- Open Access
Diversification and independent domestication of Asian and European pears
Genome Biology volume 19, Article number: 77 (2018)
Pear (Pyrus) is a globally grown fruit, with thousands of cultivars in five domesticated species and dozens of wild species. However, little is known about the evolutionary history of these pear species and what has contributed to the distinct phenotypic traits between Asian pears and European pears.
We report the genome resequencing of 113 pear accessions from worldwide collections, representing both cultivated and wild pear species. Based on 18,302,883 identified SNPs, we conduct phylogenetics, population structure, gene flow, and selective sweep analyses. Furthermore, we propose a model for the divergence, dissemination, and independent domestication of Asian and European pears in which pear, after originating in southwest China and then being disseminated throughout central Asia, has eventually spread to western Asia, and then on to Europe. We find evidence for rapid evolution and balancing selection for S-RNase genes that have contributed to the maintenance of self-incompatibility, thus promoting outcrossing and accounting for pear genome diversity across the Eurasian continent. In addition, separate selective sweep signatures between Asian pears and European pears, combined with co-localized QTLs and differentially expressed genes, underline distinct phenotypic fruit traits, including flesh texture, sugar, acidity, aroma, and stone cells.
This study provides further clarification of the evolutionary history of pear along with independent domestication of Asian and European pears. Furthermore, it provides substantive and valuable genomic resources that will significantly advance pear improvement and molecular breeding efforts.
Pear (Pyrus), one of the most economically important temperate fruit tree species, with an annual worldwide production of ~ 18 million tons (2015, FAOSTAT), belongs to the subtribe Malinae of the Amygdaloideae subfamily within Rosaceae . The genus Pyrus includes at least 22 recognized species , with more than 5000 accessions maintained worldwide. These accessions display wide morphological and physiological variability, as well as broad adaptations to wide agro-ecological ranges. As a self-incompatible flowering plant, pear is an obligate outcrosser. It is important to note that hybridization in pear occurred not only intraspecies but also interspecies, despite its wide geographic distribution. Although many pear groups are deemed as different species, they are in fact rather similar to subpopulations based primarily on their distinguishable phenotypes. Therefore, it is likely that inter-‘species’ hybridizations and genetic admixtures must have occurred among pear groups without reproductive barriers . Nevertheless, they have long been widely recognized and deemed as “species” in pear research studies [4, 5].
The ancient Pyrus lineage probably arose during the Tertiary period, between 65 to 55 million years ago (MYA), in the mountainous regions of southwestern China . Subsequently, it was dispersed across mountainous ranges both eastward and westward. This oriental and occidental geographical distribution of pear led to the respective development of Asian and European pears . The earliest cultivation of Asian pears can be traced back to about 3300 years ago , with commercial orchards known to have existed for more than 2000 years in China . Similarly, European pears have been cultivated for more than 3000 years, with distinct named cultivars recorded as early as 300 B.C. .
Pyrus communis, the predominant cultivated species of European pear, bears typical pear-shaped fruits with soft and smooth flesh, few stone cells, along with strong aroma and flavor. The major cultivated species in Asia, including P. pyrifolia, P. bretschneideri, P. sinkiangensis, and P. ussuriensis, bear round-shaped fruits with crisp flesh, high sugar content, low acid content, minimal aroma, and mild flavor. The genetic variations and domestication processes responsible for these observed phenotypic differences in fruit trait characters between European and Asian pears are not well understood.
Indeed, these wide genetic variations present in pear accessions, belonging to various Pyrus species, have made it quite difficult to identify relationships among pear germplasm collections. Consequently, current available pear DNA sequence data are inadequate to delineate clear population-level relationships among various pear species [10,11,12]. In the past two decades, whole-genome sequencing tools have revolutionized the field of life sciences as they have provided unprecedented new means and opportunities to explore and understand genetic variation, evolution, and domestication processes of agricultural crops . Owing to its self-incompatibility and long generation cycle, among other factors, genetic and molecular analyses of pear have been rather challenging and slow. However, recent completion of whole-genome sequencing of ‘Dangshansuli’, an Asian pear , and “Bartlett”, a European pear , has yielded new knowledge, including characterization of the genomic structure, chromosome evolution, and patterns of genetic variation related to important agricultural traits.
