Identification of the promising Persian walnut (Juglans regia L.) genotypes among seedling‐originated trees

Abstract Considerable genetic diversity among the native populations of Persian walnut (Juglans regia L.) provides a great opportunity to identify genotypes with valuable traits. In the present study, morphological and pomological diversity assessments of 362 walnut seedling origin genotypes were performed to identify superior genotypes. Significant differences were observed among the genotypes investigated in terms of the evaluated characters. Nut weighted from 5.53 to 19.24 g with an average of 10.67. The range of kernel weight was 1.78–9.28 g with an average of 4.83. Kernel percentage in 107 out of 362 genotypes studied was more than 50.00%. Multiple regression analysis (MRA) showed that kernel percentage was associated with kernel weight, nut weight, kernel filled, nut width, and shell thickness. Principal component analysis (PCA) showed that the characters related to nut and kernel size were correlated with the first component (PC1). Hierarchical cluster analysis (HCA) showed that the genotypes were clustered into two major clusters. Based on the most important commercial characters considered by breeders to select ideal walnuts, 15 superior genotypes were selected and are recommended for cultivation in the orchards and also can be used in breeding programs.

is morphological and pomological evaluations. One of the basic needs of germplasm management and its proper use in practical purposes of breeding is the accurate identification of genotypes (Solar et al., 2002). Morphological studies provide a guideline for selecting genotypes that are suitable for specific growth conditions. Morphological traits can be ideal for grouping walnut genotypes because they have considerable diversity and are also easy to use (Asadian & Pieber, 2005). Accordingly, the International Union for the Protection of New Varieties of Plants (UPOV, 1999) has provided guidelines for determining the differentiation, homogeneity, and stability of new cultivars in most plants (Solar & Stampar, 2011).
There is considerable diversity in the native populations of walnuts in Iran, due to propagation through seeds, high heterozygosity, and digocamy, and consequent cross-pollination Lansari et al., 2001). Significant genetic diversity provides an excellent opportunity to identify genotypes with valuable traits. Several studies have been done to identify and introduce the superior walnut genotypes (Jafari Sayadi et al., 2012;Hajnajari et al., 2012;Sarikhani et al., 2021) and also to investigate genetic diversity among populations of this species in Iran (Karimi et al., 2010;Mohsenipoor et al., 2010;Vahdati et al., 2015). In the present study, morphological and pomological TA B L E 1 Statistical descriptive parameters for morphological traits used to study walnut genotypes

| The characters evaluated
Morphological assessments were done according to the walnut descriptor (IPGRI, 1994). The range of phenotypic diversity of genotypes was evaluated using 33 characteristics ( For the traits related to nut and kernel, 30 replicates were considered, and their mean was used for analysis. Nut and kernel dimensions were measured using a digital caliper (Anyi Instrument). Also, an electronic balance with the precision of 0.01 g (Ohaus SP602 AM, Fotronic Corporation) was used to measure the weight of nut and kernel. Kernel percentage was estimated using the "kernel weight/ nut weight × 100" formula. The qualitative traits were evaluated based on the coding and scoring mentioned in the walnut descriptor (IPGRI, 1994), the list of which is shown in Table 2.
TA B L E 2 Frequency distribution for the measured qualitative morphological characters in the studied walnut genotypes

| Statistical analysis
Analysis of variance (ANOVA) was performed to evaluate the variation among the genotypes based on the traits measured using SAS software (SAS® Procedures, 1990). Simple correlations between traits were determined using Spearman correlation coefficients (SPSS Inc. Norusis, 1998). Principal component analysis (PCA) was used to investigate the relationship between genotypes F I G U R E 1 The pictures of some walnut genotypes studied pointing out diversity of nut and kernel-related characters and determine the main traits effective in genotype segregation using SPSS software. Hierarchical cluster analysis (HCA) was performed using Ward's method and Euclidean coefficient with PAST software (Hammer et al., 2001). The first and second principal components (PC1/PC2) were used to create a scatter plot with PAST software. Besides, independent traits affecting the kernel percentage as a dependent trait were determined through multiple regression analysis (MRA) using the "linear stepwise" method with SPSS software.

