Wide cultivation of almond, often under the more severe environments of Central Asia and the Mediterranean region, was possible because of the availability of a highly diverse gene pool, genetic recombination promoted by its self-incompatibility, and possibly, by interspecific hybridization and gene introgression involving other members of the Amygdalus subgenus. As a result, almond is an extremely variable species, with a high morphological and physiological diversity. This variability, measured with biochemical and molecular markers , has revealed that almond is the most genetically variable of the diploid Prunus cultivated species. In the Mediterranean Region, 2000 years of almond culture concentrated production to specific areas, where well-defined seedling ecotypes and local cultivars evolved. By the turn of the 20th century, most of these almond-producing countries had identified locally desirable cultivars that were often seedling selections of unknown origin. Thus, growers selected cultivars and landraces, which represented a rich genetic diversity. Most of these Mediterranean local cultivars have largely disappeared from cultivation in the last 50 years. Modern almond cultivation is based on a reduced number of cultivars grafted onto soiladapted clonal root stocks and cultivated under irrigated conditions when possible. Modern almond breeding started in the 1920s with the making of controlled crosses and seedling selections to meet changing agronomic and market demands. Currently,growing blueberries there are six active public breeding programs worldwide: the USA , Spain , Australia , and Israel .

Some private breeding programs exist also in the USA. In addition, there were various breeding initiatives in Russia, France, Greece, Italy, and Argentina. Different breeding objectives were developed according to regional agronomic, commercial, and market requirements. One of the main differences in the objectives is nut shell hardness. Two types of almonds are bred: soft-shelled and hard shelled . Common aims of Mediterranean breeding programs are self-compatibility and late-blooming, as most traditional almond cultivars are self-incompatible and early blooming.During the last 50 years, almond breeding for self-compatibility has mainly used two sources of Sf, local landraces originated in Italy and related species such as P. persica and P. webbii. Almond breeders have relied mainly on out crossing and, occasionally, on introgression from other Prunus species, for the development of new cultivars. Initially, in the USA and later in Russia and Mediterranean region , rapid genetic advances were achieved. In California, “Carmel” , as “Nonpareil” pollinizer, was the first cultivar release with extensive commercial impact. In Russia and the former Soviet Union, several late-flowering and frost-hardy cultivars were obtained in the 1950s with Primorskyi later used extensively for breeding in Europe. In the Mediterranean region, late flowering, productive, well-adapted, and resilient cultivars like Ferragnès or Masbovera were released with great success. The French self-compatible cultivar Lauranne showed a broad environmental adaptation, high production, and regular cropping. Although improved cultivars continued to be released, the amount of progress per generation diminishes since parents were continually drawn from the same gene pool. This situation has resulted in a potential loss of genetic variability in new breeding stocks and cultivars. Inbreeding depression in almond, expressed as low vigor, reduced flower number and fruit set, increased fruit abortion, lowered seed germination and seedling survival, increased leaf and wood abnormalities, and loss of disease resistance have been reported.

In addition, low self fruitfulness in self-compatible almond genotypes was suspected to be due to inbreeding. Regarding breeding for self-compatibility, male parents carrying the Sf allele and sharing the other S allele with the female parent are commonly used. In addition, crossing heterozygous self-compatible parents in breeding programs has been suggested to obtain homozygous self compatible genotypes to be used in further breeding. Such breeding strategies can narrow the genetic variability of crops when they lead to a reduced number of genotypes utilized as parents. Summarizing, modern almond breeding and production are dominated by a small number of widely distributed and related cultivars. This situation can lead to a potential increase of inbreeding depression and genetic vulnerability, i.e., susceptibility of most of the grown cultivars to biotic and abiotic stresses due to similarities in their genotypes. Therefore, it is needed to have up-to-date information of the relationships among genotypes used at breeding and production levels. Several almond populations have been analyzed with molecular markers in order to determine genetic variability and relatedness. However, these studies were performed with material from limited geographic areas and do not represent the current worldwide status of almond breeding stocks. Although genomic measures of inbreeding are more accurate than those obtained from pedigree data, pedigree-based analysis is a cost effective technique to estimate these parameters inbreeding populations and an alternative when genomic measures are unviable. Several reports have evaluated inbreeding based on pedigree data in breeding populations of fruit and nut tree crops. In almond, a pedigree analysis of 123 different genotypes from the USA, France, Spain, Israel, and Russia was reported. However, their work was mainly focused on North American genotypes and did not include many cultivars that have subsequently been released worldwide.

