 −0.052 −0.007 −0.069 0.008 67.083 7.710 0.043 0.431 0.054 0.037 −0.146 −0.007 0.178 −0.093 0.007 0.021 −0.091 −0.058 −0.054 0.567 −0.029 0.445 0.447 −0.011 0.001 0.094 0.050 73.590 6.506 Trait abbreviation: Stone volume [StV], Yield [YL], Ratio volume stone/volume fruit [VSt/VFr], Ratio weight stone/weight fruit [WeSt/WeFr], Petiole length [PL], Petiole width [PeWi], Ratio petiole length/width [PL/Wi], Blade length [BlLe], Blade width [BlWi], Ratio blade length/petiole length [BlLe/ PL], Stone length [StLe], Stone width [StWi], Stone thickness [StTh], Soluble solids [SoSo], Titratable acidity [TiAc], ratio titratable acidity/soluble solids [TiAc/ SoSo], Pedicel length [PeLe], Fruit polar diameter [PoDi], Fruit equatorial diameter [EqDi], Fruit width [FrWi] and Fruit weight [FrWe]. and ratio of blade length/petiole length. The sixth component explained 6.50% of the total variation, included yield, soluble solids, pedicel length and ratio of titratable acidity/soluble solids. Furthermore, a PCA scatter plot was constructed based on the first three components (Fig. 2). The plot grouped the cultivars according to their phenotypic similarity and morphological traits. For instance, cultivars Bakirtzeika and Lemonidi with the highest fruit weight, largest fruit width and longest fruit length were placed closely in the upper right plane. These results demonstrate that fruit equatorial diameter, fruit weight and fruit width are highly positively correlated and as a result, these morphological traits led to the highest loading factors in this PCA analysis. Thus, PCA showed that some traits had the highest loadings in the first two components. These traits included stone volume, petiole length, blade width, fruit polar diameter, fruit equatorial diameter, and fruit width and fruit weight. These results indicate that such traits are suitable both for the assessment of genetic diversity and for phenotypic characterization of sweet cherry germplasm (Fig. 3). 3.3. Correlations for qualitative traits There were more significant negative correlations than positive ones among the 14 qualitative traits (Table 5). Negative correlations were observed between the following traits: FiFl and TrBr (−0.297), StSi and TrBr (−0.293), FrSh and FiFl (−0.269), TrVi and FrSi (−0.262), FrSh and FrSi (−0.244), TrBr and TrVi (−0.175), and LaSh and TrVi (−0.174). Positive correlations included TrBr and NuNe (0.364), FrSi and StSi (0.342), FrSi and LaSi (0.283) LaSh and NuNe (0.280) and FiFl and FrSi (0.229). 3.4. Principal component analysis (PCA) of qualitative variables Using Kaiser’s criterion (“Eigenvalue” >1) (Kaiser, 1958), we obtained 6 significant components, which explained 65.11% of the total variation (Fig. 1). The first component, which accounted for 17.24% of the total variation, included tree branching, firmness of flesh, fruit size, number of nectaries and lamella size. The sec- Fig. 2. (A) Three-dimensional PCA plot of the 146 sweet cherry cultivars with regard to the first three principal components. (B) Three-dimensional PCA plot of the 21 quantitative traits with regard to the first three principal components. Variability explained: F1 (22.91%), F2 (16.76%), F3 (10.16%). Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model ARTICLE IN PRESS HORTI-6088; No. of Pages 11 8 I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx Fig. 3. (A) Three-dimensional PCA plot of the 146 sweet cherry cultivars with regard to the first three principal components. (B) Three-dimensional PCA plot of the 14 qualitative traits with regard to the first three principal components. Variability explained: F1 (17.25%), F2 (10.95%), F3 (10.43%). Table 5 Correlation coefficients (Spearman) among 14 qualitative traits in 146 sweet cherry cultivars. Variables TrVi TrHa TrVi TrHa TrBr LeSi LaSh LaSi NuNe PeSh FrSh FrSi FrSCo FiFl StSh StSi 1 0.108 −0.175 −0.042 0.012 −0.174 −0.055 0.076 0.017 −0.262 0.007 −0.028 −0.114 0.037 1 −0.115 −0.172 −0.071 −0.141 −0.025 −0.151 −0.009 −0.051 0.099 −0.065 −0.073 0.035 TrBr 1 −0.262 0.005 −0.112 0.364 −0.021 0.251 −0.138 0.127 −0.297 −0.105 −0.293 LeSi 1 0.107 −0.015 −0.067 −0.086 −0.016 0.115 −0.046 0.135 0.160 0.060 LaSh LaSi NuNe 1 −0.160 0.280 0.128 0.043 −0.145 0.097 −0.075 −0.128 −0.168 1 −0.119 −0.029 −0.161 0.283 −0.031 0.181 0.069 0.279 1 0.008 0.217 −0.075 0.208 −0.259 −0.175 0.051 PeSh FrSh FrSi FrSCo FiFl StSh StSi 1 −0.087 −0.105 0.122 0.115 −0.059 −0.059 1 −0.244 −0.166 −0.269 −0.237 −0.084 1 −0.014 0.229 −0.067 0.342 1 0.069 0.158 −0.062 1 0.138 0.152 1 −0.143 1 Values in bold are different from 0 with a significance level alpha = 0.05. Trait abbreviation: Tree vigour [TrVi], tree habit [TrHa], tree branching [TrBr], lenticels size [LeSi], lamella shape [LaSh], number of nectaries [NuNe], petal shape [PeSh], fruit shape [FrSh], fruit size [FrSi], fruit skin colour [FrSCo], firmness of flesh [FiFl], stone shape [StSh] and stone size [StSi]. ond component, which explained 10.94% of the total variation, was determined by fruit size, stone size and stone shape. The third component, explained 10.43% of the total variation, and was mainly correlated to characters related to tree vigour, fruit skin colour and tree habit. The fourth component, accounted for 9.31% of the total variation included lamella shape and stone shape. The fifth component, explained 9.14% of the total variation, included tree habit, lenticels size and fruit skin colour. The sixth component, explained 8.03% of the total variation, and included petal shape, tree habit and lenticels size (Table 6). Table 6 First 7 components from the PCA analysis of 14 qualitative traits in 146 sweet cherry cultivars. F1 F2 F3 F4 F5 F6 F7 TrVi TrHa TrBr LeSi LaSh LaSi NuNe PeSh FrSh FrSi FrSCo FiFl StSh StSi 0.077 0.043 0.394 −0.184 0.218 −0.329 0.341 0.032 0.345 −0.365 0.057 −0.389 −0.179 −0.301 −0.329 −0.057 0.252 −0.128 −0.164 0.302 0.272 −0.275 0.243 0.387 −0.204 −0.184 −0.352 0.363 −0.447 −0.385 0.306 0.094 0.224 0.165 0.267 0.207 -0.204 0. 0.406 0.121 0.285 −0.182 −0.290 −0.010 0.223 −0.045 −0.431 0.017 −0.327 −0.392 0.073 −0.117 −0.195 −0. 0.440 -0.380 −0.063 −0.484 −0.155 0.586 0.287 −0.076 −0.046 −0.045 0.287 −0.049 −0.460 −0.019 −0.033 −0.066 0.076 −0.411 0.147 −0.415 −0.202 0.224 −0.260 0.559 0.034 −0.122 −0.290 0.103 −0.197 −0.099 −0.458 0.316 −0.052 −0.151 0.349 −0.254 −0.244 0.056 −0.193 0.274 −0.246 0.119 −0.328 −0.348 Cumulative% Variability (%) 17.249 17.249 28.196 10.947 38.627 10.431 47.937 9/310 57.086 9.149 65.117 8.031 71.329 6.212 Trait abbreviation: Tree vigour [TrVi], tree habit [TrHa], tree branching [TrBr], lenticels size [LeSi], lamella shape [LaSh], number of nectaries [NuNe], petal shape [PeSh], fruit shape [FrSh], fruit size [FrSi], fruit skin colour [FrSCo], firmness of flesh [FiFl], stone shape [StSh] and stone size [StSi]. Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model HORTI-6088; No. of Pages 11 ARTICLE IN PRESS I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx 9 Fig. 4. Dendrogram using agglomerative hierarchical clustering (AHC) for 146 sweet cherry cultivars based on 21 quantitative and 14 qualitative traits. 3.5. Dendrogram using agglomerative hierarchical clustering (AHC) Unsupervised agglomerative hierarchical cluster analysis was used in order to divide the available data up into groups of increasing dissimilarity. The Euclidean distance was used as a metric to measure the genetic dissimilarity of the 146 sweet cherry cultivars, based on the combined quantitative and qualitative data, and the Ward’s method was used for the agglomeration. The dendrogram of Fig. 4 pointed out that, all sweet cherry cultivars were different from each other, and several clusters are identified. The dendrogram revealed three distinct groups. C1 contained 75 cultivars, C2 had 5 cultivars, and finally C3 included 66 cultivars. Among all 146 cultivars from different world regions, there were no specific clusters based on locality. The highest genetic distance exists between C2 and C3 (115.60), followed by C1 and C2 (71.14) and C1 and C3 (46.06). 