In this study, we conducted population-level analysis of genetic variation of pears based on the resequencing of genomes of a diverse group of 57 wild and 56 cultivated Pyrus accessions from wide geographical regions. A total of 18,302,883 genome-wide SNPs were identified and used in multiple analyses. Findings were then used to propose a model to explain divergence, dissemination, and domestication of Asian and European pears. Of particular note, analysis of the evolutionary rate and balancing selection of S-locus genes highlights the impact of self-incompatibility on the genetic diversity of pear, which likely had a strong influence on gene flow and observed genetic variations in Pyrus. Furthermore, selective sweeps associated with agriculturally important genes were detected in cultivated Asian and European pears. In addition to evolutionary and functional genomics insights, this study provides an unprecedented amount of genomic data that will almost certainly enable important advances in modern pear improvement and molecular breeding programs.
Results and discussion
Sequencing and mapping of pear accessions
Genomes of 113 Pyrus accessions, including 63 Asian (31 cultivated and 32 wild) and 50 European (25 cultivated and 25 wild) pears, were sequenced (Additional file 1: Figure S1). A total of 661 Gb of high-quality cleaned sequence data were generated, with an average of 5.85 Gb per accession (equivalent to approximately 11× coverage of the ~ 527-Mb pear genome). These sequences were aligned to the Asian pear genome ‘Dangshansuli’ , with an average mapping rate of 61.39% (Additional file 2), and a total of 18,302,883 SNPs were identified with ~ 90 SNPs per kb (Table 1, Additional file 1: Figure S2). It is important to note that considerable care was taken in selecting ‘Dangshansuli’ over the European pear genome Bartlett as the reference genome for read alignment in this study. This was a necessary step as there were two published pear genomes under consideration. However, a pseudo-alignment between these two genomes revealed high divergence, which would result in a low mapping rate for cross-species alignment. Specifically, we first looked at the quality parameters of the two published pear genomes, including the contig N50, the scaffold N50, and the scaffold size values, as well as the scaffold-to-chromosome anchoring ratios. Results revealed that the quality of the ‘Dangshansuli’ genome was higher than that of the Bartlett genome (Additional file 1: Table S1). This also suggested that no matter which genome was selected, it is unlikely that the mapping rates for divergent samples would be improved (Additional file 1: Note 1). It was ultimately decided that the high-quality ‘Dangshansuli’ Asian pear genome would be the reference in this study.
Among all identified SNPs, a total of 14.1% were located in coding regions: 7.7% were non-synonymous and 6.4% were synonymous, with a non-synonymous/synonymous ratio of 1.2 (Additional file 1: Table S2). These SNPs had potential effects on a total of 13,838 genes (32.3% of pear genes; Additional file 1: Table S3). The proportion of non-synonymous SNPs in coding regions of pear (7.7%) and apple (10.5%)  was higher than that detected in either soybean  (1.9%) or maize  (0.66%), underscoring the presence of higher levels of genetic variation in pear and other fruit trees compared to some annual crops (Additional file 1: Note 2). To validate SNP calling, both PCR amplification and Sanger sequencing were conducted for 510 randomly selected SNPs in 55 pear accessions, and a 97.5% consistency for SNP calling was obtained (Additional file 3).
