Impaginato 581 Adv. Hort. Sci., 2019 33(4): 581­592 DOI: 10.13128/ahsc­7908 Morphological and molecular charac­ terization of ancient pomegranate (Punica granatum L.) accessions in Northern Italy D. Beghè 1 (*), A. Fabbri 1, R. Petruccelli 2, M. Marieschi 3, A. Torelli 3,, T. Ganino 1, 2 1 Department of Food and Drug Science, University of Parma. Parco Area delle Scienze, 27/a, 43124 Parma, Italy. 2 Institute for the Bioeconomy (IBE), Italian National Research Council (CNR), 50019 Sesto Fiorentino, Italy. 3 D e p a r t m e n t o f C h e m i s t r y , L i f e S c i e n c e s a n d E n v i n r o n m e n t a l Sustainability, University of Parma. Parco Area delle Scienze, 11/a, 43124 Parma, Italy. Key words: genetic diversity, phenotype, pomegranate, RAPDs, SSRs, varieties. Abstract: The Italian research on P. granatum L. is still limited, although the study of local germplasm is extremely important in order to preserve the exist­ ing biodiversity and to identify potential useful characters for a renewed indus­ try. The study aimed at characterizing for the first time ancient pomegranates, grown in Emilia Romagna (Italy), through 38 quantitative morphometric descriptors related to leaf, flower, fruit and seed, 42 RAPD and 12 SSR markers. Morphological analyses showed large variation of traits among accessions and the descriptors related to fruit and seed had the highest power of discrimina­ tion. The considerable variation found was consistent with ANOVA and PCA results. Among all RAPDs tested, 7 were selected for their polymorphism; whereas among selected SSRs, 8 presented differences in the genetic profiles allowing a good discrimination of the local pomegranate accessions. The genet­ ic relationships among pomegranates were studied by UPGMA cluster analysis and the accessions were clearly regrouped in four different genotypes. The study has highlighted significant differences and interesting pomological char­ acteristics in the local pomegranates, which confirmed the good potential of this germplasm for the pomegranate industry. 1. Introduction Pomegranate (Punica granatum L.) is one of the world most ancient domesticated fruit crops and it is believed to have been first grown in the region between the Caspian Sea and the Caucasus (Zohary and Spiegel­ Roy, 1975). Its diffusion occurred across the millennia due to man’s activi­ ties or gene flow in quite varied environments as concerns climatic condi­ tions, has produced a rich and diversified germplasm. P. granatum L. has (*) Corresponding author: deborah.beghe@unipr.it Citation: BEGHÈ D., FABBRI A., PETRUCCELLI R., MARIESCHI M., TORELLI A., GANINO T., 2019 ­ Morphological and molecular characterization of ancient pome‐ granate (Punica granatum L.) accessions in Northern Italy. ­ Adv. Hort. Sci., 33(4): 581­592. Copyright: © 2019 Beghè D. , Fabbri A., Petruccelli R., Marieschi M., Torelli A., Ganino T. This is an open access, peer reviewed article published by Firenze University Press (http://www.fupress.net/index.php/ahs/) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Competing Interests: The authors declare no competing interests. Received for publication 27 June 2019 Accepted for publication 22 October 2019 AHS Advances in Horticultural Science http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ Adv. Hort. Sci., 2019 33(4): 581­592 582 a l a r g e g e n e ti c p o o l , r e p r e s e n t e d b y o v e r 5 0 0 described cultivars throughout the world, and by a wide amount of wild plants, and its germplasm is so far only partially explored (Beghè et al., 2016). In spite of the wide genetic diversity, only 50 cultivars were widely cultivated in the main growing areas at the time of the last official survey made by interna­ tional institutions (IPGRI, 2001). Consequently, the risk of a drastic loss of the existing biodiversity is high. In the last decade the interest on pomegranate has grown and the world production (estimated to be around 1.5 million tonnes, of which 90% provided by the main producers: India, China, Iran) has rapidly increased (da Silva et al., 2013). The renewed interest in this crop is to be ascribed to socio­economical and cultural factors, which led to a change in food habits in the West, with a growing attention to the nutrition­ al quality of foods (Negri, 2003). Pomegranate, in this respect, is considered one the fruits with the most valuable nutritional properties (Calani et al., 2013). Pomegranate cultivation is also growing in Italy, and in a ten­year period (since 2008) the surface cov­ ered, from a mere 7 ha (2 ha in Calabria and 5 in Sicily), passed to 1142 ha (mainly in Southern Italy, Sicily (364 ha) and Apulia (363 ha) and, with minor productions, also in Venetia (152 ha), Latium (101 ha), Emilia Romagna (44 ha), Tuscany (21 ha) and Lombardy (20 ha). Productions rose from only 69 tons to over 12531 tons (AGRISTAT, 2017). One of the trends of recent years has been to import from Israel pomegranate cultivars created and patented by Israel breeders, to be utilised in Italian orchards. Although this strategy has led to an increase in production, it hinders any attempt at exploiting Italian cultivars, which have been known for centuries but are not utilised in commercial culti­ vation. In Italy, there are several local cultivars which are little known and scarcely diffused in the country. Scarce research, mostly confined to Southern Italy, has been devoted to the characterization of P. grana‐ tum L. Italian diversity, by means of molecular, mor­ phological and biochemical markers (Adiletta et al., 2018). A research on the whole national territory is unavoidable to select autochthonous genotypes suit­ ed to the different environments, and to promote a pomegranate industry able to satisfy the internal demand and to compete with the international product. The best suited Italian areas for pomegranate growing are the central and southern regions, char­ acterized by a Mediterranean climate, unlike the northern regions, which have a continental climate, with cold and snowy winters. However, in some zones of Emilia­Romagna (44°­45° N and 