Multivariate analysis of oriental apple (Malus orientalis Uglitzk.) based on phenotypic and pomological characterizations

Abstract Oriental apple (Malus orientalis Uglitzk.) is rich in valuable traits, such as later blooming, adaptation to a wider array of habitats, and capacity for longer storage of the apples, that should be considered important to the genetic makeup of the domestic apple. Here, the morphological diversity of this species was evaluated. There were significant differences among the accessions studied as revealed by the recorded traits. Ripening date ranged from late April to mid‐May. Fruit skin ground color showed strong diversity, including light cream, cream, yellow, light green, and green‐white. Also, fruit skin over color was highly variable, including white, cream, yellow, light green, and green‐white. Fruit weight ranged from 3.45 to 16.74 g. Principal component analysis (PCA) showed 13 PCs which contributed to 83.30% of total variance and fruit‐related characters were the most effective traits for separating and identifying the studied accessions. The Ward dendrogram reflected the similarities and dissimilarities among the accessions based on the qualitative and quantitative variables measured. A high phenotypic diversity within the collected material of M. orientalis was indicated. The diversity existing in the indigenous wild M. orientalis could further add new genetic information in the global gene pool of Malus species. The present study confirmed the necessity of preserving this unique genetic resource and continuing its study no matter the fact that in practice.

M. orientalis was described by a lower diversity of fruit quality, but due to the high variability in populations, it could have contributed to the domestication of apple by introgression of some traits (Buttner, 2001). Other valuable traits, such as later blooming, adaptation to a wider array of habitats, and capacity for longer storage of the apples, should be considered important to the genetic makeup of the domestic apple (Forsline et al., 2003). Zhukovsky (1965) highlighted not only the late ripening and good transportability of fruits and high sugar content but also the low winter hardiness.
Collections of apple genetic resources are relatively long-term investments that have been established for a variety of different purposes, including research, breeding, conservation, distribution, and public interest. National, subnational, or local gene banks, botanical gardens, arboreta, private companies, NGOs, and private individuals all maintain significant apple collections. Some apple gene banks have plantings of seedlings or grafted trees that represent wild species populations. These accessions offer the community access to flowers and fruits of wild species that can immediately be used in breeding programs or for phenotypic and genotypic evaluations (Bramel & Volk, 2019).
Morphological traits are the basic information for breeding programs, the management of genetic resources, the protection of cultivars, and the selection of candidates to diversify local production. Descriptors for morphological traits and identifying apple cultivars have been developed by the International Union for the Protection of New Varieties of Plants (UPOV, 2005). The objective of the present study was to assess the phenotypic diversity of M. orientalis accessions from three distinct populations in the Sistan-va-Baluchestan province from the southern part of Iran using quantitative and qualitative traits.

| Plant material
A total of 45 accessions of M. orientalis were sampled from three natural habitats of the Sistan-va-Baluchestan province from the southern part of Iran, including Maki, Rahmanabad, and Bonab.
Maki area is located at 26°58'12"N latitude, 60°52'60"E longitude, and 932 m height above sea level. Rahmanabad area is located at 26˚58'30"N latitude, 60˚53'18"E longitude, and 943 m height above sea level. Bonab area is located at 26°44'35"N latitude, 61°06'45"E longitude, and 1133 m height above sea level. The appropriate distances were considered between the accessions in each collection site to avoid the possibility of sampling and collecting clones of the selected trees.

| The characters evaluated
A total of 48 quantitative and qualitative morphological and pomological traits (Table 1) were used for phenotypic evaluations of the accessions selected. Fifty replicates per accession for leaf and fruit were used for measurements and the mean values were used for analysis. Fourteen characters, including leaf length, leaf width, petiole length, petiole width, fruit length, fruit width, fruit stalk length, fruit stalk diameter, endocarp diameter, mesocarp diameter, fruit flesh thickness, seed length, seed width, and seed thickness, were measured using a digital caliper (Model 10-754-500, Mitutoyo Cooperation). Fruit weight was measured using an electronic balance with 0.01 g precision (Model JTS-JSA, Kia Cooperation). Also, the remaining characters were qualitatively estimated based on rating and coding according to the apple descriptor (UPOV, 2005; Table 2).