Nucleotide diversity and linkage disequilibrium in genomes of Asian and European pears
The nucleotide diversity (ϴπ) of pear at the whole-genome level across all pear accessions was 5.5 × 10− 3 (Table 1). This was higher than those reported for other perennial crops such as peach (1.5 × 10− 3) , cassava (2.6 × 10− 3) , and grapevine (5.1 × 10− 3) , but lower than that reported for date palm (9.2 × 10− 3) . Notably, both wild and cultivated Asian pears had higher nucleotide diversity (5.21 × 10− 3 and 4.76 × 10− 3, respectively) than either wild (3.57 × 10− 3) or cultivated (3.53 × 10− 3) European pears. However, in both Asian and European pears, similar levels of nucleotide diversities were detected for wild and cultivated accessions. This is in sharp contrast to findings reported in both soybean  and rice , wherein strong positive selection has contributed to wide differences in nucleotide diversity observed between wild and cultivated populations .
To explore relationships among various cultivated and wild pear accessions, principal component analysis (PCA) of all 113 accessions was conducted using ~ 18 M SNPs, which revealed the presence of two distinct groups, namely Asian and European pears (Fig. 1a). Linkage disequilibrium (LD) analysis showed that pear genomes have relatively short LD distances and relatively rapid LD decays (Fig. 1b). The average r2 value among pear SNPs, corresponding to LD levels of the population, was relatively low (0.386; Additional file 1: Table S4). The average distance over which LD decayed to ~ 50% of its maximum value in pear was very short, 553 bp in cultivated Asian pears and 154 bp in cultivated European pears. r2 values for cultivated groups were higher for wild accessions of both Asian and European pears (Additional file 1: Table S4). Moreover, cultivated Asian pears showed the highest LD, followed by cultivated European pears. The LD decay distances for pears (211 bp) and apples (161 bp)  are much shorter than those reported for annual crops, such as soybean (~ 150 kb)  and rice (123 kb) .
Our findings of similar levels of nucleotide diversity between wild and cultivated pears, short LD distances, and rapid LD decay in cultivated pears all support relatively weak selection during pear domestication. This could be explained by high outcrossing rates that are maintained by self-incompatibility, as well as the short domestication history and long generation time of this perennial fruit crop . Furthermore, in view of the fact that the major method for propagation of pear cultivars is by grafting, this would contribute to low numbers of sexual generations during the domestication history, and would also contribute to the weak selection during pear domestication.
The phylogeny and structure of Asian and European pear populations
A phylogenetic analysis indicated that pear accessions were clustered into two groups, Asian and European pears. This was consistent with the result of the ΔK analysis, which revealed an optimal K of 2 in the pear populations (Additional file 1: Figure S3), and was also similar to the pattern observed in our PCA score plot (Fig. 1a). Both phylogenetic analysis and population structure analysis (K = 5) revealed that Asian pear accessions were clustered into four groups (Fig. 2a). Asian group I consisted of the largest number of accessions and included accessions from both P. bretschneideri and P. pyrifolia. Asian group II included wild accessions from China, Japan, and Korea. Notably, the close phylogenetic relationship of cultivated P. bretschneideri and P. pyrifolia with wild P. pyrifolia provided whole-genome-level evidence to support the hypothesis that the two cultivated species of Asian pear, P. bretschneideri and P. pyrifolia, were derived from a common ancestor, the wild P. pyrifolia [27, 28]. Asian group III included both cultivated and wild accessions of P. ussuriensis, and these accessions were adapted to and cultivated in extremely cold regions of China. Finally, Asian group IV included all cultivated accessions derived from P. sinkiangensis, and revealing an admixed genetic background between Asian and European pears. Thus, it was not surprising that the Pyc-co6 accession, a recently bred hybrid from a cross of a cultivated Asian pear and a cultivated European pear, clustered in between the Asian and European groups.
