Characterization of frost-tolerant plum genotypes (Prunus domestica L.) in Nishabur region, Iran: a morphological and phenological assessment following natural spring frost event

Abstract

Background

The Nishabur region in Iran is an ancient hub for plum production, home to numerous seedling orchards and indigenous plum varieties. In 2020, an evaluation was conducted in the primary plum-growing zones of Nishabur following a harsh spring frost. Forty-one plum genotypes and local varieties capable of bearing fruit after frost incidents were selected for further examination. These plum selections were evaluated based on 60 morphological, pomological, and phenological traits related to flowers, fruits, and trees, in accordance with the UPOV (2020) plum descriptor.

Results

Among the 41 genotypes evaluated, 35 exhibited high yields, demonstrating their potential as viable options for cultivation in frost-prone areas. The highest coefficient of variation (39.45%) was observed in the fruit color. Several genotypes demonstrated acceptable pomological traits. The genotype ‘Kh.Da.cv.04‘ has the highest values in terms of fruit weight (56.2 g), fruit width (46 mm), and fruit length (61 mm). Significant positive correlations were found between fruit length and leaf length and fruit weight and leaf blade length. Factor analysis revealed that fruit weight, size, and leaf length are the most influential factors, accounting for 52% of the total variance. In the principal component analysis (PCA), genotypes were clustered into four main groups, with ‘Kh.Da.s.02’ and ‘Kh.Da.cv.04’ positioned at the positive end of the second axis, separate from the other genotypes. Cluster analysis indicated that these genotypes, along with ‘Ha.Bokh.cv.30’, formed distinct clusters, considerably distant from the other genotypes, which were grouped into four main clusters.

Conclusions

This research highlights the presence of frost-tolerant genotypes and varieties with suitable environmental adaptability in the Nishabur region, demonstrating relatively acceptable diversity. These genotypes hold potential for breeding and germplasm conservation purposes.

Background

Plum, belonging to the diverse species in Prunus under the subfamily Prunoideae and the family Rosaceae, is considered one of the most important stone fruit crops across temperate regions worldwide [1]. While it encompasses various fruit-bearing species, commercially cultivated plum varieties primarily consist of two species: European plum, P. domestica L, which is hexaploid with a chromosome count of 48 (2n = 6x = 48), and Japanese plum, P. salicina L, which is diploid with a chromosome count of 16 (2n = 2x = 16). The European plum group is native to areas between the Black Sea and Caspian Sea regions neighboring Iran and parts of Asia Minor. On the other hand, the Japanese plum group originated in China but was subsequently domesticated in Japan before spreading to different parts of the world [1,2,3].

Plums are recognized for their significant contributions to human health and nutrition, primarily due to their rich content in B vitamins, potassium, magnesium, and essential minerals [4]. They are particularly valued for their antioxidant properties, as well as their high levels of phenolics and flavonoids, which provide numerous health benefits, including support for heart health and improved digestion [4, 5]. According to 2022 FAO data, global plum production reached approximately 12 million tons on 2.60 million hectares, while Iran produced around 332,231 tons on 1275 hectares [6]. In addition to fresh consumption, plums are processed into various products, including jams, juices, compotes, and brandies, reflecting their versatility [4].

The European plum tree boasts a wide array of varieties in colors such as red, yellow, purple, and more, featuring fruits that are round, oval, egg-shaped, and elongated, known for freestone and clingstone [7, 8]. Genetic resources from ancient plum orchards are highly prized in plant breeding. However, these valuable resources are facing a continuous decline and risk of genetic erosion as a limited number of new cultivars are replacing them. In the extensive pomological groups of plums, it has been estimated that there are over 6000 varieties of different species spread across Asia, Europe, and America, as reported in the early 20th century [9]. Studies by HW Fogle  [10] indicate that in the early 1970s, over 2800 plum varieties of different types were utilized globally. However, currently, only a few of these varieties hold economic significance for the global plum industry [11]. The gradual disappearance of traditional old varieties and the narrowing of the genetic base weaken the breeding potential and increase vulnerability to biotic and abiotic stresses. This scenario also raises concerns about genetic erosion within plum varieties [11].

