Self-Incompatibility in Apricot: Identifying Pollination Requirements to Optimize Fruit Production

In recent years, an important renewal of apricot cultivars is taking place worldwide, with the introduction of many new releases. Self-incompatible genotypes tolerant to the sharka disease caused by the plum pox virus (PPV), which can severely reduce fruit production and quality, are being used as parents in most breeding programs. As a result, the self-incompatibility trait present in most of those accessions can be transmitted to the offspring, leading to the release of new self-incompatible cultivars. This situation can considerably affect apricot management, since pollination requirements were traditionally not considered in this crop and information is lacking for many cultivars. Thus, the objective of this work was to determine the pollination requirements of a group of new apricot cultivars by molecular identification of the S-alleles through PCR amplification of RNase and SFB regions with different primer combinations. The S-genotype of 66 apricot cultivars is reported, 41 for the first time. Forty-nine cultivars were considered self-compatible and 12 self-incompatible, which were allocated in their corresponding incompatibility groups. Additionally, the available information was reviewed and added to the new results obtained, resulting in a compilation of the pollination requirements of 235 apricot cultivars. This information will allow an efficient selection of parents in apricot breeding programs, the proper design of new orchards, and the identification and solution of production problems associated with a lack of fruit set in established orchards. The diversity at the S-locus observed in the cultivars developed in breeding programs indicates a possible genetic bottleneck due to the use of a reduced number of parents.


Introduction
Apricot (Prunus armeniaca L.) belongs to the genus Prunus in the Rosaceae family. This crop was domesticated in different domestication events ca. 2000-3000 years ago in Central Asia, which is considered the center of origin of the crop [1,2]. Later, apricot was introduced into the Mediterranean Basin from the Caucasus, a secondary center of diversification [3]. Nowadays, apricot world production has reached 3.7 million tons [4] and it is considered one of the most economically important fruit crops in temperate regions [5].
Apricot cultivars have been traditionally classified into six eco-geographical groups according to their geographical origin: Central Asian, East Chinese, North Chinese, Dzhungar-Zailij, Irano-Caucasian, and European [6]. Apricot cultivars from the Central Asian, which is the oldest and most diverse, the Dzhungar-Zailij, and the Iranian-Caucasian groups, are mostly self-incompatible. Commercial cultivars of Europe, North America, South Africa, and Australia are mainly self-compatible and belong to the European group [7], which has two main gene pools: Continental Europe and Mediterranean Europe [8].
the combination of the primers SRc-F/SRc-R amplified the first intron of the RNase, and the different alleles were classified according to the sizes of the fragments previously established by Vilanova et al. [25] ( Figure 1A; Supplementary Table S1). To our knowledge, the sequence of the first intron has only been reported in the S 1 , S 2 , S 4 , S 6 , S 7 and S c -alleles [22]. The PruC2/PruC4R primer combination amplified the second intron and was used to differentiate the S 6 and S 9 alleles in 15 genotypes, whose sequences were also previously reported [21,46]. The SHLM1 and SHLM2 primer combination amplified a fragment of 650 bp in 17 cultivars, indicating the presence of the S 1 -allele. A 413 bp fragment, that corresponds to the S 7 -allele, was only detected in 'Charisma' and 'Ninfa' using the primer pair SHLM3/SHLM4 [22]. Table 1. Incompatibility group (I.G.) and S-genotype of 103 apricot cultivars.
In order to establish the compatibility relationships among cultivars, the 12 selfincompatible cultivars were allocated in their corresponding incompatibility groups according to their S-genotypes. In five cultivars ( Dama Toronja , Tornado , Vitillo , Fuego , Mogador ), self-(in)compatibility could not be established by their S-genotype because only one allele other than S c could be identified and, consequently, they were considered unclassified ( Table 1).