11°­12° E, Northern Italy) microclimate conditions are such as to allow the cultivation of this species and of other Mediterranean species like olive (Lona et al., 1981; Calani et al., 2013). Ancient pomegranate trees, survived for hun­ dreds of years and adapted to local conditions have been retrieved in these areas. This local germplasm is of particular interest, being the result of a selection process occurred during many centuries in the unfa­ vorable conditions of this territory. The present research is part of a multidisciplinary project aiming at valorising ancient pomegranate cul­ tivars of Northern Italy. We utilized morphological a n d m o l e c u l a r m a r k e r s ( R a n d o m A m p l i fi e d Polymorphic DNA (RAPDs) and Simple Sequence Repeat (SSRs)) to characterize ancient pomegranate accessions present in Emilia­Romagna Region. The pomological comparisons also had the purpose of determining the peculiar features and the potential of these plants, for a possible introduction in commer­ cial plantings or for their use in breeding programs. 2. Materials and Methods Plant materials The pomegranate germplasm subject of this study was represented by very ancient trees located in a small hill area of Parma province (Emilia­Romagna) (44°69 N, 10°02 E), at an altitude from 150 to 250 m a.s.l. Eight accessions of P. granatum L., tagged with an alpha­numerical code (ID): ME1, ME2, ME3, ME4, ME5, ME6, ME7 and ME8 (Table 1), were studied during two seasons, 2014 and 2015. The selected plant material was maintained at pomegranate germplasm collection field in the low hills of the Emilia Appennins, where each accession was repli­ cated 4 times. The pomegranate trees were planted at a spacing of 5 x 5 m and trained to form a bush. Morphological characterization The accessions were characterized according to the guidelines proposed by the project EC Project GENRES 29 “Conservation, evaluation, exploitation and collection of minor fruit tree species” (Bellini and Giordani, 1998), integrated by the list of characters p r o p o s e d b y B e l l i n i e t a l . ( 2 0 0 7 ) a n d b y t h e I n t e r n a ti o n a l U n i o n f o r t h e P r o t e c ti o n o f N e w Varieties of Plants (UPOV, 2012). P l a n t m a t e r i a l w a s r a n d o m l y s a m p l e d f r o m Beghè et al. ‐ Characterization of ancient pomegranates in Northern Italy 583 around the canopy of four plants per each accession, by collecting 40 flowers (20 hermaphrodite also called “long­styled” and 20 male also called “short­ styled”) at full bloom, on June 1st, 40 adult leaves, from the middle part of the shoot in summer and 12 fruits, at ripening, in the first decade of October. All seeds were extracted from each fruits and 25 of them were randomly selected; arils (the seed fleshy coats, containing edible juice, that represent the seed outer integument or testa) were hand removed to analyze also the tegmen (seed lignified inner integument). The morphological characters evaluat­ ed included 38 quantitative traits (Table 2). The linear dimensions were determined with a caliper, and the weight was measured using a semi­analytic electronic scale. From these values other indices have been cal­ culated as indicated in Table 2. Furthermore, some qualitative characters were observed; these traits are reported in Table 1. DNA extraction and molecular characterization Total cellular DNA was extracted from young leaves following the CTAB (cetyl trimethylammonium bromide) as reported in Ganino et al. (2008). F o r t y ­ t w o d e c a m e r o l i g o n u c l e o ti d e p r i m e r s belonging to the AI, AH, OPA, OPC, OPX and OPK series (Table S1 of supplementary data) and twelve couples of SSR primers belonging to the PgAER (Çalişkan et al., 2017), PGKVR (Ravishankar et al., 2015), Pom (Hasnaoui et al., 2012), POM­AGC (Currò et al., 2010) and PG (Ebrahimi et al., 2010) were used for polymorphism detection on the samples. RAPD a m p l i fi c a ti o n s w e r e p e r f o r m e d a s r e p o r t e d i n Marieschi et al. (2016). The RAPD profiles obtained with each utilized primer were analyzed by comparison with Gene Ruler 100 pb DNA Ladder plus marker (M­Medical, Milano, IT), with the Kodak digital sciences 1 D Images Analysis Software, calculating the size in base pairs (bp) of each amplicone present in the elec­ trophoretic run of each sample. SSR amplification reaction was performed as reported in Ganino et al. (2008). The amplification condition, for the PGKVR, PgAER, POM­AGC and PG series, were: a first step at 95°C for 5 min followed by 35 cycles of 45 s at 94°C, 45 s at 57°C, 45 s at 72°C, for denaturation, annealing, and primer extension; the last step included 8 min of incubation at 72°C. For the “Pom” serie, the following thermal cycling proto­ col was used: a first step at 95°C for 3 min followed by 10 touchdown cycles of 30 s at 94°C, 40 s at 65°C (­1°C per cycle), 30 s at 72 and 25 cycles of 30 s at 94°C, 30 s at 55°C, 40 s at 72°C with final extension time of 8 min at 72°C. The amplification products were separated with a CEQ 2000 Genetic Analysis System (Beckman Coulter, Inc.) sequencer on acry­ lamide gel CEQ Separation Gel LPA­1 (Beckman Coulter, Inc.). A marker CEQ DNA Size Standard kit 400 (Beckman Coulter, Inc.) was used to estimate the molecular weight of the amplified products. Data analysis The quantitative morphological characters were evaluated: means, minimum and maximum, standard deviation. The coefficient of variation (CV) was calcu­ lated as indicator of variability. All data were subject­ ed to one way analysis of variance (ANOVA) followed by Tukey test to determine the statistically significant differences (p≤0.05). Correlation analyses between descriptors to reveal possible relationships were car­ Table 1 ­ Punica granatum L. accessions studied, coded (ID) and main qualitative characteristics of their fruit, leaf and flower ID Fruit shape Size Epicarp colour Calyx type Leaf shape Petiol colour Mucro Blade colour Flower petal colour Shape short­ stiled Shape long­ stiled ME1 oblate/rounded­ spheroid large/very large reddish­yel­ low/red semi­closed/ closed elliptic yellow no yellow red/orange medium bell sinuolate jug ME2 oblate/rounded­ spheroid large/very large reddish­yel­ low/red semi­closed/ closed elliptic red no yellow red/orange medium bell sinuolate jug ME3 oblate/rounded­ spheroid large reddish­yel­ low/red semi­closed/ closed elliptic red no yellow red/orange medium bell sinuolate jug ME4 oblate/rounded­ spheroid large/very large reddish­yel­ low/red semi­closed/ closed elliptic red no yellow red/orange medium bell sinuolate jug ME5 oblate/rounded­ spheroid large/very large reddish­yel­ low/red semi­closed/ closed elliptic red no yellow red/orange medium bell sinuolate jug ME6 oblate/rounded­ spheroid large reddish­yellow semi­ closed/ open elliptic red no yellow red/orange broad bell jug with base ME7 oblate/rounded­ spheroid very small reddish­yel­ low/red open elliptic yellow no yellow red/orange narrow bell sinuolate jug ME8 oblate/rounded­ spheroid very small reddish­yel­ low/red open elliptic yellow no yellow red/orange narrow bell sinuolate jug Adv. Hort. Sci., 2019 33(4): 581­592 584 ried out using a bilateral Pearson correlation. The same characters were also submitted to a principal component analysis (PCA) to evaluate the relation­ ship between pomegranate accessions. The analysis w a s p e r f o r m e d u s i n g X L S T A T 2 0 0 9 s o ft w a r e (AddinsoftTM1995­2009). RAPD bands were treated as binary characters (present = 1 or absent = 0), XLSTAT 2009 software was used to estimate genetic similarities/dissimilari­ ties using Jaccard’s similarity coefficient and cluster analysis by using the unweighted pair­group method with arithmetic mean (UPGMA) algorithm. The size of SSR fragments was determined using a conservative binning approach (Kirby, 1990) through the statistical R software. The information content of the SSR markers under study was evaluated according to number of alleles per locus (Na), observed (Ho) and expected (He) heterozygosity, and polymorphic infor­ mation content (PIC) (Botstein et al., 1980) using the Cervus 3.0 software (Kalinowski et al., 2007). The level Table 2 ­ Quantitative traits used for characterizing pomegranate accessions and their descriptive statistics analysis using the mean, minimum, maximum, standard deviation (SD) and coefficient of variation (CV) (z) Aril weight were calculated by subtracting tegmen fresh weight from whole seed fresh weight. Trait Trait code Mean Minimum Maximum SD CV (%) Leaf Leaf fresh weight (g) LFW 0.10 0.070 0.160 0.029 29.13 Leaf length (cm) LL 5.44 4.450 6.920 0.828 15.21 Leaf width (cm) LW 1.62 1.260 2.150 0.278 17.21 Leaf shape (length/diameter) LS 3.50 2.970 3.910 0.341 9.76 Flower Flower diameter long­styled (cm) FDL 1.51 1.220 1.810 0.221 14.57 Flower length long­styled (cm) FLL 4.75 2.900 5.950 1.193 25.11 Petal number long­styled (cm) PNL 6.52 5.670 7.500 0.713 10.93 Pistil length long­styled (cm) PLL 1.75 1.500 2.030 0.231 13.20 Flower diameter short­styled (cm) FDS 1.48 1.200 1.770 0.163 11.06 Flower length short­styled (cm) FLS 3.73 2.840 4.490 0.564 15.10 Petal number short­styled PNS 6.50 6.000 7.400 0.510 7.86 Pistil length short­styled (cm) PLS 0.43 0.310 0.600 0.105 24.43 Fruit Fruit weight (g) FW 274.02 87.010 388.730 120.303 43.90 Fruit diameter equatorial (cm) FD 8.22 5.070 9.600 1.799 21.88 Calyx diameter equatorial (cm) CD 2.14 1.430 2.840 0.545 25.50 Fruit height without calyx (cm) FL1 6.64 4.400 8.120 1.380 20.80 Total fruit length (cm) FL2 8.11 5.570 9.530 1.568 19.34 Calyx height (cm) CL 1.56 1.170 2.220 0.323 20.67 Fruit skin thickness equatorial (mm) FT 0.42 0.300 0.550 0.096 22.69 Fruit skin and carpellary membranes weight (g) SCW 129.94 28.980 191.810 66.526 51.20 Number of carpel in equatorial section NC 6.88 5.330 8.000 0.993 14.43 Fruit shape index (height/diameter) FSI 0.81 0.730 0.890 0.059 7.21 Calyx shape index (height/diameter) CSI 0.80 0.610 0.880 0.095 11.93 % Skin and carpellary membranes SC (%) 45.10 33.300 53.400 7.499 16.63 % Seeds S (%) 54.92 46.600 66.700 7.491 13.64 Total seeds weight (g) STW 141.918 58.076 202.556 53.394 37.62 Seed weight (g) SW 0.27 0.144 0.381 0.080 29.57 Seed length (cm) SL 0.96 0.786 1.080 0.113 11.73 Seed diameter (cm) SD 0.69 0.547 0.804 0.090 13.04 Tegmen weight (g) TW 0.02 0.016 0.029 0.005 19.95 Tegmen length (cm) TL 0.65 0.544 0.769 0.074 11.35 Tegmen diameter (cm) TD 0.31 0.253 0.416 0.052 16.88 Woody portion index (tegment weight/aril weight ) WPI 0.09 0.071 0.121 0.021 22.12 Seed shape (length/diameter ) SL/SD 1.39 1.310 1.540 0.083 5.93 Tegment shape (length/diameter) TL/TD 2.15 1.880 2.410 0.176 8.19 Aril weight (g) z AW 0.25 0.120 0.350 0.080 32.49 Aril weight/tegmen weight AW/TW 10.04 7.400 12.390 2.162 21.53 % Aril A (%) 90.35 88.200 92.600 1.653 1.83 Beghè et al. ‐ Characterization of ancient pomegranates in Northern Italy 585 WPI, AW, AW/TW). These results are consistent with previous studies (Zamani et al., 2007, Mansour et al., 2011). The mean leaf quantitative values are reported in Table S2 and the traits values presented significant differences between the accessions. Moreover, leaf blade margin color and petiole color next to the shoot was red for accessions ME1, ME7 and ME8, and yellow for the others. All accessions have an elliptic shape and absence of mucro (Table 1). The flower characteristics are reported in Table 1 and Table S3 and the observed values are comparable with those of Lebanese genotypes studied by Dandachi et al. (2017). The flowers of ME7 and ME8 accessions showed a much smaller size than that of the flowers of the other, but presented a similar style length, a feature that favors pistil pollination. Moreover, the long­styled flowers presented a “sinu­ ate jug”, except the flowers of ME6 accession that pre­ sented “jug with base”. ME7 and ME8 presented a “narrow bell” shape in the short­styled flowers where­ as ME6 presented “broad bell” shape and other acces­ sions presented flowers with “medium bell” shape. The mean values of quantitative fruit and seed traits are reported in Tables 3 and S4. Significant vari­ ability was observed in total fruit weight (FW), in maximum