The region of Nishabur and ZebarKhan in the central plateau of Iran, boasting around 4000 hectares of ancient plum orchards scattered across various villages, is a key hub for plum production in Iran [12]. Traditional orchards in this area predominantly feature seedling trees (un-grafted) or Indigenous plum varieties [13]. Consequently, this region stands out as a prime location for researching the genetic diversity of plum species. On the other hand, spring frost is indeed a significant challenge for stone fruit trees, such as plums. Breeding objectives in regions affected by this issue often focus on identifying genotypes that are resilient to low temperatures during blooming, late-blooming varieties, and those with extended flowering periods [14]. Assessing the performance of plum varieties under spring frost events in a specific regional climate zone over consecutive years is a suitable criterion for selecting cold-tolerant genotypes. In line with this, following a natural spring frost, we assessed the plums in the Nishabur region and identified and evaluated resilient genotypes and varieties.

The sensitivity of plum flower buds to frost varies significantly depending on their developmental stages [15]. As buds progress from the swelling stage to the small fruitlet stage, they become more vulnerable to low temperatures. Notably, plum flower buds exhibit less resilience to cold than those of some stone fruit species [16], and this sensitivity becomes even more pronounced due to fluctuations in winter temperatures, which can further diminish their cold hardiness. During the flowering stages, the pistil becomes the most susceptible part of the flower, particularly the stigma, which is highly vulnerable to damage from spring frosts [15]. The visible consequences of such frost damage on reproductive structures can lead to flower and fruit drop, along with morphological abnormalities in fully developed fruit. The initial signs of frost damage in plum fruits may appear as the separation of pulp from the peel [17]. This detachment can indicate the onset of internal damage, which might manifest as cracks in the seed. If the epidermis remains intact, these internal injuries could have minor long-term implications. Under favorable environmental conditions, such as moderate temperatures and adequate moisture, the undeveloped fruit may be able to recover relatively quickly. However, external damage typically becomes evident within a few days following frost exposure. Affected fruits may show pronounced signs of swelling or browning [17]. Plum growers typically mitigate frost impacts by choosing late-blooming plum varieties and employing measures such as wind machines, heaters, or orchard management practices that involve soil or water management to protect against frost damage [18]. In the long term, developing new cultivars with enhanced cold tolerance through breeding programs that focus on the genetic basis of frost resilience is crucial.

Critical temperatures that cause damage to reproductive organs can differ widely among species and cultivars in stone fruit trees [14]. Variability among cultivars often exceeds the differences observed between species [14]. However, separating genetic factors from other influencing variables presents a significant challenge. Research has shown that critical temperatures not only vary with the phenological stage of flower buds but are also influenced by microclimate, tree age, nutritional status, and can even differ within individual trees [19]. This variability indicates that the critical temperatures for specific species or cultivars should not be treated as fixed biological constants [19]. Nevertheless, genetic differences have always been a focal point for fruit tree growers and breeders, as these differences can significantly affect cold resistance and enhance the environmental adaptability of trees [14, 15].

The primary objective of this study is to identify plum genotypes and indigenous varieties that exhibit tolerance to spring frost in the Nishabur region of Iran, with a secondary focus on examining the morphological and pomological diversity among selected plum genotypes and varieties. This research can inform future breeding programs and enhance germplasm conservation efforts in agricultural sciences.