4. Discussion The main goal of germplasm management is to collect and to characterize diverse species or forms of the same species at national and regional level. To do so it is important to perform, as a first step, the evaluation of morphological and agronomic traits of interest. Plant breeders then routinely use this morphological characterization for the initial description and classification of the germplasm under consideration. We estimated the morpho-physiological traits variation in sweet cherry cultivars established in the Greek GeneBank and provided basic knowledge on the range of variation of several morphological traits. The studied cultivars displayed significant variation in fruit weight traits as well as moderate variation for yield, stone volume, ratio of titratable acidity/soluble solids, titrat- able acidity, and ratio of volume stone/volume fruit (Pommer, 2012). Herein, the largest variation among the fruit traits corresponded to fruit weight and yield as it has been previously reported (Christensen, 1974; Ganopoulos et al., 2011; Hjalmarsson and Ortiz, 2000; Petruccelli et al., 2013). In addition, these authors have also suggested that the fruit weight could represent the utmost significant trait for distinguishing sweet cherry cultivars. Furthermore, the fruit weight (fruit size) is a very crucial trait due to its economic importance: some local cultivars, e.g., ‘Lemonidi’ and ‘Mpakirtzeika’, have shown relative large fruits with considerably elevated weights and volumes. The morphological traits that were observed as crucial for sweet cherry cultivars characterization in this study were fruit weight and fruit size which were also found to be important in sweet cherry from Greece (Ganopoulos et al., 2011; Petruccelli et al., 2013), which indicates similar diversity patterns in the Mediterranean region. The fruit weight showed positive correlation with stone weight (r = 0.495, p < 0.001) and also fruit equatorial diameter with fruit width (r = 0.844, p < 0.001) in agreement with (Khadivi-Khub, 2014). Our results also showed a very close correlation among fruit weight and fruit dimensions (length, width and diameter; r = 0.82, r = 0.90 and r = 0.86, respectively); therefore, these parameters could potentially be used to predict each other. A degree of correlation between fruit weight and fruit dimensions was also detected earlier in cherries (Khadivi-Khub, 2014; Rakonjac et al., 2010). All variables related to fruit size (fruit weight and dimensions) showed significant positive correlation with leaf traits (petiole length and petiole width) at p = 0.05, indicating a role of leaf in the increased fruit size (Demirsoy and Demirsoy, 2004). A positive correlation between the sweet cherry fruit size and leaf traits was also determined before by Rakonjac et al. (2010). Posi- Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model HORTI-6088; No. of Pages 11 10 ARTICLE IN PRESS I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx tive correlations were also obtained between leaf traits with each other. Most of the qualitative traits evaluated are considered crucial for species registration and discrimination in the test guidelines proposed by UPOV (UPOV, 1976). Notably, the fruit characteristics, such as skin color, flesh colour and firmness, are those that typically differentiate sweet cherry cultivars (Antonius et al., 2012; Hjalmarsson and Ortiz, 2000; Petruccelli et al., 2013). The skin color of the fruit is a very important quality characteristic, which also helps the assessment of