equatorial diameter (FD) and in fruit length, with calyx (FL2) and without calyx (FL1). In particular, these characters have lower values for ME7 and ME8. All accessions showed fruits with shape “oblate or rounded­spheroid” and “closed or semi­ closed calyx”, except ME7 and ME8 that have a majori­ ty of fruits with “open calyx”. The number of locules (NC) was higher in fruits of higher total weight. of similarity/dissimilarity among examined accessions was obtained through the genetic similarity matrix uti­ l i z i n g M a n h a tt a n d i s t a n c e a n d c l u s t e r a n a l y s i s (UPGMA) algorithm, with XLSTAT 2009 software. Finally, to test the correlations between genetic distance matrices and between the morphological and genetic distance matrices among accessions, Mantel tests were performed (Mantel, 1967). Each matrix dis­ tance was obtained by calculating Pearson’s index. Mantel tests were performed with 100,000 permuta­ tions (p = 0.05). Pearson’s r­value was used to mea­ sure linear correlation between two matrices. 3. Results and Discussion Morphological characterization The data resulting from the 2­year study were grouped and the average values were used for statis­ tical analysis. The accessions showed significant vari­ a b i l i t y i n m a n y o f t h e c h a r a c t e r s a n a l y z e d . Descriptive values for each quantitative trait are recorded in Table 2. The coefficient of variation (CV) was used to determine the total variability present in each trait. The CV varied from 1.83% (A%) to 51.20% (SCW%), with seven traits having CV between 15 and 20% and fourteen traits with CV value higher than 20%. According to Audergon (1987), the descriptors with a high CV are more discriminating than the other ones, and can be reliable markers for the char­ acterization of pomegranate accessions. The highest CV values were evident in traits involving fruits and the lowest were in flowers (except FLL and PLS), leaves (except LFW) and seeds (except STW, SW, Table 3 ­ Mean values, standard deviation and ANOVA analysis for fruit characteristics The same letter show no statistically significant differences (P<0.05). (z) For explanation of character symbols, see table 2. ID FW (z) FD CD FL1 FL2 CL FT SCW NC FSI CSI SC% S% ME1 373.48 ±49.78 A 9.60 ±0.62 A 2.68 ±0.51 A 7.00 ±1.14 A 9.22 ±1.28 A 2.22 ±0.62 A 0.55 ±0.10 A 191.81 ±19.64 A 8.00 ±1.58 A 0.73 ±0.12 A 0.84 ±0.27 A 53.4 ±0.03 A 46.6 ±0.03 D ME2 303.53 ±93.46 A 8.83 ±1.17 A 2.40 ±0.62 A 7.83 ±0.72 A 9.53 ±1.06 A 1.70 ±0.43 A 0.53 ±0.11 AB 144.42 ±51.87 A 6.67 ±0.58 AB 0.89 ±0.07 A 0.70 ±0.03 A 47.2 ±0.03 AB 52.8 ±0.03 CD ME3 335.73 ±13.03 A 9.10 ±0.52 A 2.43 ±0.55 A 7.10 ±0.17 A 8.53 ±0.61 AB 1.43 ±0.49 A 0.44 ±0.10 BC 163.25 ±11.59 A 7.67 ±0.58 A 0.78 ±0.06 A 0.61 ±0.23 A 48.6 ±0.02 AB 51.4 ±0.02 CD ME4 388.73 ±89.49 A 9.40 ±0.84 A 2.15 ±0.78 A 8.12 ±1.03 A 9.00 ±1.27 AB 1.61 ±0.41 A 0.48 ±0.11 ABC 185.73 ±46.34 A 7.60 ±0.55 A 0.86 ±0.06 A 0.82 ±0.31 A 48.3 ±0.03 AB 51.7 ±0.02 CD ME5 345.68 ±87.85 A 9.50 ±1.14 A 2.84 ±0. 94 A 7.47 ±0.67 A 9.24 ±0.81 A 1.63 ±0.55 A 0.41 ±0.09 BC 181.74 ±47.63 A 7.25 ±0.58 AB 0.80 ±0.05 A 0.85 ±0.27 A 52.5 ±0.01 A 47.5 ±0.01 D ME6 266.81 ±23.78 AB 8.60 ±0.71 A 1.70 ±0.14 A 6.45 ±0.64 AB 7.95 ±0.78 AB 1.50 ±0.14 A 0.37 ±0.08 BCD 111.03 ±6.31 AB 7.00 ±0.45 AB 0.75 ±0.01 A 0.88 ±0.01 A 41.9 ±0.06 BC 58.2 ±0.06 BC ME7 87.01 ±21.77 B 5.67 ±0.57 B 1.43 ±0.06 A 4.73 ±0.38 B 5.82 ±0.58 B 1.25 ±0.11 A 0.30 ±0.09 D 28.98 ±7.48 B 5.33 ±0.58 B 0.84 ±0.02 A 0.88 ±0.19 A 33.3 ±0.01 D 66.7 ±0.01 A ME8 91.21 ±27.45 B 5.07 ±0.11B 1.47 ±0.06 A 4.40 ±0.10 B 5.57 ±0.23 B 1.17 ±0.15 A 0.30 ±0.79 D 32.54 ±10.48 B 5.50 ±0.55 B 0.87 ±0.03 A 0.80 ±0.13 A 35.5 ±0.02 CD 64.5 ±0.02 AB 586 Adv. Hort. Sci., 2019 33(4): 581­592 According to the list of “pomegranate descriptors” of Bellini et al. (2007), the fruits were classified as “large or very large” (ME1, ME2, ME3, ME4, ME5, ME6) and “very small” (ME7, ME8). The first group had a total weight mean values of about 335 g, comparable to those of the fruits of many Italian and Spanish cultivars (Martinez et al., 2006 and Ferrara et al., 2014). Moreover, the epicarp (or “skin”) of the local fruits has presented different colors, ranging from reddish­yel­ low to red. The size of the fruit and the color of the epicarp are two important parameters considered in the international market as concerns the quality of the fresh product (Mansour et al., 2011). Another impor­ tant parameter for fruit quality was the descriptor “skin thickness”. ME7 and ME8 fruits have a thinner “skin thickness” than other accessions, and in field their fruits were more subjected to cracking at the first rainfall in October. Significant variability was observed in seeds total weight (STW), skin and carpellary membranes weight (SCW) and seeds percentage (S%) (Tables S4 and 3). The mean S% found was of 54.91% and it was similar to that reported in another Italian study (Cristofori et al., 2011). Furthermore, our mean values were lower than those reported on Italian and Iranian cultivars (Ferrara et al., 2014). The seed descriptors showed, for the majority of traits, significant differences between two groups of accessions: ME1, ME2, ME3, ME4, ME5, ME6 and ME7, ME8; the former had SL, SD, SW, TL, TW greater than the latter (Table S4). Tegmen index (WPI), aril percentage (A%), aril weight (AW) and aril/tegmen ratio (AW/TW) are very important parameters from a qualitative point of view. The WPI is a parameter that refers to the quantity of lignified tissue contained in the seed compared to total seed weight, and consumers greatly appreciate seeds with a limited amount of lignified tissue (Martinez et al., 2006). Accessions ME1, ME2, ME3, ME4, ME5 had a higher quantity of aril (AW) and a lower percentage of the tegmen index (WPI) than ME6, ME7 and ME8. The WPI presented an average value of 7.7% in the first group of accessions, and an average value of 11.9% in the accessions ME6, ME7 and ME8. These values were in agreement with those of other Italian (5.4 to 10%), Spanish (7.4 to 9.7%), Moroccan (6.1 to 10.7%) and Iranian (5.4 to 7.5%) accessions (Martinez et al., 2006; 2012; Sarkhosh et al., 2009). The aril is a tissue valued for the high pro­ duction of juice; AW and A%, with reference to the individual seed, were high in all accessions, and A% showed an average value of about 90%; a value com­ parable and higher than that of other Italian and Spanish genotypes (La Malfa et al., 2009). Correlation among morphometric traits The correlations found between the quantitative variables, significant at p < 0.05, are reported in fig­ ure 1. The correlation coefficient can provide infor­ Fig. 1 ­ Pearson's correlation matrix of the quantitative traits in pomegranate accessions, visualized as a heat map plot. For explanation of character symbols, see table 2. Beghè et al. ‐ Characterization of ancient pomegranates in Northern Italy 587 mation on the traits that are most important in assessing accessions (Norman et al., 2011). Skinner et al. (1999) recommended analyzing correlation coeffi­ cients close to 0.7: in these conditions, the variance of one trait is strongly dependent on the others. According to this criterion, we estimated two hun­ dred and fifteen valid correlations. Most of the signif­ icant correlations among traits coincide with those from the same plant organ, in particular the variables relative to fruits and seeds. A strong correlation value was found for fruit and seed descriptors, and these were also positively correlated with each other (e.g. FW was positively correlated with other fruit descrip­ tors, FD, FL1, FL2, CL, SCW, NC, SC%, S% and seed descriptors, STW, SL, SD, SW, TW, A%, AW, AW/TW). The character WPI was negatively correlated with fruit traits relative to dimensions and with SW, SL, SD. The same results were observed in several for­ mer works that analyzed accessions from different countries (Martinez et al., 2006; Zarei et al., 2013). Moreover, the trait WPI could be an index of seed hardness, because this feature is strongly dependent on the trait WPI, as reported in Martinez et al. 2012 (r= 0.63 p ≤ 0.01). In flowers, positive correlations were found among the variables relative to dimen­ sions: between FLS and FDS and between FLL and FDL. In leaves, a positive correlation was between LD and LL, actually all accessions showed an elliptic leaf shape. Finally, some correlations were observed between seed (SW, A%, AW, TL, TW) and flower char­ acteristics (FDL, FLL, FDS, FLS, PNS); we noted that accessions with small flowers (ME7­ME8) presented fruits and seeds of smaller dimensions. In fact, fruit growth potential is largely determined genetically through the ovary size in several species (Rosati et al., 2009). These correlations between different traits could be due to genetic linkage or to a pleiotropic effect (i.e., when one gene influences two or more seemingly unrelated phenotypic traits) (Iezzoni and Pritts, 1991). Principal component analysis The results of PCA revealed the existence of large variability among accessions. The total variance explained by the first three principal components (PCs) in the model was 83.53%. A plot of the percent­ age of variance explained by seven PCs and eigenval­ ues associated with the first seven PCs for each quan­ titative trait are reported in supplementary material ( F i g . S 1 a n d T a b l e S 5 ) , r e s p e c ti v e l y . T h e P C 1 explained the 54.60 % of total variance and the traits with the greatest weight on this component were related to fruit (FW, FD, CD, FL1, FL2, CL, FT, SCW, NC, SC%), seed (STW, SL, SD, SW, TW, WPI, A%, S%) and some flower traits (FDL, FLL, FLS). The PC2 explained 15.82% of the variability, and showed a strong negative load for TD, LD and PNS, whereas a strong positive load was present for SL/SD ,TL/TD, FSI, and LS. Finally, LD, LL relating to the leaf and PLS relating to the flower showed the highest contribu­ tion to PC3 (13.11% of the variability). The compari­ son of plot scores for PC1, PC2 and PC3 in figure 2 permits to obtain a view of accession dispersion and their clustering based on morphological traits. The accessions, for the first two PCs, were grouped into two main groups highly dissimilar: the first group consists of ME7 and ME8, the second group consists of three sub­groups (sub­group ME6; sub­group ME2, ME3, ME4, ME5 and sub­group ME1). The groups plotted for the PC1 and PC3 were very similar to t h o s e o n P C 1 ­ P C 2 p l o t , t h o u g h a c c e s s i o n M E 6 showed less differences with the sub­group ME2, ME3, ME4, ME5. As already reported in the literature (Martinez­Nicolas et al., 2016), fruit characteristics Fig. 2 ­ Loading plots of the first, second and third Principal Component showing the position of accessions. Adv. Hort. Sci., 2019 33(4): 581­592 588 had the highest loading values for the first compo­ nent in principal component analysis. Our results confirmed that the traits related to fruit and seed had the highest power of discrimination, and were, therefore, the most useful for characterization of this local germplasm. RAPDs and SSRs characterization and genetic rela‐ tionships Only 7 RAPD oligonucleotides (AI08, AI12, AI105, AH17, OPA19, OPB08, OPC16) out of 42 belonging to the AI, AH, OPA, OPC, OPX and OPK series, showed polymorphism in two or more accessions, by produc­ ing polymorphic and reproducible amplification pat­ terns. The 7 oligonucleotides amplified a total of 84 RAPD fragments, 14 of which were polymorphic, making 16.67% polymorphism (Table 4). The number of polymorphic fragments found per primer was between 1 (OPB8, AI05 and AH17) and 3 (AI12, OPC16, AI08), with a mean of 2 (Table 4) and their size ranged from 250 to 3000 bp. The RAPD markers have been applied in many investigations aimed at the study of polymorphism in pomegranate, for their simplicity and low cost (Kathuria et al., 2017), but these markers have often shown low polymorphism in this species. As reported in literature it is neces­ sary an initial screening by a high number of RAPD primers to detect a discrete number of discriminating markers (Zamani et al., 2010). The level of polymor­ phism detected in our study is lower than that reported in other