On the other hand, the phylogenetic analysis revealed that European cultivated pears formed a clade that was nested within wild European pear accessions (Fig. 2a). However, it was interesting to note that there was little change in the population structure for European pears, with increasing K values from 2 to 7 (Additional file 1: Figure S3a). In view of the high polymorphism and diversity of Asian pears, which might influence the population structure of European pears, the population structure analysis for European pears was conducted independently. It was found that European pears could be classified into two groups (Fig. 2b). European group I included wild accessions from Europe and North Africa. Almost all of the cultivated European pears clustered together into European group II, except for the wild accession Pyw-ni1 belonging to P. nivalis, an atypical pear used in the production of “perry” cider. Among 13 wild accessions in European group I, P. pyraster accessions were those most closely related to European group II accessions. Thus, it seems that P. pyraster, which grows widely throughout Europe , is likely the progenitor species from which cultivated European pears are derived. It is important to note that very few changes in clustering of European pears were observed at different K values (Additional file 1: Figure S3a). This is supported by findings obtained from ϴπ analyses revealing that, relative to sampled Asian pears, there was lower genetic diversity present among sampled European pear accessions. Thus, for subsequent gene flow and identity-by-descent (IBD) analyses, European group I accessions were split into two subgroups based on their geographical sampling locations. European group I subgroup 1 included accessions that were relatively closer to east Asia, while European group I subgroup 2 accessions were relatively farther away from east Asia (Additional file 1: Table S5; Fig. 2a).
Subsequently, gene flow within and between Asian and European pears were explored. First, we used groups from the phylogenetic tree to conduct a gene flow analysis using Treemix within Asian pears. While we detected a relatively strong gene flow from Asian group I to Asian group IV (P. sinkiangensis), we did not detect significant gene flow among the other pairings of Asian pear groups (Fig. 2c). As for European pears, extensive gene flow (P value = 2.2e-308; F_statistic = 0.988) was detected between European group II and European group I subgroup 2 (Fig. 2d). To further investigate gene flow between Asian and European pears, an analysis was conducted using all groups with no significant gene flow between Asian and European groups. A weak gene flow was detected from European pear group accessions to the Asian pear accession of P. sinkiangensis (Fig. 2e). All these findings are consistent with the earlier hypothesis that P. sinkiangensis is derived from a hybridization between Asian and European pears.
An IBD analysis of Asian group IV (P. sinkiangensis) was conducted to verify gene flow into this species from both cultivated Asian and cultivated European pears (Fig. 2f). It was observed that the proportion of genetic background from Asian cultivated pears was 45.3–61.8% in P. sinkiangensis, which was higher than that detected from European group II (17.9–35.3%). IBD analysis of European group I revealed that P. sinkiangensis contained 10.9–23.0 and 11.8–26.7% of the genetic backgrounds of European subgroups 1 and 2, respectively. This was lower than that detected for European group II, thus indicating that cultivated European pears contributed a higher proportion of genetic background to P. sinkiangensis compared with wild European pears. Therefore, this IBD analysis further supported Treemix results noting that P. sinkiangensis was the product of a hybridization that occurred between cultivated Asian and cultivated European pears. These findings are reasonable to expect from a historical perspective, as there was extensive cultural contact along the Silk Road from 207 BCE to 220 CE . Interestingly, there is a historical record from about 2000 years ago of a Han dynasty diplomat, Qian Zhang, bringing over cultivated Asian pear to the Xinjiang region . Given our finding that the admixed P. sinkiangensis species must have resulted from hybridization between cultivated Asian and cultivated European pears, we can speculate that historical and commercial influences may have contributed to the development of this unique species of cultivated pear.
Origin and dissemination of wild pears
To explore the origin and dissemination of pears, we divided the 57 wild pear accessions into three groups, which were the same as the aforementioned European group I, Asian group II, and Asian group III groups (Fig. 3a). The filtered SNP dataset of 57 wild pear accessions (excluding all cultivated pear accessions to avoid gene flow effects), consisting of 16,320,215 SNPs, were subjected to both population structure analysis (Fig. 3b) and PCA (Fig. 3c). Results strongly supported a population structure of K = 3. Of particular note, P. ussuriensis accessions that comprised Asian group III, known to be highly cold-tolerant, were collected from extremely cold environments in what is currently the northeast region of China. Mantel tests showed that this population structure was highly correlated with geographical distribution (P = 1e-04). This suggests that geographical factors have been highly influential in generating the observed genetic variation present among wild pear populations across the Eurasian continent. Interestingly, this also strongly suggests that several accessions originally annotated to belong to P. regelii, P. armeniacaefolia, P. xerophila, P. hopeiensis, and P. fauriei species were instead highly admixed (Fig. 3d). This finding was not surprising as these “species” were not reproductively isolated. Therefore, it is likely that inter-“species” hybridizations and genetic admixtures occurred among these pears.