Diversity in the S-Locus Region
The S-genotypes of 235 apricot accessions, including the 60 cultivars analyzed herein in which two S-alleles could be identified together with 175 accessions previously reported (Tables 1 and 2; Supplementary Table S2) were used to assess the diversity and differentiation at the gametophytic self-incompatibility S-locus.
Fourteen S-alleles in 32 S-locus combinations were identified within 70 traditional apricot accessions, whereas 36 S-genotype combinations with 17 S-alleles were found in the group of 157 releases from breeding programs (Tables 3 and 4). Both groups of accessions showed the same value of the average number of alleles per country (N a = 4). Thirteen S-alleles were present in both groups ( Figure 3). However, the alleles S 4 , S 10 , S 24 , and S 31 were only identified in the commercial cultivars 'Ezzine (S 24 ), 'Harmat (S 10 ), 'Harcot (S 4 ), 'Cow-1 (S 31 ), and 'Cow-2 (S 31 ), and the S 5 -allele was only found in the traditional cultivars 'Búlida' and 'Velázquez', from Spain, and 'Shalakh', from Armenia. Table 3. Genetic parameters of apricot traditional cultivars including landraces and local selections. Number of cultivars, number of alleles (N a ), allelic richness (A r ), and number of private alleles (P a ) for each country of origin. SD: standard deviation; SE: standard error. In both traditional cultivars and cultivars released from breeding programs, S c was the most frequent S-allele as it was found in 38 and 101 cultivars, respectively. The S c -allele was not detected in Armenian and Turkish germplasm but was present in more than 50% of the genotypes from Australia, France, Greece, Romania, Spain, and Ukraine ( Figure 3A; Supplementary Table S3). A similar trend was observed in cultivars from breeding programs, except for cultivars from North America (Canada and the USA), in which the S c -allele appeared in less than 20% of the cultivars ( Figure 3B; Supplementary Table S4). Table 4. Genetic parameters of apricot cultivars released from breeding programs. Number of cultivars, number of alleles (N a ), allelic richness (A r ), and number of private alleles (P a ) for each country of origin. SD: standard deviation; SE: standard error.  The existence of private alleles, S-alleles that are found only in a single population, was used as an indicator of genetic differentiation between both groups. Three alleles (Pa) were found in the group of traditional cultivars. S3 was only found in 'Perfection' and 'Sun Glo' from the USA, S12 in 'Bedri Ahmar', 'Bouthani Ben Friha', 'Oud Rhayem', and 'Oud Hmida', from Tunisia, and S19 in 'Mari de Cenad' from Romania. In the group of releases from breeding programs, seven private alleles were identified, three of them, S10, S11 and S12, present in Hungarian accessions ('Harmat', 'Korai zamatos', and 'Voski'). In traditional cultivars, S 2 was the second most frequent S-allele (n = 24), followed by S 9 (n = 13), S 6 and S 8 (n = 8), S 7 (n = 7), S 12 and S 20 (n = 5), S 1 , S 3 , S 5 , S 11 , and S 13 (n = 3), and S 19 (n = 1). In cultivars from breeding programs, S 9 was the second most frequent S-allele (n = 35) followed by S 2 (n = 26), S 1 (n = 24), S 6 (n = 23), S 8 (n = 21), S 3 (n = 14), S 7 (n = 8), S 13 and S 20 (n = 3), S 11 and S 31 (n = 2) and S 4 , S 10 , S 12 , S 19 , S 24 (n = 1) (Supplementary Tables S2-S4).
A higher percentage of single alleles was observed in cultivars from breeding programs compared to traditional cultivars: 17% in France, 24% in Spain, and 25% in Switzerland for the cultivars from breeding programs compared to 10% in Greece and 6% in Spain for the traditional cultivars (Figure 3).
Regarding the allele frequencies, the chi-squared tests showed a statistically significant relationship (p < 0.05) between S-alleles and countries in the groups of both traditional cultivars and releases from breeding programs ( Supplementary Tables S3 and S4).
The existence of private alleles, S-alleles that are found only in a single population, was used as an indicator of genetic differentiation between both groups. Three alleles (P a ) were found in the group of traditional cultivars. S 3 was only found in 'Perfection' and 'Sun Glo' from the USA, S 12 in 'Bedri Ahmar', 'Bouthani Ben Friha', 'Oud Rhayem', and 'Oud Hmida', from Tunisia, and S 19 in 'Mari de Cenad' from Romania. In the group of releases from breeding programs, seven private alleles were identified, three of them, S 10 , S 11 and S 12 , present in Hungarian accessions ('Harmat', 'Korai zamatos', and 'Voski').
Although the number of genotypes in the group of cultivars from breeding programs was nearly twice than those in the group of traditional cultivars, the same value of average allelic richness (A r = 1.70) was found in both groups. Comparisons of traditional cultivars with cultivars from breeding programs revealed a slight loss of diversity for the S-locus in Hungarian, Tunisian, and Turkish modern cultivars. On the contrary, higher allelic richness in cultivars from breeding programs than in traditional cultivars were observed in France, Italy, and Spain.