Materials and methods

Location and plant materials

To evaluate and identify the genotypes and varieties of plums that are tolerant to frost, a field evaluation was conducted in the main plum-production areas of Zebarkhan, Nishabur, Iran, after the spring frost event in 2020 (Fig. 1). Nishabur is situated in the northeastern part of Iran, within Khorasan Razavi Province. It features a semi-arid climate characterized by hot summers and cold winters. In the Nishabur region, spring frosts typically occur from April 6 to May 6. An analysis of frost events from 1989 to 2011 indicates that Nishabur is among the areas in northeastern Iran with a high frequency of frost occurrences [20]. On average, from 2016 to 2020, there have been over five significant incidents of spring frost damage in the region [20]. The Zebarkhan, Nishabur district covers an area of 1,139 square kilometers, predominantly in its eastern and southeastern regions, and includes 103 villages as well as the cities of Darood, Ghadamgah, and Kharv. The orchards of Zebarkhan are mostly owned by small owners and the trees include a mixture of grafted local varieties as well as seedling trees (Fig. 1). Most of the famous local varieties in this area are Alu-Bokhara, Kobraee, Shabloon, Qatre-tala, Shams, Targhabeh and Darghazi. On April 1, 2021, while the plum trees were in full bloom and beginning to fruit set, a cold air mass swept through the region. The temperature dipped to -5 °C (as reported by the Nishabur synoptic station), and the trees endured sub-zero conditions for six hours. As a result, local experts and insurance companies reported complete (100%) frost damage to the plum orchards. Two months later, during field visits, plum trees that bore fruit were selected, and each identified tree was tagged, labeled, and given a unique code (Table 1). Their geographic coordinates and elevations were recorded using a GPS device, leading to the selection of 41 plum genotypes from the area (Table 1).

Fig. 1

Map of Zebarkhan, Nishabur, Iran, highlighting the main plum production areas and ancient plum orchards

Table 1 Location, altitude, genotype code and source names of selected plum after natural spring frost from Zebarkhan, Nishabur, Iran

Morphological and pomological evaluations

The selected Plum genotypes were assessed using 60 morphological and phenological traits (Table 2). These traits were evaluated in accordance with the plum descriptor (UPOV 2020). For the analysis of leaf and fruit quantitative traits, 10 samples of fully developed and ripe fruits were randomly collected from around the tree in each genotype. Weight-related traits were measured using digital scales with a precision of 0.01, while length-related traits were assessed with digital vernier’s calipers. Phenological traits, including the timing of fruit ripening, were assessed in the same year, while traits related to buds and flowers were evaluated in the following year for each selected tree.

Table 2 Morphological, phenological, and pomological traits used to evaluate plum genotypes in the Neyshabour region

Statistical analysis

Statistical parameters, including minimum, maximum, mean, standard deviation (SD), and coefficient of variation (CV), were calculated using RStudio 1.0.136 software. The frequency distribution for the measured qualitative morphological traits was determined. For correlation analysis, the cor function was used to calculate correlation coefficients, applying Pearson’s method for quantitative traits and Spearman’s method for qualitative traits. The relationships among progenies were analyzed through principal component analysis (PCA). Additionally, scatter plots were generated using the first and second principal components (PC1/PC2) in RStudio. Cluster analysis of genotypes was performed based on Ward’s method, and heat map analysis was utilized to visualize distributions and distances among samples, employing the ggplot2 and d3heatmap packages within RStudio 1.0.136.