| Statistical analysis
Analysis of variance (ANOVA) was performed to evaluate the variation among accessions based on the traits measured using SAS software (SAS Institute, 1990). Simple correlations between traits were determined using Pearson correlation coefficients (SPSS Inc., Norusis, 1998). Principal component analysis (PCA) was used to investigate the relationship between accessions and determine the main traits effective in accession segregation using SPSS software.
Hierarchical cluster analysis (HCA) was performed using Ward's method and Euclidean coefficient using PAST software (Hammer et al., 2001). The first and second principal components (PC1/PC2) were used to create a scatter plot with PAST software.

| RE SULTS AND D ISCUSS I ON
There were significant differences among the accessions studied as revealed by the recorded traits. Thirty-five out of 48 characters measured showed a CV of more than 20.00%. The highest CV belonged to fruit symmetry (190.91%), in agreement with Khadivi et al. (2020) who observed these findings in M. orientalis from the Isfahan province, Iran. Also, the CV was more than 50.00% in petiole color (72.61%), branch skin color (63.04%), color of spots on fruit skin (56.52%), trunk type (55.46%), the tendency to form suckers (54.83%), seed presence (54.17%), leaf serration depth (53.17%), and ripening date (52.45%). The lowest CV belonged to leaf length (11.42%), fruit stalk diameter (11.46%), leaf shape index (12.23%), and seed length (12.83%) ( Table 1). Five characters were the same and stable among the accessions, including leaf margin dentation (present), leaf serration shape (serrate), petiole cross-section (round), fruit flesh color (white), and seed color (brown), in agreement with Khadivi et al. (2020) who observed these findings in M.
orientalis from the Isfahan province, Iran.

TA B L E 3 (Continued)
Principal component analysis (PCA) was used to identify the patterns of variability among the accessions studied. For PCA, components with eigenvalues more than 1.00 were retained to uphold reliability of the final output. Thus, 13 PCs were observed which contributed 83.30% of total variance and the first three PCs explained 32.76% of the variance (Table 3). Khadivi et al. (2020)  The PCA has been widely used to analyze the phenotypic diversity of apple (Forsline, 2000;Hofer et al., 2013;Khadivi et al., 2020;Luby et al., 2001;Reig et al., 2015).
The scatter plot prepared according to the PC1 and PC2 (25.09% of total variance) reflected the relationship among the accessions in terms of phenotypic characteristics. The plot distributed accessions into four sides and showed significant differences for some traits ( Figure 2). Also, the Ward dendrogram reflected the similarities and dissimilarities among the accessions based on the qualitative and quantitative variables measured. The most significant result was represented by the identification of two separate main clusters of the accessions based on morphological characteristics (Figure 3).
Cluster I included 17 accessions, while cluster II included 28 accessions forming two subclusters.
The results showed that the studied germplasm have high phenotypic variation. Similarly, high phenotypic variability was reported in M. orientalis collections from different countries (Aldwinckle et al., 2002;Ercisli et al., 2004;Hofer et al., 2013;Khadivi et al., 2020).
These studies indicated that high diversity in pomological and leafrelated traits could be used as an efficient marker system to discriminate between the apple accessions.
M. orientalis has been reported as one of the possible ancestors  (Ponomarenko, 1992) and also with resistance to fire blight (Erwinia amylovora), apple scab, and cedar-apple rust (Gymnosporangium juniperi-virginianae) have been revealed (Volk et al., 2008(Volk et al., , 2009  characterization is always needed and should be included in any program of conservation and use of genetic resources (Aldwinckle et al., 2002;Ercisli et al., 2004;Hofer et al., 2013).

| CON CLUS ION
The

CO N FLI C T O F I NTE R E S T
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.