Self-and Cross-Incompatibility in Apricot
The S-genotype of 66 apricot cultivars was reported herein, 41 for the first time. A total of 49 cultivars were characterized as self-compatible since their genotype contained the S c -allele, which is associated with self-compatibility [56]. However, the self-compatibility of two cultivars was determined through controlled pollinations due to mismatching of PCR primers resulting in no amplification. When self(in)compatibility cannot be determined by identifying the S-genotype, laboratory pollination experiments have proven to be an accurate method because they avoid weather-related failures under field conditions [23,57,58].
According to their S-allele composition, 12 self-incompatible cultivars were allocated in their corresponding cross-incompatibility groups together with the 77 self-incompatible cultivars previously analyzed [16,21,22,24,32,34,35,[47][48][49][50]. To date, a total of thirty-five incompatibility groups (I to XXXV) have been described in apricot [17,21,22,32,35,48]. Selfincompatible cultivars within the same incompatibility group have the same S-genotype and are genetically incompatible with each other. On the other hand, cultivars from different incompatibility groups are inter-compatible, since at least one of the S-alleles of their genotype is different [60]. Although lower yields have been related to semi-compatible pollinizers (when one S-allele is identical and the other differs) in Japanese plum and sweet cherry [61][62][63], there is no information on the effects of semi-compatibility in apricot. In addition, 16 cultivars are included in group 0 since no other cultivars with the same S-genotype have been reported until now. Thus, the cultivars from group 0 as well as the self-compatible cultivars could act as universal pollinizers. This could be highly valuable information since the knowledge of incompatibility relationships aims to help breeders to choose parental genotypes for breeding programs and fruit growers to select compatible pollinizers coincident at flowering time.

Current Genetic Diversity at the S-Locus
To provide an overview of the genetic diversity at the S-locus of currently grown apricot cultivars, results herein have been combined with previous results [16,21,22,24,25,32,34,35,39,[47][48][49][50]52,54,55] to analyze the frequency and distribution of the S-alleles. A total of 33 S-alleles (S 1 to S 20 , S 22 to S 30 , S 52 , S 53 , S c , S v , and S x ) have been described in apricot cultivars [16,[24][25][26][27]. Fourteen S-alleles were detected within the group of 70 traditional cultivars from Armenia, Australia, France, Greece, Hungary, Italy, Romania, Spain, Tunisia, Turkey, Ukraine, and the USA, whereas 17 S-alleles were identified in the group of cultivars from breeding programs, reflecting lower S-allele diversity in the group of traditional cultivars. Twenty S-alleles (20) were reported in a group of 67 cultivars from Europe and North America [16], but fewer S-alleles were found when local accessions or landraces were studied separately in Tunisia [17,35], Turkey [32,36,37], and Morocco [33].
Four S-alleles (S 10 , S 24 , S 31 and S 4 ) were exclusively found in the group of cultivars from breeding programs, which could be related to the use of landraces instead of commercial cultivars in some breeding programs. Two of these S-alleles showed a clear relationship with a specific breeding program: S 24 was found in 'Ezzine' from INRAT (Tunisia) [64], and S 31 was found in 'Cow-1 and 'Cow-2 from INRA (France) [65]; S 4 was present in the North American cultivar 'Harcot' [49]. Although this cultivar has been used as a parental genotype in several breeding programs to introduce Sharka-resistance to new releases [29], the S 4 -allele has not been found in any recent releases. On the other hand, the S 5 -allele has been previously reported only in some traditional cultivars from Spain and Armenia. Our results differed in the S-genotype of the Spanish landrace 'Corbato' in which this allele has been previously reported, (S c vs. S 2 S 5 [16]). The presence of the S 5 -allele in this cultivar, as well as in other traditional cultivars from Spain and Armenia, has been reported by Muñoz-Sanz et al. [16]. They suggested that the presence of the S 5 -allele in those populations could be the result of a connection between Southern-Spanish accessions with the Armenian and Eastern-Turkish accessions; in addition, they suggested that Moroccan accessions are a part of the Southwest-Mediterranean apricot diffusion route.
Traditionally, the main marker for self-compatibility in apricot has been S c [49]. For this reason, one of the main objectives of breeding programs has been to introduce this S-allele into new releases [11]. Our results showed that S c was the most frequent S-allele in both the groups of traditional and cultivars from breeding programs. Although we did not detect the S c -allele in traditional cultivars from Armenia and Turkey, previous reports found some self-compatible Turkish cultivars carrying this S-allele [32,37]. The S c -allele might have evolved in Southeastern Turkey as a result of a pollen-part mutation within SFB 8 , causing the pollen with the mutated S 8 haplotype to be self-compatible [52]. Subsequently, it is hypothesized that the allele S c was disseminated to the Mediterranean Basin since the cultivars from Central Asia, the center of origin of apricot, are self-incompatible [52,66]. The S c -allele was present in cultivars from breeding programs in most countries except for Turkey, probably due to the low number of new releases from this country studied in this work. However, the low presence of the S c -allele in North American cultivars (the USA and Canada) is probably due to the fact that most of these cultivars are self-incompatible [28,67].
The presence of just one allele has been observed in a high number of cultivars, mostly cultivars released from breeding programs, mainly as genotype S c -(Supplementary Table S2) that would presumably correspond to S c S c . Homozygote cultivars can arise as a result of self-and cross-pollinations with self-compatible parents in breeding programs. Although this genotype has been frequently found in apricot cultivars from most European countries [16,25,34,52], results herein as well as those from previous studies could not confirm homozygosity since sequencing would be needed [22].
The distribution of S-alleles varied considerably between countries, with a significant association (p < 0.05) between S-alleles and geographical origin, both in the group of traditional accessions and in the group of breeding program releases. Within traditional cultivars, the S 3 -allele was found exclusively in North American cultivars [16]. The S 12 -allele was only found in four cultivars from Tunisia [35], being one of the most frequent S-alleles in this country [17]. The S 19 -allele was only found in one cultivar from Romania, 'Mari de Cenad', despite having been associated with Hungarian and Turkish local apricots [32]. Muñoz-Sanz et al. [16] reported that the S 19 -allele found in this cultivar could be S 20 ; however, further sequencing would be required to confirm this.
Thus, in addition to providing useful information to know the self-compatibility of cultivars, S-genotyping can be a valuable tool in elucidating the evolution and dissemination of the crop.