the stage of fruit maturity as it has been shown when (Esti et al., 2002) evaluated the changes in color of sweet cherry and used this trait for monitoring pigment evolution. Moreover, fruit color has a significant impact on consumer perception of fruit quality, especially as regards the attractiveness of fruit (Ruiz and Egea, 2008). Consumers generally seem to prefer dark red cherries (Petruccelli et al., 2013). Another important characteristic is fruit firmness which is relevant to an assessment of the quality of fruit, fruit shelf life, and consumer acceptance. Fruit firmness is a combination of skin and flesh strength, and in general cultivars with the firmest fruit are preferred by consumers increasing as a result their value in the market (Kappel et al., 1996; Petruccelli et al., 2013). Multivariate statistical methods such as PCA and cluster analysis could be valuable tools for monitoring cultivars, and characterizing and classifying plant germplasm in GenBank collections (Iezzoni and Pritts, 1991). Further, PCA is useful in defining the number of main factors, thus decreasing the number of efficient parameters to differentiate genotypes. Additionally, associations between characteristics emphasized by this method may correspond to genetic linkage between loci controlling traits with a pleiotropic effect (Iezzoni and Pritts, 1991; Rakonjac et al., 2010). Many studies, in agreement with our study, indicated that fruit and leaf traits are crucial factors in phenotyping and morphologically characterizing the diversity in sweet cherry breeding materials (Antonius et al., 2012; Ganopoulos et al., 2011; Hjalmarsson and Ortiz, 2000; Lacis et al., 2009; Rakonjac et al., 2010). Furthermore, PCA analysis has been widely used for evaluation of sweet cherry germplasm (Beyer et al., 2002; Christensen, 1974; Ganopoulos et al., 2011; Hjalmarsson and Ortiz, 2000; Khadivi-Khub, 2014; Lacis et al., 2009; Petruccelli et al., 2013; Rodrigues et al., 2008; Sánchez et al., 2008). We have used PCA for the identification of the most significant variables in the data set presented herein. For each factor, a principal component loading of more than 0.55 was considered significant, which indicated that seven components explained 73.59% of the total variance. The first three components, consisted of 21 quantitative variables, explained 49.81% of the total variability observed (Table 4), indicating that these attributes have the highest variation between the cultivars and had the greatest impact on separation of the cultivars (Iezzoni, 2008; Khadivi-Khub, 2014; Lacis et al., 2009). PCA analysis of the qualitative variables has shown similar results (Table 5). Further, PCA analysis achieved the characterization of cultivars groups by identifying highly discriminating variables like traits regarding the fruit. Thus, PCA analysis suggested that future evaluations can rely on a reduced number of traits with a minimum loss of information for the discrimination and characterization of the different varieties/cultivars resulted in a reduction in labour, time and cost. Fig. 4 depicts the unsupervised hierarchical cluster analysis of 146 sweet cherry cultivars. The cultivars analyzed grouped into three main clusters. (Khadivi-Khub, 2014) characterized 41 sweet cherry genotypes with morphological descriptors and found that all genotypes were divided into five clusters. The dendrogram produced showed high diversity between sweet cherry cultivars indicating that analyzed germplasm collection could be assumed in breeding programs as a good gene pool for contrasting traits. For instance, sweet cherry cultivars ‘Cuglyeva Acacra’ from Bulgaria and ‘Merton Bigarreau’ from U.K. were highly distinguished from other cultivars. 