| Morphological and pomological description
The ANOVA (p < 0.01) revealed significant differences among the genotypes investigated in terms of the evaluated characters. The coefficient of variation (CV) ranged from 7.45 (in nut width) to 83.97% (in kernel color). The CV in 22 out of 33 characters was more than 20.00%, so that 11 characters, including kernel color, ease of kernel removal from nuts, shell seal, shell hardness, shell surface serration, shell color, kernel vein, nut shape, yield, kernel plumpness, and tree growth vigor, had the CVs more than 50.00% (Table 1).
Tree growth vigor and tree height were mainly high (133 and 172 genotypes, respectively) ( fruiting was lateral only in two genotypes. Poggetti et al. (2017) reported that all walnut trees turned out to be terminal bearing in their survey. Previous studies in different countries have reported less than 5.00% of genotypes with lateral fruiting (Atefi, 1997;Botu et al., 2010;Korac et al., 1997;Solar et al., 2002). Only Rouskas & Zakynthinos (2001) reported that Greek walnuts were more laterally bearing in their study, but the trees they studied were a panel of selected genotypes.
Thus, shell thickness in 185 out of 362 genotypes studied was lower than 1.50 mm. The ease of kernel removal from nuts is affected by shell thickness, and the quality of kernels is increased with the easier kernel removal from nuts from a commercial point of view (Sharma & Sharma, 1998).
Kernel color was predominantly light (297 genotypes). A primary object in walnut breeding programs is finding the genotypes with light kernel color (Kabiri et al., 2018). Kernel removal from nuts was easy in the majority of genotypes (251). The ease of kernel removal from nuts is an essential pomological feature. Korac et al. (1998) state that nuts with a thinner shell have a kernel easy to remove, especially if their shell is smooth, which is the most common case.
Kernel plumpness was low (103 genotypes), moderate (107) 2.36-6.64 g for kernel weight in Kazakh walnuts. The minimum of 6.00 g for kernel weight is accepted in walnut breeding programs (Germain, 1997;Korac et al., 1997;Sharma & Sharma, 1998). Thus, 53 out of 362 genotypes studied had kernels with more than 6.00 g.  Karadag & Akca, 2011). The kernel percentage is a prominent character that is of great importance to walnut breeders (Cosmulescu & Botu, 2012). The kernel percentage is influenced by the weight of nut and kernel. The genotypes with a kernel percentage above 50.00% are preferred (Germain, 1997;Korac et al., 1998). Thus, kernel percentage in 107 out of 362 genotypes studied was more than 50.00%. The diversity of nut and kernel-related characters is shown in Figure 1.

| Multiple regression analysis
The effect of independent traits on kernel percentage as a dependent trait was investigated with MRA (Table 3). The MRA showed that kernel percentage was associated with kernel weight, nut weight, kernel filled, nut width, and shell thickness. Thus, these key variables are the main traits accounting for kernel percentage, and they should be considered together in breeding with aiming increasing kernel percentage. Significant regression associations between kernel percentage and nut and kernel weights have been previously reported with MRA in walnut (Khadivi et al., 2019a;Khadivi-Khub et al., 2015a, b).

| Principal component analysis
The PCA showed that the first 11 components accounted for 68.21% of the total variance (Table 4). The characters, including nut length, nut width, nut weight, kernel length, and kernel weight, were correlated with PC1, accounting for 11.38% of the total variance, called fruit size. The PC2 included leaf length, leaf width, terminal leaflet length, and terminal leaflet width, explaining 9.96% of the total variance called leaf size. Three characters, including shell thickness, shell hardness, and ease of kernel removal from nuts, formed the PC3, accounting for 7.57% of the total variance. In the previous studies, PCA has been used to investigate the phenotypic diversity of walnut genotypes (Cosmulescu & Stefanescu, 2018;Ebrahimi et al., 2015;Ghasemi et al., 2012;Hussain et al., 2016;Khadivi et al., 2019b;Khadivi-Khub et al., 2015a, b;Norouzi et al., 2013;Rezaei et al., 2018).
The scatter plot created using PC1/PC2 showed phenotypic variations among the genotypes (not shown). Starting from negative to positive values of PC1, the genotypes showed gradual increases in nut length, nut width, nut weight, kernel length, and kernel weight.
Also, starting from negative to positive values of PC2, the characters, including leaf length, leaf width, terminal leaflet length, and terminal leaflet width, showed gradual increases among the genotypes studied.
F I G U R E 2 Biplot for the studied populations of walnut based on the morphological characters

| Hierarchical cluster analysis
The genotypes were clustered into two major clusters by the Ward dendrogram (not shown). The first cluster (I) was divided into two subclusters. Subcluster I-A included 36 genotypes, while 81 genotypes were grouped into subcluster I-B. The remaining genotypes were placed into the second cluster (II), forming three subclusters.
Also, population analysis showed that the studied areas were divided into four main groups ( Figure 2). Group I included Vashe area, while group II consisted of Khalaj and Ghale areas. Besides, group III included Shahbaz and Hesar areas, while group IV consisted of Emarat area.
The obtained results showed considerable phenotypic diversity in the Iranian walnut germplasm compared with that of other countries. This variability may be due, first, to genotypic variation or environmental conditions (Ghasemi et al., 2012). Second, walnuts species are monoecious and heterodichogamous, favoring outcrossing over selfing . Third, propagation by seeds of this species causes a significant genetic vaariability that appears at the flowering period for pomological characters, tree vigor, and type of fructification, which allow each geographic region to maintain a diverse population (Lansari et al., 2001). Morphological traits have been used effectively in detecting genetic variation on walnut in various studies (Cosmulescu & Stefanescu, 2018;Ebrahimi et al., 2015;Ghasemi et al., 2012;Hussain et al., 2016;Khadivi et al., 2019b;Khadivi-Khub et al., 2015a, b;Norouzi et al., 2013;Rezaei et al., 2018).

ACK N OWLED G M ENT
The authors would like to thank Arak University in Iran for its financial support.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

R E S E A RCH I N VO LV I N G H U M A N PA RTI CI PA NTS A N D/ O R A N I M A L S
None.

I N FO R M E D CO N S E NT
None.