Self-Compatibility and Diversity
Self-compatibility has been described as a cause of loss of genetic diversity since it promotes inbreeding [68]. Self-incompatibility not only reduces inbreeding by preventing self-fertilization but also reduces mating between close relatives, ensuring the exchange of genetic material [69]. It has been suggested that the loss of genetic diversity affecting the S-locus is due to crop dissemination [16]. In fact, a bottleneck has been observed in apricot diversity as a consequence of the domestication and diffusion of the apricot throughout the history of the crop [8]. Additionally, a decrease in genetic diversity from the eastern (Iran-Caucasian area) to the south-western (North Mediterranean Basin and South Mediterranean Basin areas) distribution of the crop has been detected, analyzing local cultivars from Algeria, France, Italy, Morocco, Spain, Tunisia, and Turkey [70]. A lower number of S-alleles were found in accessions from Moroccan oases as compared to the whole allele pool in this country, probably due to the pressure to increase production and self-compatibility in the genotypes, allowing a higher level of endogamy [33]. A similar situation was described in landraces from Central Europe [52].
Apricots belonging to the European group have been traditionally considered to be self-compatible [71]. However, the number of self-incompatible commercial cultivars in the European group increased rapidly over the last two decades due to the use of selfincompatible North American cultivars as parentals in breeding programs [21,28]. Recent studies show that about half of the new releases are self-compatible: 51.1% (47 of 92) [21] and 49.6% (61 of 123) [20]. Recently, additional sources of self-compatibility have been described in apricot in addition to the S c -allele. Thus, some works have reported the existence of an additional mutation in the M-locus not linked to the S-locus, which causes a loss of pollen S-activity [16,55,56,72]. Pollen-part mutations (PPMs) in the M-locus were mapped at the distal end of chromosome three in 'Canino' (called m [72]) and in 'Katy' (called m' [55]). Muñoz-Sanz et al. [73] proposed the ParMDO gene as a relevant gene involved in pollen part SI function. In order to optimize the screening of self-compatible genotypes, a new useful method based on both loci has been recently developed [74].
Although a clear trend towards releasing self-compatible cultivars is shown, our results exhibited similar allelic richness values in both groups, traditional cultivars, and cultivars from breeding programs. A clear differentiation between apricot landraces and cultivars from breeding programs was recently revealed using SSR markers, showing an unexpected higher diversity in cultivars from breeding programs, which was related to the use of North American genotypes as parentals [75]. Our results showed a slight loss of diversity for the S-locus in Hungarian, Tunisian, and Turkish cultivars comparing landraces with releases from breeding programs. However, this situation was not observed in the countries included in the North Mediterranean Basin group such as France, Italy, and Spain. This could be due to the higher number of new releases from breeding programs of these countries.
Although there is no evidence of a reduction in the diversity at the S-locus in cultivars developed in breeding programs, results herein suggests that the use of a reduced number of parents in breeding programs can lead to a genetic bottleneck.