works (Sarkhosh et al., 2006; Zamani et al., 2007). However, a percentage of poly­ morphism similar to ours was detected by Hasnaoui et al. (2010). In agreement with these authors we hypothesized that the slightly lower percentage of polymorphism detected could be due to the reduced dimension of the sample collection and to the limited v a r i a b i l i t y i n t e r m s o f g e o g r a p h i c a l o r i g i n . Relationships among accessions were studied by clus­ ter analysis (UPGMA) based on Jaccard’s coefficient, and following statistical analysis a dendrogram was produced (Fig. 3A). The genetic distance among the accessions ranged from 0 to 0.8, showing genetic diversity among the pomegranate accessions. In the dendrogram, two main clusters could be identified: the first cluster included only two accessions (ME7 and ME8) at a 0 dissimilarity level (genetic identity). The second cluster comprised all other accessions which presented a level of dissimilarity that varied between 0 and 0.6. This last cluster presented three subgroups: ME1, ME6 and a subgroup that included accessions with genetic identity or with a very little (0.10) dissimilarity’s distance (ME2, ME3, ME4, ME5). The SSR molecular technique was utilized as a sec­ ond molecular method to discriminate the pome­ granates and to characterize their genetic profile. Although for P. granatum there is not yet a set of the best SSR primers with validity recognized at the inter­ national level, these markers have been successfully employed by several researchers to characterize the pomegranate germplasm (Beghè et al., 2016). SSRs utilized in this study have been chosen between the primers which had shown a high discriminating capacity and yielded a total number of 31 repro­ ducible fragments, which allowed a good discrimina­ Fig. 3 ­ UPGMA clusters of 8 pomegranate accessions generated by RAPD markers using the Jaccard similarity coefficient (A) and by SSR markers using Manhattan distance (B). Table 4 ­ Primer sequence of the most informative primers and level of polymorphism found by the RAPD analysis RAPDs Sequence Total n° of bands N° of polymorphic bands OPA19 5ʹ­d[CAAACGTCGG]­3ʹ 13 2 OPB08 5ʹ­d[GTCCACACGG]­3ʹ 12 1 OPC16 5ʹ­d[CACACTCCAG]­3ʹ 13 3 AH17 5ʹ­d[CAGTGGGGAG]­3ʹ 13 1 AI08 5ʹ­d[AAGCCCCCCA]­3ʹ 15 3 AI12 5ʹ­d[GACGCGAACC]­3ʹ 15 3 AI105 5ʹ­d[GTCGTAGCGG]­3ʹ 3 1 TOT 84 14 mean 12 2 Beghè et al. ‐ Characterization of ancient pomegranates in Northern Italy 589 tion of the local accessions (Table 5). Among 12 selected microsatellites, 10 showed polymorphism and 8 showed differences in the pomegranates genetic profiles. The alleles obtained by amplification of SSRs loci produced four different genetic profiles from eight ancient accessions analyzed (Table S6 of supplementary data). The number of alleles at each locus (Na) varied between 1, for loci PGKUR114 and PGKUR127, and 5, for loci ssrOeUA­PG6 and Pom021, with an average value of 2.58 and their size ranged from 155 to 319 bp. The values of expected (He) and observed (Ho) heterozygosity were always above 0.500, except for the locus PGKVR027, where was not observed heterozigosity, and obviously for the two monomorphic loci (PGKVR114 and PGKVR127). It is important to underline that four primers (PG6, Pom021, Pom045, PgAER154) were highly polymor­ phic, showing a PIC> 0.5, as defined by Botstein et al. (1980) (Table 5). These last markers showed genetic parameters (Ho, He, PIC) higher (or similar) to those reported in previous studies where they have been developed (Ebrahimi et al., 2010; Hasnaoui et al., 2012; Caliskan et al., 2017). Instead, the PGKVR­ primers presented genetic parameters lower to these reported in literature (Ravishankar et al., 2015); these primers had low polymorphism; of 6 primers 2 r e s u l t e d m o n o m o r p h i c m a r k e r s a n d o n l y 3 (PGKVR027, PGKVR064 and PGKVR165) showed dif­ ferences in the studied accessions. Similarly, the Na, He, Ho and PIC values in other studies of pomegranate varieties also varied according to the primers tested and the geographic origin of population analyzed (Caliskan et al., 2017). These results confirm the necessity to test different series of SSRs to obtain the best set of markers for each local germplasm. The UPGMA cluster based on SSR data divided the set of pomegranate accessions into two main cluster at a dissimilarity level of 45 (Fig. 3B). In the SSRs den­ drogram, the first cluster included only two acces­ sions (ME7 and ME8) with genetic identity. The sec­ ond cluster instead presented all other accessions which a level of dissimilarity that varied between 0 and 38%. This last cluster comprised three genetic subgroups: ME1, ME6 and a subgroup that included accessions with genetic identity (ME2, ME3, ME4, ME5). As confirmed by Mantel’s test (r = 0.399; p ≤ 0.033), the SSRs clustering was very similar to that performed by RAPD markers. In fact, it showed the distinction of the same four genetic groups (Fig. 3A and B). Comparison between morphological and molecular based clusters and potential use of pomegranate genetic resources It is known that RAPD fragments derived from any r e g i o n o f t h e g e n o m e a n d t h a t S S R f r a g m e n t s derived only from non­transcribed regions, therefore post­transcriptional modifications and non­nuclear inheritance of some characteristics can’t be detected by these markers (Sarkhosh et al., 2006). For these reasons, in literature there are contrasting results about the correlation between these molecular and morphological descriptors (Sarkhosh et al., 2009; Zamani et al., 2010; Basaki et al., 2013). However, researchers agreed that the combination of morpho­ logical and molecular techniques are essential for a p r o p e r a n d c o m p l e t e c h a r a c t e r i z a ti o n o f t h e germplasm of this species (Beghè et al., 2016). In this study, the RAPD and SSR analysis reflected the main morphological differences observed among the local accessions studied; the molecular cluster analyses confirmed the same two main clusters detected with PCA analysis using quantitative mor­ phological