Selection of superior late‐blooming almond (Prunus dulcis [Mill.] D.A.Webb) genotypes using morphological characterizations

Abstract Almond (Prunus dulcis [Mill.] D.A.Webb) is one of the earliest domesticated trees and the evidence dates back to 3000–2000 BC. In the present study, 198 almond seedling origin trees were studied to select late‐flowering genotypes having high kernel quality. Significant variabilities were exhibited among the genotypes investigated based on the recorded traits. Full‐blooming date ranged from mid‐March to mid‐April. The Ward dendrogram clustered the genotypes into two major clusters forming several subclusters. After clustering the genotypes based on the full‐blooming dates, 68 late‐blooming genotypes were recognized and reanalyzed based on the quantitative characters to select the superior ones. Nut‐related characters were as follows: nut length: 22.34–43.05 mm, nut width: 14.07–24.34 mm, nut thickness: 9.21–18.00 mm, nut weight: 1.88–6.62 g, and shell thickness: 2.26–4.59 mm. Kernel‐related characters were as follows: kernel length: 16.73–25.91 mm, kernel width: 8.50–13.64 mm, kernel thickness: 3.56–7.37 mm, and kernel weight: 0.35–1.41 g. Kernel weight was positively and significantly associated with nut weight, kernel thickness, kernel length, kernel width, nut length, and branch leaf width. Thus, these key variables are the main traits accounting for kernel weight, and they should be considered together in breeding with aiming at increasing the kernel weight. Based on ideal values of the important and commercial characters of almond, such as fruit yield, nut weight, shell hardness, kernel shape, kernel weight, and kernel taste, 19 late‐blooming genotypes were promising and are recommended for cultivation in orchards.

edible part. So almonds are unique in this respect. Also, the whole part of the almond fruit (seed, mesocarp, and endocarp) can be consumed during maturity, while in most species of the genus Prunus, the fruit can only be consumed during maturity. Almond kernels have a high nutritional value and it has a high demand in world markets, and its endocarp (shell), and mesocarp are used as animal feed (Rabinowitz, 2006).
Almonds are usually preferred in the form of shelled nut or consumed as roasted kernels. Also, it can be used as almond milk or oil (Font i Forcada et al., 2011). Almonds are also used in cooking and baking (Sang et al., 2002). Almond kernels are considered as one of the most valuable substances in terms of health and nutrition, and they are free of cholesterol. In total, 100 g of almond kernel contains 575 kcal, 12.20 g of fiber, 26 mg of vitamin E, 949 mg of total fat, 21 g of protein, 670 mg of potassium, 268 mg of magnesium, 484 mg of phosphorus, 265 mg of calcium, and 3.50 mg of iron. Its saturated fatty acid content is low, but it has other important nutrients (Blanca, 2007).
In general, breeding of other plants is easier than fruit trees.
The first step for investigating and classifying indigenous and local plant resources is the use of morphological traits (Salim et al., 2018).
Success in breeding programs depends on genetic diversity (Arslanoglu et al., 2011). Genetic diversity is used to investigate taxonomic relationships, select promising cultivars and genotypes, and introduce them for use in breeding programs as well as cultivation in orchards (Balkaya & Ergun, 2008).
One of the most important factors limiting the distribution and production of horticultural crops in the world is damage caused by spring frosts. Almond flowers open early, which results in spring frost damages (Imani & Khani, 2011). Therefore, the cultivation and production of this nut in regions with spring frosts are limited (Kodad and Socias i Company, 2004). Therefore, identification and introduction of late-blooming genotypes and cultivars are one of the most important goals in breeding programs of almonds (Kester & Asay, 1975). Iran is one of the most important almond-producing countries in the world and it has a long history of almond cultivation as well as rich almond seedling origin trees. In the present study, the seedling-originated populations were assessed to identify late-blooming almond genotypes. In the first step, the selected trees were examined in terms of flowering time. In the second step, morphological diversity of the selected genotypes was investigated using the traits related to tree, leaf, and fruit. In the third step, late-flowering genotypes were identified and then the traits related to their kernel quality were used to select the superior late-flowering genotypes.