5. Conclusions PCA combined with unsupervised cluster analysis for agronomic, morphological and fruit quality traits revealed a wide diversity in sweet cherry maintained in Germplasm Bank at the Institute of Pomology (Greece). Understanding the phenotypic diversity among the sweet cherry cultivars is crucial for the conservation of traditional genetic material that is endangered. Furthermore, this information is very useful for registration of new sweet cherry cultivars carried out by EU-Community Plant Variety Office. The study presented herein provides useful information on agronomic, morphological and fruit quality traits of sweet cherry cultivars grown under the climatic conditions of Imathia area in North Greece. Future breeding programs could take advantage of those genetic materials and use them as parental genotypes in order to achieve genetic recombination and improvement in agronomic, morphological and fruit quality traits. Thus, new strategies could be created for new breeding programs of sweet cherry cultivars with better adaptation to the limiting agro-climatic conditions of Greece. When separate traits are analyzed, information of paramount importance about the internal similarities of cultivars and genetic structure is lost. Contrariwise, when combined application of PCA and cluster analysis is used, important traits for sweet cherry cultivars characterization are revealed and provide comprehensive information. Hence, in order to thoroughly analyze the diversity of a GeneBank collection, a multivariate statistics approach taking into account both the genetic and the phenotypic data is of utmost importance. Conflict of interest The authors declare that they have no conflict of interest. References Antonius, K., Aaltonen, M., Uosukainen, M., Hurme, T., 2012. Genotypic and phenotypic diversity in Finnish cultivated sour cherry (Prunus cerasus L.). Genet. Resour. Crop Evol. 59, 375–388. Beyer, M., Hahn, R., Peschel, S., Harz, M., Knoche, M., 2002. Analysing fruit shape in sweet cherry (Prunus avium L.). Sci. Hortic. 96, 139–150. Christensen, J.V., 1974. Numerical Studies of Qualitative and Morphological Characteristics of 41 Sweet Cherry Cultivars 2. Tidsskrift for Planteavl. de Oliveira, E.J., Dias, N.L.P., Dantas, J.L.L., 2012. Selection of morpho-agronomic descriptors for characterization of papaya cultivars. Euphytica 185, 253–265. Demirsoy, H., Demirsoy, L., 2004. A study on the relationships between some fruit characteristics in cherries. Fruits 59, 219–223. Díez, C.M., Imperato, A., Rallo, L., Barranco, D., Trujillo, I., 2012. Worldwide core collection of olive cultivars based on simple sequence repeat and morphological markers. Crop Sci. 52, 211–221. Esti, M., Cinquanta, L., Sinesio, F., Moneta, E., Di Matteo, M., 2002. Physicochemical and sensory fruit characteristics of two sweet cherry cultivars after cool storage. Food Chem. 76, 399–405. Furones-Pérez, P., Fernández-López, J., 2009. Morphological and phenological description of 38 sweet chestnut cultivars (Castanea sativa Miller) in a contemporary collection. Span. J. Agric. Res., 829–843. Ganopoulos, I.V., Kazantzis, K., Chatzicharisis, I., Karayiannis, I., Tsaftaris, A.S., 2011. Genetic diversity, structure and fruit trait associations in Greek sweet cherry cultivars using microsatellite based (SSR/ISSR) and morpho-physiological markers. Euphytica 181, 237–251. Hedrick, U.P., 1915. The Cherries of New York. J. B. Lyon, Albany, N.Y. Hjalmarsson, I., Ortiz, R., 2000. In situ and ex situ assessment of morphological and fruit variation in Scandinavian sweet cherry. Sci. Hortic. 85, 37–49. Iezzoni, A.F., 2008. Cherries Temperate Fruit Crop Breeding. Springer, pp. 151–176. Iezzoni, A.F., Pritts, M.P., 1991. Applications of principal component analysis to horticultural research. HortScience 26, 334–338. IPGRI, 1985. Cherry Descriptors. International Plant Genetic Resources Institute, Rome, Italy, pp. 33. Kaiser, H.F., 1958. The varimax criterion for analytic rotation in factor analysis. Psychometrika 23, 187–200. Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061 G Model HORTI-6088; No. of Pages 11 ARTICLE IN PRESS I. Ganopoulos et al. / Scientia Horticulturae xxx (2015) xxx–xxx Kappel, F., Fisher-Fleming, B., Hogue, E., 1996. Fruit characteristics and sensory attributes of an ideal sweet cherry. HortScience 31, 443–446. Kazantzis, K., 2013. Monograph of Sweet Cherry Cultivars Evaluated from Institute of Pomology in Greece. Institute of Pomology, Elgo ‘Demeter’, 218. Khadivi-Khub, A., 2014. Assessment of cultivated cherry germplasm in Iran by multivariate analysis. Trees 28, 669–685. Khodadadi, M., Fotokian, M.H., Miransari, M., 2011. Genetic diversity of wheat (Triticum aestivum L.) genotypes based on cluster and principal component analyses for breeding strategies. Aust. J. Crop Sci. 5, 17–24. Koukourojiannis, V., 1996. O␫ ␶˛␴ ´ ␫ς ␴␶␩␯ ␲␣␳␣␥␻␥´ ␬␣␫ ␮␲о␳´␫␣ ␶␻␯ ´ Ŵ␻␳␥´␫␣-K␶␩␯о␶␳о␸´␫␣ (Greek) 2, 24–31. ␬␳␣␴␫␻␯. Lacis, G., Kaufmane, E., Trajkovski, V., Rashal, I., 2009. Morphological variability and genetic diversity within Latvian and Swedish sweet cherry collections. Acta Univ. Latv. 753, 19–32. Marshall, R.E., 1954. Cherries and cherry products. Economic Crops, vol. 5. Inter-science, New York. Mehmood, A., Jaskani, M.J., Khan, I.A., Ahmad, S., Ahmad, R., Luo, S., Ahmad, N.M., 2014. Genetic diversity of Pakistani guava (Psidium guajava L.) germplasm and its implications for conservation and breeding. Sci. Hortic. 172, 221–232. Mohammadi, S.A., Prasanna, B.M., 2003. Analysis of genetic diversity in crop plants—salient statistical tools and considerations. Crop Sci. 43, 1235–1248. Peeters, J.P., Martinelli, J.A., 1989. Hierarchical cluster analysis as a tool to manage variation in germplasm collections. Theor. Appl. Genet. 78, 42–48. Petruccelli, R., Ganino, T., Ciaccheri, L., Maselli, F., Mariotti, P., 2013. Phenotypic diversity of traditional cherry accessions present in the Tuscan region. Sci. Hortic. 150, 334–347. 11 Pommer, C.V., 2012. Guava world-wide breeding: major techniques and cultivars and future challenges. In III International Symposium on Guava and other Myrtaceae 959, 81–88. Rakonjac, V., Akšić, M.F., Nikolić, D., Milatović, D., Čolić, S., 2010. Morphological characterization of ‘Oblačinska’sour cherry by multivariate analysis. Sci. Hortic. 125, 679–684. Rodrigues, L.C., Morales, M.R., Fernandes, A.J.B., Ortiz, J.M., 2008. Morphological characterization of sweet and sour cherry cultivars in a germplasm bank at Portugal. Genet. Resour. Crop Evol. 55, 593–601. Ruiz, D., Egea, J., 2008. Phenotypic diversity and relationships of fruit quality traits in apricot (Prunus armeniaca L.) germplasm. Euphytica 163, 143–158. Sánchez, R.P., Sánchez, M.A.G., Corts, R.M., 2008. Agromorphological characterization of traditional Spanish sweet cherry (Prunus avium L.), sour cherry (Prunus cerasus L.) and duke cherry (Prunus x gondouinii Rehd.) cultivars. Span. J. Agric. Res. 6, 42–55. Trujillo, I., Ojeda, M.A., Urdiroz, N.M., Potter, D., Barranco, D., Rallo, L., Diez, C.M., 2014. Identification of the worldwide olive germplasm bank of Córdoba (Spain) using SSR and morphological markers. Tree Genet. Genomes 10, 141–155. UPOV, 1976. Guidelines for the Conduct of Test for Dis- Tintness, Homogeinity and Stability of the Cherry. International Union for the Protection of New Varieties of Plants, Genova, Italy, pp. 15. Webster, A.D., 1996. The Taxonomic classification of sweet and sour cherries and a brief history of their cultivation. In: Webster, A.D., Looney, N.E. (Eds.), Cherries. Cab International, Wallingford OX 10 8DE, UK, pp. 3–25. Please cite this article in press as: Ganopoulos, I., et al., Diversity of morpho-physiological traits in worldwide sweet cherry cultivars of GeneBank collection using multivariate analysis. Sci. Hortic. (2015), http://dx.doi.org/10.1016/j.scienta.2015.09.061