traits. Moreover, molecular analyses allowed to detect clearly four different genotypes. Analysis of correlation between distance matrices (morphological traits and molecular markers) by Mantel’s test confirmed a high statistical significance (r = 0.412; p ≤ 0.034 and r = 0.583; p ≤ 0.002 for SSRs and RAPDs, respectively). It is important to stress that in populations adapt­ ed to difficult ecological conditions, as the germo­ plasm in study, the polyphenolic content was high (Calani et al., 2013). In this previous study the ME1, ME3, ME5 and ME8 accessions were subjected to phytochemical discrimination fingerprinting in pome­ SSRs Na Size He Ho PIC PGKVR027 2 236­242 0.429 0 0.305 PGKVR064 2 239­241 0.536 0.750 0.359 PGKVR065 2 202­204 0.571 0 0.375 PGKVR114 1 258 ­ ­ ­ PGKVR127 1 246 ­ ­ ­ PGKVR165 2 307­319 0.571 1.000 0.375 POM­AGC11 2 183­185 0.536 0.250 0.359 PG4 2 198­244 0.571 1.000 0.375 PG6 5 191­199 0.786 0.750 0.653 Pom021 5 203­211 0.857 1.000 0.712 Pom045 3 155­163 0.750 0.750 0.581 PgAER154 4 262­300 0.786 1.000 0.630 TOT 31 Mean 2.58 0.533 0.542 0.394 Table 5 ­ Number of alleles (Na), Size (bp) expected (He) and observed (Ho) heterozigosity, polymorphic information content (PIC) at 12 loci in pomegranate accessions Adv. Hort. Sci., 2019 33(4): 581­592 590 granate juices. The Emilian pomegranates have pre­ sented interesting and peculiar phytochemical pro­ files. Moreover, the juices were rich in ellagitannins and had high total phenol content and total antioxi­ dant capacity, especially ME8 pomegranate. For these reasons, the local germplasm studied could be considered a source of useful traits (e.g. resistance to diseases, frost tolerance, polyphenol synthesis) for genetic improvement of this species. According to pomological descriptors and phytochemical charac­ teristics, we could appreciate peculiar features of these plants. Indeed, the ME7 and ME8 accessions showed some characteristics (e.g. small size of fruits, high woody portion in seeds, low pH) that make the fruits unlikely to be used for direct consumption, but have a juice with high antioxidant capacity, and could be successfully employed for the preparation of nutraceutical products or for industrial blending of juices. The other pomegranates (except ME6 acces­ sion that presented a high woody portion in seeds) presented good pomological characteristics for which a fresh use of the fruit could also be expected. 4. Conclusions The present work is a first contribution to the genetic and morphological characterization of the pomegranate germplasm still present in Emilia Romagna region. The morphological traits, in particu­ lar those related to fruit and seed, seven RAPDs and eight SSRs have allowed to characterise the genetic diversity of ancient pomegranate accessions. The study, although preliminary and limited to a restrict­ ed area, highlighted significant differences and inter­ esting pomological traits in local pomegranates. These results, presented in association with the study of Calani et al. (2013), clearly demonstrate a good p o t e n ti a l o f t h i s g e r m p l a s m f o r a c o m m e r c i a l exploitation as fresh or processed fruits. The acces­ sions could be used for new pomegranate plantings and could contribute to cross­breeding and the pro­ duction of new genotypes suited to marginal environ­ ments. References ADILETTA G., PETRICCIONE M., LIGUORI L., PIZZOLONGO F., ROMANO R., DI MATTEO M., 2018 ­ Study of pomo‐ logical traits and physico‐chemical quality of pome‐ granate (Punica granatum L.) genotypes grown in Italy. ­ Eur. Food Res. Technol., 244(8): 1427­1438. AGRISTAT, 2017 ­ Agriculture and livestock. ­ ISTAT, Istituto Nazionale di Statistica, via C. Balbo 16, 00184, Rome, Italy. AUDERGON J.M., 1987 ­ Elements de reflexion pour une strategie dans l’amelioration varietale des arbres fruitiers (exemple de l’abricotier). ‐ Fruits, 42(12): 725­ 734. BASAKI T., NEJAT M.A., NEJAD R.J., FARAJI S., KEYKHAEI F., 2013 ­ Identification of simple sequences repeat (SSR) informative markers associated with important traits in pomegranate (Punica granatum L.). ­ Int. J. Agron. Plant Prod., 4(3): 575­583. BEGHÈ D., FABBRI A., GANINO T., 2016 ­ Pomegranate: botany, histology and genetic resources. pp. 1­26. In: C A L I G I A N I A . ( e d . ) ­ P o m e g r a n a t e . C h e m i s t r y , P r o c e s s i n g a n d H e a l t h B e n e fi t s . ­ N o v a S c i e n c e Publishers Inc. pp. 259. B E L L I N I E . , G I O R D A N I E . , 1 9 9 8 ­ D e s c r i p t o r L i s t f o r Pomegranate (Punica granatum L.). Project on “Minor F r u i t T r e e S p e c i e s C o n s e r v a ti o n ” : ­ R E S G E N 2 9 . Horticulture Department, University of Florence, Italy. BELLINI E., GIORDANI E., GIANNELLI G., PICARDI E., 2007 ­ Le Specie legnose da Frutto. Liste dei caratteri descritti‐ vi (The Fruit Woody Species. Descriptor List). ­ Agenzia Regionale per lo Sviluppo e l’Innovazione nel settore Agricolo forestale ARSIA, Florence, Italy, pp.1070. BOTSTEIN D., WHITE R.L., SKOLNICK M., DAVIS R.W., 1980 ­ Construction of a genetic linkage map in man using restriction fragment length polymorphisms. ­ Am. J. Hum. Genet., 32(3): 314­331. CALANI L., BEGHÈ D., MENA P., DEL RIO D., BRUNI R., FABBRI A., DALL’ASTA C., GALAVERNA, G., 2013 ­ Ultra‐ HPLC–MS n (poly) phenolic profiling and chemometric analysis of juices from ancient Punica granatum L. culti‐ vars: A nontargeted approach. ‐ J. Agr. Food. Chim., 61(23): 5600­5609. ÇALIŞKAN O., BAYAZIT S., ÖKTEM M., ERGÜL A., 2017 ­ Evaluation of the genetic diversity of pomegranate accessions from Turkey using new microsatellite mark‐ ers. ‐ Turk. J. Agric. For., 41(2): 142­153. CRISTOFORI V., CARUSO D., LATINI G., DELL’AGLI M., CAMMILLI C., RUGINI E., BIGNAMI C., MULEO R., 2011 ­ Fruit quality of Italian pomegranate (Punica granatum L.) autochthonous varieties. ­ Eur. Food Res. Technol., 232: 397­403. CURRÒ S., CARUSO M., DISTEFANO G., GENTILE A., LA MALFA S., 2010 ­ New microsatellite loci for pomegra‐ nate, Punica granatum (Lythraceae). ‐ Am. J. Bot., 97(7): e58­e60. DA SILVA J.A.T., RANA T.S., NARZARY D., VERMA N., MESHRAM