Results and discussion

Descriptor analysis

The results of descriptive analysis of morphological traits among 41 plum genotypes are reported in Table 3. Traits that have a high coefficient of variation have a wider range of variability of the trait and more selection possibility for those traits in breeding programs. Out of 60 traits measured, 32 traits had a coefficient of variation above 30%, which indicates high diversity in 41 plum genotypes, which includes pomologically important traits related to fruit shape (FST, FDSC, FShLV, and FShVV), fruit weight (FW), traits related to fruit color (FHOC and FCS), ratio of stone weight to fruit flesh (FSWR), adherence of stone to flesh (ASF) as examples. The highest coefficient of variation is related to the hue of over color in fruit (FHOC) (45.39%) and the lowest is related to the ratio lateral width to ventral width of the fruit (FLVW) (8.18%). The FW is indeed a crucial pomological trait, influenced by a combination of factors including nutritional conditions, the number of fruits on the tree (thinning), and the position of buds on branches [7]. However, as pointed out, genotype plays a significant role and can lead to notable differences among varieties [8]. FW with an average of 30.62 g varies from 12.68 g to 56.9 g. The FW has a variance of 108.52 and a coefficient of variation of 34.02%. The highest FW is in the genotype coded ‘Kh.Da.cv.04’ and the lowest is related to the genotype ‘Ar.GH.Cv.37’. The highest FW for Kashmir plums were reported in two cultivars, Friar and Santa Rosa, at 50.47 g and 50.23 g, respectively, while the lowest was observed in the Chatsuma cultivar, which weighed 38 g [21]. In the Kelkit Valley, Türkiye, the outstanding genotype E9 was reported to have a FW of 25.92 g [3]. Additionally, Ö Beyhan [22] reported plum FW in Turkey ranging from 12.63 g to 29.17 g, while M Arvas [23] documented weights from 7.58 g to 52.22 g. A Küden, N Kaska, A Özgüven and A Küden [24] reported weights from 31.81 g to 50.47 g. E Ganji Moghaddama, S Hossein Avab, S Akhavanc and S Hosseinid [25] investigated plum genotypes in Iran and reported the weight of plum fruit to be 23 to 50 g. JH Kwon, JH Jun, EY Nam, KH Chung, IK Yoon, SK Yun and SJ Kim [26] reported the average weight of Asian plums studied in South Korea with an average of 77.2 g ranging from 32.4 to 139 g. In general, the FW of Asian plums is higher than that of European plums [2]. The plum varieties investigated in this research are of the European plum type.

Table 3 Descriptive analysis of 60 traits measured in 41 plum genotypes tolerance to spring frost in Neishabur region

The leaf length (LL) was variable between 55 and 100 mm with an average of 75.6 mm and a variance of 103.69. The leaf width (LW) was varied at least 23 to 51 mm with an average of 39.75 and a variance of 36.23 (Table 3). The shape of the petals is oval, broad oval, and round with a coefficient of variation of 38.06 (Table 3; Fig. 2).

Fig. 2

Color variation, fruit size, and floral diversity in plum genotypes tolerant to spring cold in Nishabur, Iran: The upper image showcases the diversity in fruit color and size, while the lower image highlights the arrangement of sepals and petals in the flowers

Qualitative trait frequency

The frequency values of qualitative traits in 41 plum genotypes are detailed in Table 4, which includes 43 quantitative traits. Over 68% of the evaluated genotypes were classified as very high-yield trees, with 51.2% considered vigorous and 17.2% extremely vigorous. Additionally, more than 70% exhibited a spreading growth habit. The leaf blade shape (LGCo) was primarily oval, occurring in 60% of the samples, while the angle of the leaf tip (LAA) was acute in 50% of the samples. Most genotypes displayed an oval fruit shape, and fruit symmetry was prevalent across the varieties. The fruit skin color (FCS) varied from yellow to red, with dark blue being the most common color, accounting for 40% of the genotypes (Table 4; Fig. 2). Dark blue is a very common fruit skin color in P. domestica but very rare in P. salicina [2]. Many seedling (non-grafted) plum genotypes studied in this study had large and suitable fruits (Fig. 2). The flesh color of selected plum genotypes varied from yellow to red. M Vizzotto, L Cisneros-Zevallos, DH Byrne, DW Ramming and WR Okie [27] reported that red-fleshed plums generally have higher anthocyanin and phenolic content compared to those with whitish or yellow flesh. It’s noteworthy that high levels of phenolic compounds are often strongly associated with astringency [5]. Some genotypes exhibited clingstone characteristics; however, most varieties were freestone. Freestone is a quantitative trait, but several reports indicate that freestone is genetically monogenic and dominant [1]. Freestone is considered an advantage for consumers and for drying prunes. Among the measured genotypes, the tree height varied from 3 to 8 m. The petal arrangement (Fp) was free in the majority of genotypes. Notably, 40% of the genotypes exhibited late flowering, while 10% were classified as very late flowering (Table 4). The high degree of phenotypic variability observed suggests that targeted selection can effectively exploit this diversity to produce superior plum cultivars tailored to specific growing conditions and consumer preferences.