| Plant material
A total of 198 seedling origin almond trees were selected in the Shazand region of Markazi province, Iran, and were studied for two consecutive years (2021 and 2022), with the aim of selecting superior late-flowering genotypes in terms of kernel quality. In the first step, the selected trees were examined in terms of full blooming dates (50% of flowers completely open; Sakar, Yamani, Boussakouran, & Rharrabti, 2019;. The control cultivar used was "Tardy-Nonpareil," which is a late flowering cultivar. In the second step, morphological diversity of the selected genotypes was investigated using the traits related to tree, leaf, and fruit. In the third step, late-flowering genotypes were identified and then the traits related to their kernel quality were used to select the superior lateflowering genotypes. Shazand region is located at latitude 33°57′44″N, longitude 49°25′15″E, and 1913 m height above sea level, with mean annual temperature of 13.90°C and annual precipitation of 320 mm.
The genotypes were named based on their studied areas and started with a number. The collections under study were suitable in terms of growth conditions such as irrigation, nutrition, and pest and disease control, and the trees were mature, healthy, and in full cropping.

| The characteristics evaluated
The number of 50 replicates was used to evaluate the traits related to leaf, nut, and kernel. A digital caliper was used to measure leaf, nut, and kernel dimensions. Also, an electronic scale with an accuracy of 0.01 g was used to measure the weight of nuts and kernels.
Almond descriptor (International Plant Genetic Resources Institute, IPGRI, Gulcan, 1985) was used to estimate qualitative traits and based on that traits were coded and ranked.

| Statistical analysis
Analysis of variance (ANOVA) using SAS software (SAS® Procedures, 1990) was used to determine the phenotypic variation among genotypes based on the recorded traits. Pearson correlation coefficients with Student's t-distribution test were used to determine the relationship between the recorded traits using SPSS software (SPSS Inc.;Norusis, 1998). To identify the most important influencing traits in the grouping of genotypes, principal component analysis (PCA) was applied using SPSS software. Hierarchical cluster analysis (HCA) based on Ward's method and Euclidean distance coefficients using PAST software (Hammer et al., 2001) was used to classify genotypes. The first and second principal components (PC1/PC2) were used to draw a two-dimensional plot by determining the distribution of genotypes. Multiple regression analysis (MRA) with "stepwise linear" method was used to determine the effect of independent traits on kernel weight as a dependent trait, using SPSS software.

| Characterization of the 198 genotypes investigated
Significant variabilities were exhibited among the genotypes investigated based on the recorded traits (ANOVA, p < .01). The coefficient of variation (CV) ranged from 11.25 (in kernel length) to 201.79% TA B L E 1 Statistical descriptive parameters for morphological traits used to study almond genotypes. Shell hardness was moderate in most of the genotypes (117).
One of the most important goals in almond breeding programs is shell softness. The resistance to pests and diseases is higher in hardshell genotypes than in soft-shell genotypes (Gradziel & Martinez-Gomez, 2002;Khadivi-Khub & Etemadi-Khah, 2015;Ledbetter & Shonnard, 1992). Therefore, the type of genotype used has a considerable effect on the breeding programs, so that in the case of soft-shell genotypes, identification of individuals that are resistant to fungi and insects is a priority, while in the case of semi-soft-shell genotypes, breeders seek to select genotypes having an acceptable shelling (Khadivi, Goodarzi, & Sarkhosh, 2019).
Light kernel color is preferred (Kodad et al., 2015;Socias i Company et al., 2008). A greater pubescence is associated with darker kernel color and is less desirable for nuts consumed raw (Kodad et al., 2015;Socias i Company et al., 2008).

TA B L E 2
Frequency distribution for the measured qualitative morphological characteristics in the studied almond genotypes.  (89) Green (109) Oval (125) Hearty (15) Elongated oval (56) -Shell hardness -Very hard (4) Hard (40) Moderate (117) Soft (30) Paper (7) Shell color intensity -Very light (14) Light (64) Moderate (93) Dark (27) (Kodad et al., 2015). Also, the degree of double kernel has a significant effect on crop quality and marketing (Kester & Gradziel, 1996). Various factors, such as environmental conditions, especially low temperature in the pre-flowering phase, are significantly influential in the occurrence of this negative phenomenon. To reduce this phenomenon, the plant nutrition program should be normal and genotypes sensitive to this phenomenon should be cultivated in milder areas in terms of climate (Kester & Gradziel, 1996).
A range of 0.00%-40.00% with a mean of 3.13% was recorded for the amount of blanked nuts. This phenomenon is greatly influenced by the fertility rate. Therefore, to reduce this phenomenon, great care should be taken in pollination, fertility, placement of bee hives, and selection of pollinizer in the orchards (Khadivi-Khub & Etemadi-Khah, 2015). Kernel taste was significantly variable, including bitter (38 genotypes), relatively bitter (19), sweet (110), and very sweet (31) ( Table 2).
Significant correlations were detected between some traits as revealed using Pearson correlation coefficient analysis (Table 3). Thus, the effect of independent traits on kernel weight as a dependent trait was investigated with MRA (Table 4). The MRA showed that kernel weight was positively and significantly associated with nut weight, kernel thickness, kernel length, kernel width, nut length, and branch leaf width, while it was as negatively and significantly associated with shell thickness. Thus, these key variables are the main traits accounting for kernel weight, and they should be considered together in breeding with aiming increasing kernel weight.
Significant regression associations between kernel weight and nut and kernel-related characters have been previously reported with MRA in almond (Asgari & Khadivi, 2021;Khadivi, Safdari, et al., 2019;Sepahvand et al., 2015). In addition, kernel weight and others geometrical traits were investigated by Ledbetter and Sisterson (2010) to determine best regression models that could describe variations TA B L E 4 The traits associated with kernel weight in the almond genotypes studied as revealed using multiple regression analysis and coefficients.
The scatter plot created using PC1/PC2 showed phenotypic variations among the genotypes (Figure 1). Starting from negative to positive values of PC1, the genotypes showed gradual increases in nut width, nut thickness, nut weight, shell thickness, kernel width, and kernel weight. Also, starting from negative to positive values of PC2, the characters, including tree growth vigor, canopy density, branching, branch density, showed gradual increases among the genotypes studied.