Plant Material
Young leaves from 66 apricot cultivars, including traditional cultivars (landraces and local selections) and releases from breeding programs of several origins (Tables 1 and 2), were collected in spring from germplasm collections and orchards in Spain. Moreover, flowers were collected from two cultivars ('Samourai' and 'Water') for pollination experiments to establish the self-(in)compatibility by microscopic observations. The apricot accessions analyzed originated from 12 countries: Armenia, Australia, France, Greece, Hungary, Italy, Romania, Spain, Tunisia, Turkey, Ukraine, and the USA.

DNA Extraction and S-Allele Identification
Genomic DNA of each sample was extracted using DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) according to Hormaza [76] and quantified by NanoDrop™ ND-1000 spectrophotometer (Bio-Science, Budapest, Hungary).
The S-genotype of each cultivar was identified through PCR amplification of RNase and SFB regions with different primer combinations (Table 5) [23]. The first intron of the S-RNase gene was amplified with the fluorescently labeled primer combination SRc-(F/R) [25,77]. PCR amplifications were carried out in 15 µL reaction volumes, containing 10× NH 4 Reaction Buffer, 25 mM MgCl 2 , 2.5 mM of each dNTP, 10 µM of each primer, 100 ng of genomic DNA and 0.5 U of BioTaq TM DNA polymerase (Bioline, London, UK). The temperature profile used had an initial step of 3 min at 94 • C, 35 cycles of 1 min at 94 • C, 1 min at 55 • C and 3 min at 72 • C, and a final step of 5 min at 72 • C. The amplified fragments were analyzed in a CEQ TM 8000 capillary electrophoresis DNA analysis system (Beckman Coulter, Fullerton, CA, the USA) and classified according to Vilanova et al. [25] and Herrera et al. [21]. Because two pairs of alleles, S 6 /S 9 and S 1 /S 7 , showed similar fragment sizes, specific primers based on the second intron of the RNase were used to distinguish between them. For the identification of the S 6 -and S 9 -alleles, the PruC2/PruC4R primer combination designed from P. avium S-RNase-cDNA sequences [78] was used to differentiate both alleles in 15 genotypes. Specific primers SHLM1/SHLM2 and SHLM3/SHLM4 were required to distinguish between S 1 and S 7 , respectively [22]. PCR reactions were carried out according to Vilanova et al. [25], but with the addition of 10 cycles and using 55 • C of annealing temperature [79]. The amplified fragments were separated on 1% (w/v) agarose gels and the DNA bands were visualized using the nucleic acid stain SYBR Green (Thermo Fisher Scientific, St Leon-Rot, Germany). For the identification of the S 1 -allele, the specific primers SHLM1 and SHLM2 were used following the protocol described by Herrera et al. [22] for Taq DNA polymerase (Qiagen, Hilden, Germany). Primers SHLM3 and SHLM4 were used for S 7 -allele identification. PCR reactions were performed with Phusion ® High-Fidelity DNA Polymerase (Thermo Fisher Scientific, St Leon-Rot, Germany) according to Herrera et al. [22].
Since PCR amplification of the S c -and S 8 -alleles using the primers SRc-(F/R) provides a fragment of similar size [24,56], the specific primers AprFBC8-F and AprFBC8-R, designed based on the V2 and HVb variable region of SFB gene, were used to distinguish between both alleles [32]. The PCR amplifications were carried out using the program previously described by Halász et al. [32].