Table 4 Frequency (%) of quality traits evaluated in 41 spring frost resistant plum genotypes in the Nishapur region, Iran

Correlation analysis

In this study, we conducted a correlation analysis of quantitative traits measured in 41 plum genotypes exhibiting spring frost tolerance. The analysis revealed significant positive and negative correlations among several traits at the one and five% significance levels, which are crucial for informing breeding improvement programs. It is important to note that these correlations indicate relationships between traits rather than direct causal effects, highlighting their potential utility for indirect measurements [28]. As demonstrated in Fig. 3, the Pearson correlation coefficients illustrate that many traits exhibit linear relationships. Notably, significant positive correlations were found between LL and LW (r = 0.74), fruit length (FH) and FWVV (r = 0.73), and the ratio of fruit weight to stone weight (FSWR) with FH (r = 0.50). Additionally, FWVV showed positive correlations with LL (r = 0.34) and Fruit: width in lateral view (FWLV) (r = 0.83). These findings suggest that certain traits, such as leaf blade length, correspond with increases in leaf blade width and fruit dimensions. Conversely, significant negative correlations were observed between the leaf length-to-width ratio (LWR) and LW. Figure 4 presents Spearman’s correlation coefficients for qualitative traits. Several traits exhibited strong correlations; for instance, the relationship between fruit symmetry in the ventral view (FsVV) and fruit shape in the ventral view (FShVV) showed a high correlation coefficient (r = 0.55). The ground color of fruit skin (FCS) had notable correlations with fruit hue of over color (FHOC), fruit intensity of over color (FiOC), and fruit ground color of skin (FCF), displaying coefficients of 0.46, 0.45, and − 0.46, respectively. Additionally, the relationship between fruit firmness of flesh (FfiF) and fruit juiciness (FTF) exhibited a very strong positive correlation (r = 0.57). Furthermore, the adherence of stone to flesh (ASF) showed a moderate positive correlation with fruit hue of over color (r = 0.46). Meanwhile, the stone shape in the lateral view (Sslv) and stone shape in the ventral view (SShvV) revealed a strong negative correlation (r = -0.47). Lastly, the vegetative bud support on one-year-old shoots (Dvb) and shape of vegetative bud (Shvb) displayed a strong positive correlation (r = 0.58). Overall, these traits can be utilized in genetic analyses and may provide insights into the causal relationships that exist between various traits in plum genotypes.

Fig. 3

Heat plot displaying the Pearson correlation coefficients between quantitative traits measured in 41 plum genotypes that exhibit spring frost tolerance. For complete trait descriptions, refer to Table 2

Fig. 4

Heat plot showing the Spearman’s correlation coefficients for qualitative traits measured in 41 plum genotypes that exhibit spring frost tolerance. For complete trait descriptions, refer to Table 2

Principal component analysis (PCA)

Given the extensive data related to the traits obtained from the morphological evaluation of the plum plant, drawing clear and straightforward conclusions using univariate analysis is challenging. To address this, PCA analysis can be employed to organize the various traits into components, each representing several attributes. PCA has been utilized to derive more insightful results from the data. Additionally, the principal components (PC) are instrumental in examining the population structure and evaluating individuals within that population. This analysis identifies the primary traits influencing the differentiation among genotypes.

In the PCA of the measured traits, the first ten PC accounted for 95.16% of the variance. As illustrated in Fig. 5, the first PC explained the highest percentage of variance at 52.81%, while the second PC followed, accounting for 18.23%. The combined explanatory percentages of these two PC total 71.04%, indicating that they collectively explain more than 70% of the variation in the data. This significant percentage underscores the importance of the first two factors in interpreting the results (see Fig. 5).