This study aimed to determine the genetic structure of current breeding stocks and breeding tendencies over the last 50 years using marker-verified pedigree data.Pedigree data of 220 almond genotypes were compiled from available bibliography and breeding records. From the 220 almond genotypes, 37 genotypes were no longer available as they were eliminated some time ago or were from discontinued breeding programs. To verify parental relationships of the rest of genotypes , we used SSRs, SNPs, and self incompatibility S-allele data from previous studies performed by the breeding programs taking part in this study . Marker data confirmed both parents of 71 genotypes and one parent of four genotypes and found three erroneous parentages. Two wrong parentages were found on the male parent of “Capella” and “Davey”, changing their pedigree to open-pollinated and a third incorrect parentage on “Yosemite” female parent, eliminating this genotype from the analysis. After the corrections made, pedigrees of 169 genotypes of known origin were analyzed . The origin of the genotypes were 59 from Spain, 56 from the USA, 16 from Russia, 11 from Israel, 10 from France, 7 from Australia, 7 from Greece, 2 from Argentina, and 2 from Italy. A pedigree data file was created. Each record in the file contained one cultivar or selection name, the female parent and the male parent, in that order. Once entered,square plant pots these data were available for inbreeding analyses such as determining the number of times a cultivar appeared in a pedigree as a male or female genitor. Genotypes of known origin were classified into two groups according to self compatibility: 104 self-incompatible and 65 self compatible.In summary, the inbreeding coefficient measures the probability that two alleles in a locus are identical by descent and so copies of the same allele from a previous generation. The pairwise relatedness measures the probability that two alleles at any locus are identical by descent between two different individuals. F and r range from 0 to 1, with values close to 0 indicating a low degree of inbreeding or relatedness and values close to 1 indicating a high degree of inbreeding or relatedness. The genetic contribution estimates the proportion of genome that comes from the same individual. Thus, a child will have 0.5 genome of either parent and a grandchild will have 0.25 genomes of his grandparents.To calculate F, r, and GC, parents of unknown origin were assumed to be unrelated and noninbred. The seed parent involved in all open pollinations was also assumed to be unrelated to the pollen parent. These assumptions, based on the fact that most almond cultivars are obligate out crossers because of their self-incompatibility, may lead to an underestimation of inbreeding. In the cases of genotypes of open-pollinated origin , numbers OP1, OP2, and OP3 were given to the pollen parent in order to be distinguishable for genetic studies. Also, all mutants were considered to have no genetic differences from the original cultivar, thus GC = 1. Since the differences between such mutants and the original cultivar are expected to be caused by a few mutations in the DNA, this simplification avoids the overestimation of inbreeding coefficients. Cultivars like Supernova and Guara were considered as “Tuono” clones. Regarding the different clones of the French paper-shell cultivar Princesse, used in both the USA and Russian breeding programs, we adopted the approach of Lansari et al.by analyzing both clones as the same cultivar. Historical reports suggest that the Hatch series “Nonpareil”, “I.X.L.”, and “Ne Plus Ultra” were seedling selections from an open-pollination progeny of the early-introduced cultivar Princesse.

This cultivar probably originated from the Languedoc region in France. Also, “Nikitskij” was selected in France in 1902. Because their specific origins remain uncertain, we analyzed these genotypes as nonrelated, which, however, could lead to an underestimation of inbreeding. Pedigree data were analyzed at four levels: worldwide, by country , by breeding program , and by genotypes carrying the Sf allele for self-compatibility.Our genetic study of almond breeding programs worldwide demonstrated that the most widely used cultivars were Nonpareil, Tuono, Cristomorto, and Mission. “Nonpareil” had a large influence in USA and Australian programs, where soft-shelled nuts are bred. This reference cultivar was present in all the breeding programs studied . The self-compatible “Tuono” and the late blooming “Cristomorto” were extensively used in the Mediterranean programs, where hard-shelled nuts are bred. “Mission” initially showed a considerable importance worldwide, but deeper analysis demonstrated that it was mainly influential in private American programs. Taking into account these results, we can establish two main breeding lines based on the use of three different founders: the European programs based mainly on “Tuono” and “Cristomorto” , and the North American–Australian programs based on “Nonpareil” . The French and Spanish breeding programs were based directly on “Tuono” and “Cristomorto”. In the French INRA program, the Italian cultivars Tuono and Cristomorto account for 60.0% of total GC and were present in the pedigree of all ten cultivars and selections evaluated. Also, the local French late-flowering and Monilinia-resistant cultivar Aï was a parent to both “Ferragnès” and “Ferraduel”. In the three Spanish breeding programs, the importance of “Tuono” and “Cristomorto” cultivars was very high, accounting to 46.2% of total GC. These two cultivars were present in the pedigree of 53 out of 59 cultivars and breeding selections from Spain. These results can be explained by the large influence of the French germplasm on the Spanish breeding programs, causing a high relationship between the programs of both countries . In the North American breeding programs, “Nonpareil” accounts for 43.7% of the total GC and was present in the pedigree of 48 out of 56 cultivars and breeding selections from the USA. In Australia, ‘Nonpareil’ accounts for 39.3% of the total GC and is present in the pedigree of 6 out of 7 cultivars and breeding selections. Also, “Lauranne” reaches an importance similar to ‘Nonpareil’, explaining the close relationship between the Australian and French programs . Even in other countries with noncontinuous breeding initiatives, such as Russia, Greece, or Argentina, the use of “Nonpareil” as a founder was common. Israel was the only country where these cultivars had a relatively low influence. This may be due to the extreme Israeli climatic conditions, forcing breeders to use locally adapted selections as parents. In Spain, the use of locally adapted cultivars such as Bertina at CITA as a donor for Polystigma ochraceum Sacc. resistance was successful but used only to a limited extent. Other examples of secondary founders include “Primorskyi”, used regularly as late-blooming and Fusicoccum-resistance donor in two of the Spanish breeding programs and “Eureka” and “Harriott” in the North American breeding programs.Pedigree analysis is a cost-effective and well-established way to monitoring inbreeding and relatedness among controlled breeding populations. However, the veracity of any analysis based on this kind of data relies on the accuracy of records collected across multiple institutions and by many breeders. In order to verify parental relationships of the genotypes under study, we used SSRs, SNPs, and self-incompatibility S-allele data from previous analysis carried out by the breeding programs taking part in this study. Our molecular marker analysis confirmed 146 parentage relationships and found three errors , which were corrected accordingly. Thus, the marker-based pedigree analysis performed showed only small parental changes and corroborates the consistency of the results reached by this study.However, several reports have demonstrated that large scale genomic analysis may provide more accurate results than pedigree analysis. This kind of genome-based pedigree analysis has already been performed in apple.