Levels of population differentiation, FST, were then estimated across all chromosomes among the three groups of wild accessions (Fig. 3e). The FST between Asian group III and Asian group II was smaller than that between European group I and Asian group II. This suggested that the divergence of Asian and European groups preceded the divergence of Asian group II from Asian group III. Furthermore, when ϴπ analysis was conducted to evaluate levels of genetic diversity of these groups, it was found that Asian group II and Asian group III showed the highest levels of diversity (Fig. 3f) when compared to other groups.
In addition, IBD analysis was conducted to assess the same DNA segments within and across accessions (Fig. 3g). Overall, IBD values were higher for accessions in both Asian group II (0.59) and Asian group III (0.54), followed by those found in European group I, including central Asia (0.47), Western Asia (0.38), and the European mainland (0.439). Results of population structure analyses of wild pear accessions along with their geographical distributions support the hypothesis that pear must have originated in what is now known as the southwest region of China. Subsequently, it was then disseminated throughout central Asia before it was further spread over to western Asia and then to Europe.
A proposed model for the evolutionary pathway of pear
The divergence time of Asian and European pears was estimated by constructing a phylogenetic tree using a total of 420 single-copy conserved genes from nine plant species (Fig. 4a). It was estimated that Asian and European pears diverged between 6.6 and 3.3 MYA, far prior to any possible human intervention, perhaps mediated by animals through the dispersal of fruit and pollen. Based on the phylogeny, population structure, gene flow, and IBD analyses, we propose the following model for divergence and independent domestication of Asian and European pears (Fig. 4b).
The hypothetical common ancestor of both Asian and European pears seems to have originated in China, subsequently disseminated through central Asia, and then eventually on to western Asia and Europe. Considering the fact that there is no reproductive isolation in the population, and given that pear is a typical self-incompatible species and an obligate outcrosser, it is likely that a “continent-wide species” must have undergone local adaptation followed by independent domestication processes for each of Asian and European pears. Each of these domestication processes must have involved selection for distinct phenotypic traits, including distinctive fruit shape, flavor, and texture traits that are now characteristic of Asian and European pears [32, 33].
This proposed model clarifies present-day complex relationships among the large numbers of so-called pear species. The five currently recognized cultivated pear species have been domesticated from three wild species. This is quite different from other crops which are often domesticated from a single wild species [34, 35]. The only species of European pear, P. communis, is derived from a wild European species, P. pyraster. One of the four species of Asian pear, the cultivated P. ussuriensis, is derived from the wild P. ussuriensis. Two other cultivated species of Asian pear, P. pyrifolia and P. bretschneideri, are derived from a common ancestor, the wild P. pyrifolia. Finally, the admixed species of the fourth cultivated Asian pear, P. sinkiangensis, is derived from hybridization that must have occurred within the last 3000 years between the cultivated European pear (P. communis) and the cultivated Asian pear, either P. pyrifolia or P. bretschneideri. The IBD analysis indicates similarly sized genome contributions from P. pyrifolia and P. bretschneideri to P. sinkiangensis.
Balancing selection along with rapid evolution of S-RNase genes have strengthened self-incompatibility in pear
Pear exhibits typical gametophytic self-incompatibility (GSI), which is controlled by a single multi-allelic locus, the S-locus. The S-locus contains the pistil determinant, S-RNase, and candidate pollen determinant S-locus haplotype F-box genes, SFB genes [14, 36]. It is commonly known that the S-RNase gene exhibits high sequence variability among different pear cultivars [37, 38].