Fig. 5

Chart of eigenvalues in the principal component analysis of 60 morphological and phenological traits measured in 41 plum genotypes exhibiting spring frost tolerance. The red dashed line represents the expected eigenvalues under the broken stick model

Table 5 presents the coefficients of the effective traits identified through PCA. These coefficients reflect the extent to which each trait contributes to the first three PC of the analysis. Based on the data in Table 4, we can ascertain which traits are most influential in this research. For the PC1, which accounts for 52% of the variance, the traits with the highest coefficients are as follows: FW (0.558), fruit-to-stone weight ratio (0.504), LL (0.390), FH (0.319), and LW (0.174). In PC 2, the most significant traits include LL (0.787), LW (0.422), fruit-to-stone weight ratio (-0.687), and FW (-0.328). For PC 3, the traits with the highest coefficients consist of FH (0.501), fruit-to-stone weight ratio (-0.687), and fruit width in the ventral view (0.305). It is important to note that the traits in each list are organized from highest to lowest based on the absolute value of their coefficients. PCA simplifies complex, high-dimensional data into more manageable, low-dimensional data. This method highlights significant features within the dataset. For instance, the first PC 1 emphasizes the importance of FW and the ratio of FW to stone. Likewise, in the second PC 2, LL and LW emerge as key traits. Overall, PCA simplifies complex and multidimensional data effectively, enabling researchers to identify crucial features for future analysis and planning effortlessly.

Table 5 Coefficients of effective traits in the first three factors of factor analysis. Significant values above 0.3

Biplot analysis

In this study, we utilized biplot analysis in conjunction with PCA to assess the genetic relationships among plum genotypes. The separation of genotypes into specific areas of the plot indicates their genetic similarity, with genotypes located close to one another exhibiting greater similarity in traits associated with the first and second PCs. The PCA analysis demonstrated that the first and second PCs explained a total of 71% of the overall variance in the plum genotypes (Fig. 5). This graphical representation visualizes the relationships between plum genotypes and the most important measured traits in a single plot.

In the biplot analysis, genotypes were distributed across four distinct regions (Figs. 4, 5, 6 and 7). In the PC1, the genotype ‘Ar.GH.Cv.37’ was situated in the end positive region of the axis, indicative of its larger fruit size. The genotypes ‘Kh.Da.S.02’ and ‘Kh.Da.cv.04’ were positioned at the positive end of the PC2, distinct from other genotypes. Additionally, genotype ‘Ka.kob.Cv.16’ was located separately in zone IV, characterized by its smaller fruits and leaves (Fig. 6). The remaining genotypes were categorized into four groups, as illustrated in Fig. 6. The first group comprises genotypes ‘Ka.kob.cv.31’, ‘Boz.Bokh.cv.40’, ‘Kh.BOKH.S.07’, and ‘Ha.kob.cv.21’, originating from the Kariznov, Kharv, and Hajji Abad and Bojmehran regions of Zebarkhan, Nishabur, Iran. The second group consists of genotypes ‘B.Kob.cv.27’, ‘Ar.Haj.cv.42’, ‘Kh.Da.s.03’, ‘B.Haj.cv.14’, ‘B.Haj.cv.24’, ‘Kh.Bokh.cv.10’, and ‘Ha.Bokh.cv.30’, found in Ardoghsh, Borj, Kharv, and Hajji Abad. The third group includes genotypes from Kariznov, Bojmehran, Hajji Abad, and Ardoghsh: ‘Ka.Haj.cv.26’, ‘Boz.kob.cv.38’, ‘Kh.Haj.cv.13’, ‘Ha.SH.cv.35’, ‘Kh.Da.s.01’, and ‘Ar.Sh.cv.18’. Finally, the fourth group encompasses the genotypes ‘Ar.Kob.cv.34’, ‘Boz.Haj.cv.32’, ‘Boz.Gh.cv.19’, ‘Kh.Bokh.cv.12’, ‘B.Bokh.cv.39’, ‘Kh.Bokh.cv.09’, ‘Kh.Bokh.s.06’, ‘Ha.Haj.cv.15’, ‘Ka.Bokh.cv.22’, ‘B.Kob.cv.20’, ‘Kh.Bokh.s.05’, and ‘Kh.Bokh.cv.11’, originating from the Bojmehran, Kariznov, Borj, Kharv, Haji-Abad, and Ardoghsh regions. The delineation of regions within the biplot enhances the ability to target specific genotypes for further investigation. Overall, the integration of PCA with biplot analysis serves as a powerful tool in evaluating the genetic diversity among plum genotypes, paving the way for improved breeding strategies and the enhancement of fruit quality.