Pollination Experiments
In cultivars 'Samourai' and 'Water', no amplification was produced with the primers AprFBC8-F/AprFBC8-R. Thus, controlled pollinations were carried out in these two cultivars to differentiate the S c -and S 8 -alleles. Self-(in)compatibility was established by laboratory-controlled pollinations and the observation of pollen tube growth under fluorescence microscopy [23]. Self-pollinations were carried out in the two cultivars. Pollen of the cultivar 'Katy', known as universal pollinizer for apricot [55], was used to pollinize another set of flowers of each cultivar as control.
Flowers from each cultivar were collected at the balloon stage one day before anthesis and emasculated to avoid self-pollination. Pistils were placed on wet florist foam and maintained at laboratory temperature. After 24 h, a group of 20-25 flowers were handpollinated with the help of a paintbrush for each self-and cross-pollination [80]. Pollen was obtained from flowers at the same balloon stage by removing and drying the anthers at laboratory temperature during 24 h. Pollen grains were then sieved by using a fine mesh (0.26 mm) and used immediately or frozen at −20 • C until further use. Seventy-two hours after pollination, pistils were fixed in ethanol (95%)/acetic acid (3:1, v/v) during 24 h, and conserved at 4 • C in 75% ethanol. After hand pollinations, pollen viability was evaluated. Pollen from each pollen donor was scattered on a solidified pollen germination medium [81] and pollen germination was observed under the microscope after 24 h. Pollen grains were considered viable when the length of the growing pollen tube was higher than the pollen grain diameter.
For histochemical preparations, the fixed pistils were washed three times for 1 h with distilled water and left in 5% sodium sulphite in distillated water at 4 • C for 24 h. Then, they were autoclaved at 1 kg/cm 2 during 10 min in sodium sulphite to soften the tissues [82]. Pistils were squashed and stained with 0.1% (v/v) aniline blue in 0.1N K 3 PO 4 [83] to observe callose. Examination of pollen tube growth was carried out by fluorescence microscopy using by a Leica DM2500 microscope (Cambridge, UK) with UV epifluorescence using 340-380 bandpass and 425 longpass filters.
Pollen tube behavior was observed in at least 10 pistils in each self-pollination. Cultivars were considered as self-incompatible when pollen tube growth was arrested along the style in most self-pollinated pistils. On the other hand, when the pollen tube reached the base of the style in most self-pollinations, the cultivars were considered as self-compatible.

S-Allele Diversity Analysis
In order to analyze the S-allele genetic diversity, the S-RNase genotypes of the 60 cultivars identified herein were compiled with those of the 175 cultivars previously reported (Tables 1 and 2; Supplementary Table S2). Cultivars were first classified into two groups according to the pedigree origin of the accessions, traditional, and releases from breeding programs. Statistical analyses were performed using the R programming environment (R Core Team, 2022, version 4.1.0, Vienna, Austria). The S-genetic profiles were stored in a csv file which was converted into a matrix of allelic frequencies stored in a genind class with the "loci2genind" function using the R package "pegas" version 1.0-1 [84]. Missing data (<0.1%) were replaced with the mean frequency of the corresponding allele, which avoids adding artefactual between-group differentiation [85].
The number of alleles (N a ), allelic richness (A r ), and private alleles (P a ) were calculated for all countries on the traditional and cultivars from breeding programs using the adegenet 2.1.3 [85], and PopGenReport 3.0.4 [86] packages. Additionally, the frequency of each S-allele was calculated in each country within each group and the results were plotted as a heatmap with the R package PopGenReport version 3.0.4 [86]. To analyze the relationship between the S-alleles and their distribution by country, a contingency table of absolute frequencies of alleles by country was created and a chi-square test was performed with the "chisq.test" function using the R package "stats" v. 4.1.2. Due to the low number of observations in some countries, a Monte Carlo simulation with 2000 replicates was indicated.

Conclusions
Results reveal that a significant proportion of new apricot releases are self-incompatible and, therefore, require cross-pollination to produce fruit. Knowing the incompatibility relationships between cultivars will help breeders to select suitable parental genotypes in crosses. This information, combined with the flowering dates in each geographical area, will allow the selection of appropriate inter-compatible pollinizers for self-incompatible cultivars in the design of new orchards. The identification of the S-alleles, in addition to the determination of the pollination requirements of the cultivars, can elucidate missing gaps in the evolution, domestication, and dissemination of the crop. The diversity at the S-locus observed in the cultivars developed in breeding programs indicates a possible genetic bottleneck due to the use of a reduced number of parents in breeding programs.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants11152019/s1, Table S1: Sizes of PCR fragment amplification using five primer pair combinations for the identification of S-alleles and S-genotype of the analyzed apricot accessions; Table S2: Origin and S-genotype of the 235 analyzed apricot accessions; Table S3: Contingency table of absolute allele frequencies by country in apricot traditional cultivars including landraces and local selections and Pearson's chi-squared test with simulated p-value (based on 2000 replicates); Table S4: Contingency table of  Funding: This research was funded by the projects PCI2020-111966, PID2019-109566RB-I00 and PID2020-115473RR-I00 funded by MCIN/AEI/10.13039/501100011033, the grant PRE2018-084962 funded by MCIN/AEI/10.13039/501100011033, and FSE "El FSE invierte en tu futuro" and Gobierno de Aragón-European Social Fund, European Union (Grupo Consolidado A12_17R).
Data Availability Statement: Not applicable.