To analyze allelic diversity of the S-RNase gene, cleaned reads of each pear accession were mapped onto the S-RNase locus of the reference pear genome of ‘Dangshansuli’. A total of 92 SNPs were detected among S-RNase alleles of wild accessions and 141 SNPs among S-RNase alleles of cultivated accessions of pear. Mean ϴπ values were 1.70 × 10− 1 for wild accessions and 1.72 × 10− 1 for cultivated accessions (Additional file 1: Table S6). These mean ϴπ values were much higher than the mean diversity of the genes (1.56 × 10− 2) detected in the whole genome. Notably, the high genetic diversity of S-RNase alleles was almost identical in both cultivated and wild accessions (Additional file 4), suggesting that this gene has not experienced strong selection pressure under human intervention. Meanwhile, both cultivated and wild Asian and European pears had positive Tajima’s D values (> 2.0) for the S-RNase gene. This indicated that a balancing selection must have contributed to maintenance of high levels of polymorphisms. This finding was also supported by high π and FST values obtained for both Asian and European pears (Additional file 1: Table S7).
We speculated that a fast evolution might help to account for the wide variability observed in the S-RNase gene. Therefore, the evolution rates of S-RNase and other genes under balancing selection were compared. The evolution rate of the S-RNase gene of pear was estimated to be at least 1.91e− 09 sites/year, whereas those for other genes under balancing selection ranged between 2.31e− 10 and 6.10e− 10 sites/year. Therefore, the evolution rate of the S-RNase gene remains higher than the estimated evolution rates of other balance-selected genes (Additional file 1: Note 3). These findings support the hypothesis that rapid evolution of the S-RNase gene may have led to its high variability, which is consistent with the theory that reproduction-related genes show higher evolution rates [39, 40]. This has likely contributed to strengthening of GSI and promoting outcrossing, thus facilitating genetic recombination among genotypes of different genetic backgrounds of pear.
Independent domestication processes for each of Asian and European pears
Human intervention via artificial selection of favorable phenotypic traits to enhance production and improve desirable agronomic traits can both reduce levels of genetic variability and skew allele frequencies . Separate selective sweeps driven by artificial selection were detected in Asian and European pears. For Asian pears, selective sweep signatures for a total of 9.29 Mb of the genome sequence, containing 857 putative genes, were detected. For European pears, there were selective sweep signatures for 5.35 Mb of the genome sequence containing 248 putative genes (Fig. 5a-d; Additional file 5). It was notable that there was only 515 kb of overlap for regions with selective sweep signatures between Asian and European pear genomes, containing 47 putative genes (Additional file 6).
The different genes identified in selective sweeps of Asian and European pears were found to be enriched for 47 and 34 biological processes, respectively, including growth, response to cold, meristem and flower development, and single-organism metabolic processes, which could be involved in the distinct domestication pathways that have contributed to different traits of Asian and European pears. For example, in Asian pears, 11 cell wall degradation-related genes were found in selective sweep regions (Additional file 7), while none were found in selective sweep regions of European pears. These domestication-related genes might contribute to the crisp fruit flesh texture observed in Asian pears, compared with the soft and fine flesh texture of European pears. Four genes associated with fruit size, including one YABBY (Pbr003157.1; Additional file 1: Figure S4), two cyclin-like genes (Pbr015160.2 and Pbr028956.1), and one EXP4 (expansin-A4-like, Pbr041772.1), were found in selective sweeps of Asian pears. In contrast, two different fruit size-related genes, Pbr012098.1 and Pbr012099.1, which are homologous to tomato fw2.2 , were found in selective sweeps of European pears. This result indicates that different genome regions were selected for fruit size in Asian and European pears.