Fig. 6

Biplot analysis of 41 frost-tolerant plum genotypes based on the first two principal components, considering 60 morphological and phenological traits in the Neishabur Region, Iran

Fig. 7

Cluster analysis diagram of 41 frost-tolerant plum genotypes based on 60 selected morphological and phenological traits from the Neishabur region, Iran. The numbers on the clades refer to the bootstrap values

Cluster analysis

This study employed cluster analysis to group different plum genotypes, utilizing all measured traits to assess similarities and relationship distances in selected plums. This analysis revealed four main groups (Fig. 7). Notably, genotypes ‘Kh.Da.S.02’ and ‘Kh.Da.Cv.04‘ formed a separate cluster, significantly distanced from the others. Genotype Ha.Bokh.Cv.30 was also identified in its own distinct group. The remaining genotypes were categorized into four primary groups, comprising 8, 10, 7, and 13 genotypes, respectively. The first group included genotypes ranging from ‘Boz.Bokh.cv.40‘ to ‘Ha.kob.cv.21‘ and largely corresponded to the group I in the biplot analysis (Fig. 6), characterized by large leaves and lower FW. The second group, spanning from ‘Kh.Bokh.cv.12‘ to ‘Boz.Haj.cv.32‘, mostly aligned with group IV in the biplot analysis, displaying smaller leaves and fruits. The third group consisted of genotypes from ‘Kh.Bokh.cv.11‘ to ‘Ka.Kob.cv.16‘, most of which were also placed in the fourth cluster of the biplot analysis. Finally, the fourth group included genotypes from ‘Kh.Da.S.01‘ to ‘B.Haj.cv.14‘, which exhibited a more pronounced separation from the second and third groups. This group was notable for producing fruits with higher weights and a greater flesh-to-stone weight ratio compared to the other plum genotypes.

The cluster analysis enhances our understanding of genetic diversity among the plum genotypes by clearly defining distinct groups based on phenotypic traits. The separation of genotypes such as ‘Kh.Da.S.02’ and ‘Kh.Da.Cv.04’ indicates the potential for these genotypes to possess unique advantages, perhaps making them prime candidates for breeding programs aimed at specific qualities. Furthermore, the alignment of groups identified in both the biplot and cluster analyses underscores the consistency of the traits observed, supporting the reliability of the analytical approaches utilized. This dual application provides a robust framework for further exploration into the genetic attributes of plum genotypes and highlights the importance of selected traits in breeding and cultivation strategies.

Heatmap analysis

Heatmap analysis is an effective data visualization technique that allows for the graphical comparison of differences among genotypes and their traits. This tool is instrumental in identifying genotypes with significant and beneficial characteristics, as well as in examining variation among them. In Fig. 8, each row represents a specific trait, while each column corresponds to a particular plum genotype. The color intensity in each section of the diagram indicates the level of genetic variation for each trait, enabling the analysis of genetic diversity based on color and intensity. The traits are classified into three main groups: A, B, and C, with a total of six sub-groups. The genotypes are divided into four main groups (I – IV), excluding genotypes labeled ‘Ha.Haj.cv.15‘, ‘Ha.Bokh.cv.30‘, ‘Ha.Sh.cv.35‘, and ‘B.Haj.cv.24‘. Group A: Subgroup A1 shows the highest trait values in group I genotypes and the lowest in genotypes ‘B.Haj.cv.24‘, ‘Ha.Bokh.cv.30‘, ‘Ha.Sh.cv.35‘, and group III. Subgroup A2, which includes pomological characteristics like FW, exhibits the highest values in group II genotypes and the lowest in group IV genotypes. Group B: In trait group B1, the lowest values are found in group I and somewhat in group IV. In group B2, traits in groups I and IV report the highest values, while group II exhibits the lowest. Group C: Group C1 displays the highest values in group IV, with the lowest in group II. Group C2 shows medium to low values mainly in group II. This analysis provides valuable insights into the genetic diversity and characteristics of the assessed plum genotypes.