F I G U R E 1
Scatter plot for the studied almond genotypes based on PC1/PC2. The numbers represent the genotypes of each area in the plot, including 1-120 (Borj) and .

TA B L E 6
Descriptive statistics for quantitative morphological traits utilized in the late-blooming almond genotypes. The Ward dendrogram clustered the genotypes into two major clusters (not shown). The first cluster (I) included 15 genotypes.
The remaining genotypes were placed into the second cluster (II), forming four subclusters. Subcluster II-A included 49 genotypes, and subcluster II-A consisted of 37 genotypes. Subcluster II-C contained 57 genotypes, while the rest 40 genotypes formed subcluster II-D.

| Characterization of the late-blooming genotypes selected
After clustering the genotypes based on the full-blooming dates, 68 late-blooming genotypes were recognized and reanalyzed based on the quantitative characters to select the superior ones. The CV ranged from 9.75% (in kernel width) to 26.83% (in kernel weight).
Branch leaf-related characters were as follows:  (Table 6).
Kernel weight with average of approximately 1.00 g is common in many European and American cultivars, and such weight is a desired trait in breeding programs (Kester & Gradziel, 1996). Based on ideal values of the important and commercial characters of almond such as fruit yield, nut weight, shell hardness, kernel shape, kernel weight, and kernel taste, 19 late-blooming genotypes were promising and are recommended for cultivation in orchards. The pictures of kernels of the superior late-blooming almond genotypes selected are shown in Figure 2.
The PCA clustered the characters into four PCs, justifying 75.77% of the total variance ( Two-dimensional distribution of the genotypes was shown based on effective traits in the PC1 and PC2 (Figure 3). Accumulation of genotypes in one area of the plot indicated a similarity between them. The HCA based on Ward's method divided the genotypes

F I G U R E 2
The kernels of the 19 promising late-blooming almond genotypes identified.
studied into two main clusters (Figure 4). The first cluster (I) contained 28 genotypes. The rest 40 genotypes were classified into the second cluster (II), forming two subclusters.
One of the most important limiting factors for the distribution and production of almond in the world, including Iran is spring frost, which causes the loss of blossoms. In areas where there is frequent spring frost, the best way to avoid, the cultivation late-blooming almond cultivars is preferred. Considering that in ancient times in Iran, almond propagation was through seeds, therefore this nut has a considerable genetic diversity and a rich genetic resource is provided, which provides many opportunities for breeding almond for different goals, including identifying late-blooming genotypes (Asgari & Khadivi, 2021). In the present study, 68 late-blooming genotypes among the studied germplasm were selected that can be used as parents in breeding programs, and 19 genotypes among these late-blooming genotypes had high kernel quality that may be recommended for cultivation.

| CON CLUS IONS
One of the most important improvement goals in most almond production areas is to increase spring frost resistance. Here, 68 late-blooming genotypes were identified among the germplasm studied and among them, based on the ideal values of almond commercial traits, such as yield, nut weight, shell hardness, kernel weight, kernel taste, and kernel shape, 19 late-blooming genotypes,

F I G U R E 3
Scatter plot for the late-blooming almond genotypes based on PC1/PC2.
Also, due to the fact that the flowering of such genotypes is after the rainy season, the condition for the activity of pollinators such as honey bees is better, and as a result, they will be more efficient, and it will significantly help the ovule fertility rate and the reduction of blanked nuts.

ACK N OWLED G M ENTS
None.

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

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.

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.