Many sugar-related genes were found in selective sweeps, indicating a preference for sweet fruit during domestication. For Asian pears, a total of 45 sugar-related genes were identified in the selected regions, including four genes (Pbr000142.1, Pbr019272.1, Pbr018801.2, and Pbr030762.1) that encode enzymes (starch synthase, fructokinase, and invertase) involved in sugar metabolism, and three genes (Pbr013451.1, Pbr037348.1, and Pbr037349.1) that encode sorbitol and hexose transporters (Additional file 1: Figure S5). Comparatively, only 11 sugar-related genes were identified in the selected sweeps of European pears, including Pbr039977.1, encoding a sorbitol transporter. Of these genes, sorbitol transporter (SOT; Pbr013451.1), starch synthase (SS; Pbr000142.1), trehalose-phosphatase (TPP; Pbr000530.1), and endoglucanase (EG; Pbr030853.1) were expressed at higher levels in the ripe fruits of cultivars than in those of wild accessions (Fig. 6), indicating that these genes could be important candidates associated with sugar transport and biosynthesis in pear fruits. In contrast, fructokinase (FK; Pbr018801.2) displayed lower expression in the ripe fruits of cultivars than in those of wile accessions, leading to a reduction of D-fructose transformed to β-D-fructose-6P, and thus promoting sugar accumulation in the fruits of cultivars.
Fewer genes involved in organic acid metabolism were identified in selective sweeps. The citric acid-related gene Pbr014969.1 (Additional file 1: Figure S6), a homolog of ACLA-3 that controls the synthesis of citrate , was identified in Asian pears, whereas three malate acid biosynthesis or transport-related genes, Pbr013232.1, Pbr013272.1, and Pbr030186.1, were identified in European pears. These findings further support an earlier observation regarding the presence of different dominant acid components in Asian and European pears .
Stone cells specifically accumulate in pear flesh and can detract from eating quality. The stone cell content shows a close relationship with biosynthesis, transfer, and deposition of lignin in cell walls. Six genes related to lignin biosynthesis in the selective sweep regions (Additional file 1: Figure S7), including two peroxidases (POD; Pbr000146.1 and Pbr15965.1), two hydroxycinnamoyl transferases (HCT; Pbr006408.1 and Pbr012356.2), one ferulate-5-hydroxylase (F5H; Pbr031416.1), and one cinnamoyl-CoA reductase (CCR; Pbr039962.1), were found in Asian pears, while only two CCR genes (Pbr013290.1 and Pbr039962.1) were detected in European pears. This finding may help to explain the higher concentration of stone cells in Asian pears than European pears.
As for volatile compounds in pear fruit, a total of 12 and three genes annotated in fatty acid metabolism pathways were identified in selective sweeps of Asian and European pears, respectively. Four alcohol dehydrogenase genes (Pbr003230.1, Pbr027590.1, Pbr027591.1, and Pbr034873.1) related to aroma biosynthesis were detected in Asian pears, while only two (Pbr013212.1 and Pbr028181.1) were found in European pears, indicating that aroma in these two groups of pears is regulated by different genes in the metabolic pathway.
It is interesting to note that the 47 genes in selective sweeps of both Asian and European pears include one POD gene (Pbr013214.1) related to stone cell formation and one cyclin-like gene (Pbr035650.1) related to fruit size development, suggesting a certain degree of convergent domestication of fruit quality in Asian and European pears.
Since the selective sweep signatures could colocalize with many important agronomic traits, potentially some of which have been implicated in quantitative trait loci (QTL) studies. We looked for enrichment in selective sweep signatures and candidate QTL regions. The selected regions included previously reported QTL with accurate chromosome information from hybridized segregating populations of pear [45,46,47] and a newly constructed pear population (“Niikata” × “Hongxiangsu”; Additional file 8). We found 208 and 14 selective sweeps from Asian and European pears that overlapped, respectively, with QTL regions (permutation test P values of 1E-5 and 0.35, respectively), indicating significant colocalization signals in Asian pears. The weak colocalization signals in European pears might be due to the fact that the QTL used here were mostly identified from Asian pear, while few fruit related QTL have been reported in European pears (Fig. 5e). In the overlapping regions of QTL and selective sweeps, a total of 151 and 17 genes were identified in selective sweeps of Asian pears and European pears, respectively (Additional file 9). Among them, 94 genes were mapped to sugar-related QTL, 20 to fruit size, 14 to acidity, ten to firmness, 18 to fruit shape, and three to stone cell content. These results strongly support that genes with selective sweep signatures in QTL regions might play important roles to regulate the distinct phenotypic traits selected in Asian versus European pears.