Fig. 8

Heatmap analysis of 41 frost-tolerant plum genotypes from Nishabur, Iran, based on 60 morphological and phenological traits. Color intensity ranges from yellow (low) to dark violet (high). The horizontal dendrogram represents the plum genotypes, categorized into four groups (I to IV), while the vertical dendrogram categorizes the traits into three main groups (A, B and C) with corresponding subgroups. For complete trait descriptions, refer to Table 2

Superior genotypes from frost-tolerant plums

Based on the genetic distance among the examined genotypes and their pomological characteristics, we try to introduce four superior genotypes (‘Kh.Da.cv.04‘, ‘Kh.Da.s.02‘, ‘Kh.Bokh.s.07, and ‘B.Haj.cv.14‘) from 41 frost-tolerant plums (Table 6). The genotype ‘Kh.Da.cv.04‘ emerged as a standout with its classification as a free-stone fruit that ripens and flowers late. It demonstrated significant genetic distance from the other genotypes based on pomological traits (Figs. 6 and 7). This genotype produces very large fruits, boasting a very high yield with a remarkable average FW of 56.9 g. Furthermore, its excellent frost tolerance indicates its suitability for cultivation in regions susceptible to low temperatures. Another promising genotype, ‘Kh.Da.s.02‘, is categorized as a semi-free stone, with similar late ripening and flowering characteristics. This variant also yields very large fruits and shows a considerable genetic distance from the other genotypes (Figs. 5 and 7), with an impressive average weight of 52.6 g and a very high yield. Its frost tolerance is rated as very high, making it an excellent candidate for frost-prone areas. The ‘Kh.Bokh.s.07‘ genotype, also semi-free stone, differs slightly in its flowering time, with medium flowering and ripening periods. Its fruit size is moderate, with an average FW of 22.4 g, but it still achieves a high yield. Despite its comparatively smaller size, it retains very high frost tolerance, positioning it as a viable option under cold conditions. Lastly, the genotype ‘B.Haj.cv.14‘, which is a semi-free stone, exhibits a late flowering time and a very late ripening period. It produces medium-sized fruits weighing approximately 35.66 g and shows a very high yield. Its frost tolerance is also rated high, suggesting strong resilience in colder climates. Overall, these genotypes represent a significant advance in plum cultivation, providing valuable options for growers as well as for germplasm collection for further evaluations.

Table 6 Characterization of selected superior genotypes from frost-tolerant plums

Conclusion

In the Nishabur region of Iran, which is a key area for plum cultivation, spring frosts have historically caused significant damage to local farmers. Given the diverse local plum varieties and the presence of non-grafted plum trees in Nishabur, coupled with a tradition of ancient farming practices, there exists considerable potential to discover plums that exhibit frost tolerance. This research investigated the productivity of plum trees from six villages in Nishabur, Iran in the aftermath of severe spring frost, ultimately identifying 41 trees that successfully fruited. We assessed these select 41 frost-tolerant plum genotypes for morphological characteristics, focusing on their productivity and tolerance to spring frost. The study uncovered significant morphological and pomological variability among the genotypes, with notable variation observed in key traits such as FW, skin color, flowering time, and tree growth characteristics. Our analysis revealed important relationships between traits that can inform future breeding practices. The application of PCA and cluster analysis identified distinct groupings among the genotypes, highlighting both genetic diversity and phenotypic similarities. Additionally, heatmap analysis provided visual insights into trait variations across genotypes, further illustrating the diversity within the studied population. Our findings indicate that several genotypes not only display high tolerance to cold but also possess significant fruiting potential. These characteristics can be strategically harnessed in breeding initiatives to improve yield and fruit quality. This study lays a strong foundation for future genetic improvement efforts in plums, promoting the selection and combination of genotypes to advance plum production in the face of environmental challenges.