D.T., RANADE S.A., 2013 ­ Pomegranate biology and biotechnology: A review. ­ Sci. Hortic., 160: 85­107. DANDACHI F., HAMADEH B., YOUSSEF H., CHAHINE H., CHALAK L., 2017 ­ Diversity assessment of the Lebanese germplasm of pomegranate (Punica granatum L.) by morphological and chemical traits. ­ Ann. Agr. Sci., Beghè et al. ‐ Characterization of ancient pomegranates in Northern Italy 591 62(1): 89­98. E B R A H I M I S . , S E Y E D T . B . , S H A R I F N . B . , 2 0 1 0 ­ Microsatellite isolation and characterization in pome‐ granate (Punica granatum L.). ‐ Iran J Biothecnol., 8 (3): 156­163. FERRARA G., GIANCASPRO A., MAZZEO A., GIOVE S.L., MATARRESE A.M.S., PACUCCI C., PUNZI R., TRANI A., GAMBACORTA G., BLANCO A., GADALETA A., 2014 ­ Characterization of pomegranate (Punica granatum L.) genotypes collected in Puglia region, Southeastern Italy. ‐ Sci. Hortic., 178: 70­78. GANINO T., BEGHÈ D., ROTONDI A., FABBRI A., 2008 ­ Genetic resources of Olea europaea L. in the Bologna province (Italy): SSR analysis and identification of local germplasm. ­ Adv. Hort. Sci., 22(2): 149­155. HASNAOUI N., BUONAMICI A., SEBASTIANI F., MARS M., ZHANG D., VENDRAMIN G.G., 2012 ­ Molecular genetic diversity of Punica granatum L. (pomegranate) as revealed by microsatellite DNA markers (SSR). ­ Gene, 493(1): 105­112. HASNAOUI N., MARS M., CHIBANI J., TRIFI M., 2010 ­ Molecular polymorphisms in Tunisian pomegranate (Punica granatum L.) as revealed by RAPD fingerprints. ‐ Diversity, 2(1): 107­114. IEZZONI A.F., PRITTS M.P., 1991 ­ Applications of principal c o m p o n e n t a n a l y s i s t o h o r ti c u l t u r a l r e s e a r c h . ­ HortScience, 26(4): 334­338. I P G R I , 2 0 0 1 ­ R e g i o n a l r e p o r t C W A N A 1 9 9 9 ‐ 2 0 0 0 ­ International Plant Genetic Resources Institute, Rome, Italy. KALINOWSKI S.T., TAPER M.L, MARSHALL T.C., 2007 ­ Revising how the computer program CERVUS accom‐ modates genotyping error increases success in paterni‐ ty assignment. ­ Mol. Ecol., 16 (5): 1099­1106. KATHURIA K., BHARGAVA R., YADAV P.K., GURJAR K., 2017 ­ Molecular studies ascertaining the phylogenetic rela‐ tionships in pomegranate (Punica granatum L.) culti‐ vars using RAPD markers. ­ Int. J. Curr. Microbiol. App. Sci., 6(9): 1282­1291. KIRBY L.T., 1990 ­ DNA Fingerprinting. An introduction. ­ Stockton Press, New York, NY, USA, pp. 365. LA MALFA S., GENTILE A., DOMINA F., TRIBULATO E., 2009 ­ Primosole: A new selection from Sicilian pomegranate germplasm. ‐ Acta Hortic., 818: 125­132. LONA F., GANDINI I.M., CORRADI M.G., 1981 ­ Il verde a Parma: aspetti significativi della cultura e della tradizio‐ ne botanica a Parma. ‐ Ed. Banca Monte Parma, pp. 191 (in Italian language). MANSOUR E., BEN KHALED A., HADDAD M., ABID M., BACHAR K., FERCHICHI A., 2011 ­ Selection of pome‐ granate (Punica granatum L.) in south‐eastern Tunisia. ‐ Afr. J. Biotechnol., 10(46): 9352­9361. MANTEL N., 1967 ­ The detection of disease clustering and generalized regression approach. ­ Cancer Res., 27: 209­220. MARIESCHI M., TORELLI A., BEGHÉ D., BRUNI R., 2016 ­ Authentication of Punica granatum L.: Development of SCAR markers for the detection of 10 fruits potentially used in economically motivated adulteration. ­ Food Chem., 202: 438­444. MARTÍNEZ J.J., HERNÁNDEZ F., ABDELMAJID H., LEGUA P., MARTÍNEZ R., EL AMINE A., MELGAREJO P., 2012 ­ Physico‐chemical characterization of six pomegranate cultivars from Morocco: processing and fresh market aptitudes. ­ Sci. Hortic., 140: 100­106. MARTÍNEZ J.J., MELGAREJO P., HERNÁNDEZ F., SALAZAR D.M., MARTÍNEZ R., 2006 ­ Seed characterisation of five new pomegranate (Punica granatum L.) varieties. ‐ Sci. Hortic., 110(3): 241­246. MARTINEZ­NICOLAS J.J., MELGAREJO P., LEGUA P., GARCÍA F., HERNÁNDEZ F., 2016 ­ Genetic diversity of pome‐ granate germplasm collection from Spain determined by fruit, seed, leaf and flower characteristics. ­ Peer J., 4(7): 1­20 NEGRI V., 2003 ­ Landraces in central Italy: Where and why they are conserved and perspectives for their on‐farm conservation. ‐ Genet. Resour. Crop. Ev., 50(8): 871­ 885. NORMAN P.E., TONGOONA P., SHANAHAN P.E., 2011 ­ Determination of interrelationships among agro‐mor‐ phological traits of yams (Discorea spp.) using correla‐ tion and factor analyses. ­ J. Appl. Bios., 45: 3059­3070. RAVISHANKAR K.V., CHATURVEDI K., PUTTARAJU N., GUPTA S., PAMU S., 2015 ­ Mining and characterization of SSRs from pomegranate (Punica granatum L.) by pyrosequencing. ­ Plant Breed., 134(2): 247­254. ROSATI A., ZIPANCIC M., CAPORALI S., PADULA G. 2009 ­ Fruit weight is related to ovary weight in olive (Olea europaea L.). ‐ Sci. Hortic., 122(3): 399­403. SARKHOSH A., ZAMANI Z., FATAHI R., EBADI A., 2006 ­ RAPD markers reveal polymorphism some Iranian pomegranate (Punica granatum L.) genotypes. ­ Sci. Hortic., 111(1): 24­29. SARKHOSH A., ZAMANI Z., FATAHI R., RANJBAR H., 2009 ­ Evaluation of genetic diversity among Iranian soft‐seed pomegranate accessions by fruit characteristics and RAPD markers. ­ Sci. Hortic., 121: 313­319. SKINNER D.Z., BAUCHAN G.R., AURICHT G., HUGHES S., 1999 ­ A methods for the efficient management and utilization of large germplasm collection. ­ Crop Sci., 39: 1237­1242. UPOV, 2012 ­ Guidelines for the conduct of tests for dis‐ tinctness, uniformity and stability: Pomegranate. ‐ International Union for the Protection of New Varieties of Plants. 34, chemin des Colombettes, Geneva, Switzerland. ZAMANI Z., SARKHOSH A., FATAHI R., EBADI A., 2007 ­ Genetic relationships among pomegranate genotypes studied by fruit characteristics and RAPD markers. ­ J. Hort. Sci. Biot., 82: 11­18. ZAMANI Z., ZAREI A., FATAHI R., 2010 ­ Characterization of progenies derived from pollination of pomegranate cv. Malase‐Tourshe‐Saveh using fruit traits and RAPD mol‐ ecular marker. ­ Sci. Hortic., 124(1): 67­73. Adv. Hort. Sci., 2019 33(3): 581­592 592 ZAREI A., ZAMANI Z., FATAHI R., MOUSAVI A., SALAMI S.A., 2013 ­ A mechanical method of determining seed‐hard‐ ness in pomegranate. ­ J Crop Imp., 27(4): 444­459. ZOHARY D., SPIEGEL­ROY P., 1975 ­ Beginnings of fruit growing in the old world. ­ Science, 187(4174): 319­ 327.