In this study, we report on genome variation mapping of 113 wild and cultivated pear accessions, collected from worldwide germplasm material. Our findings provide insights that informed our proposed model for the divergence, dissemination, and independent domestication of Asian and European pears. A rapid LD decay was identified in pear, thus revealing a characteristic weak domestication process for this perennial fruit tree. Separate selective sweep signatures identified between Asian and European pears underlined the distinct phenotypic traits observed in pear, including fruit acidity, sugar, and stone cell content, among others. Population structure analysis provided new evidence to support the admixed genetic background of some pear species, which was likely driven by self-incompatibility. Furthermore, analysis of the nucleotide diversity of the S-RNase gene controlling self-incompatibility suggested that a potential mechanism which promoted outcrossing must have accounted for the extensive genome diversity observed in pear.
Finally, it bears repeating that this study offers an unprecedentedly large amount of genomic resources for wild and cultivated pears. This, alongside our identification of candidate genes in selective sweep regions and colocalized QTLs, will significantly contribute to efforts for genetic improvement and molecular breeding of pear. Further, these findings raise intriguing questions that will almost certainly set the stage for the next phase of global pear and perennial tree fruit research.
Sampling information and sequencing
In this study, a total of 113 accessions, belonging to 33 Pyrus species from 26 countries and spanning a wide geographic distribution, were collected and sequenced. This collection covers accessions from all five of the major cultivated species of pear and from most recognized wild species (Additional file 2 and Additional file 1: Figure S1).
Genomic DNA was extracted from leaves using the CTAB method. Paired-end DNA libraries with short inserts (~ 500 bp) were constructed according to the manufacturer’s instructions and sequenced using the HiSeq™ 2000 or Hiseq™ 4000 platforms (Illumina, USA). To retain reads of high quality, reads with fewer than 5% N (missing) bases and with fewer than 50% of bases of base quality < 5 were deemed as cleaned reads. All other reads were discarded.
Reference genome selection and SNP calling
First, to facilitate the selection of an appropriate reference genome, we performed comparison of two published pear genomes: the Asian pear genome ‘Dangshansuli’ and the European pear genome Bartlett [14, 15]. For the details of our assembly quality assessment based on the contig N50, the scaffold N50, and the scaffold size values, as well scaffold to chromosome anchoring ratios, see Additional file 1: Table S1. Second, to clarify differences between these two potential reference genomes, we conducted synteny analysis of both genomes and also used all-versus-all BLASTP (E-value less than 1e-5) analysis to identify orthologous genes of the two genomes. Here, an orthologous gene was defined as a positive reciprocal BLASTP hit between the two genomes. MCScanX  was used to analyze synteny blocks. Third, for SNP calling, we used ‘Dangshansuli’ as the reference genome and the following protocol: 1) SOAPaligner (version 2.22 beta)  was used to map cleaned reads to the pear reference genome ; 2) based on genome coordinates and following removal of potential PCR duplicates, alignments were used to build a consensus sequence for each accession using SOAPsnp (version 1.04) ; 3) further filtration was conducted to obtain an accurate genotype for each site in each accession using the following criteria: (a) the quality value should be more than 20, (b) the number of unique reads for a confirmed genotype should be higher than 2, and (c) the copy number for each site had to be less than 1.5; and 4) the confirmed credible genotype from all accessions for each site and biallelic SNPs with missing rates of less than 0.5 were deemed as SNP variants in the population. Further, to ensure that the variant mapping rates of divergent samples would not deleteriously affect the analysis of θπ, θw, and Tajima’s D and so on, SNPs that were present in the syntenic blocks and that had a missing rate of < 10% among the accessions for both Asian pears and European pears were selected to validate the findings, yielding results that were consistent with mapping to the ‘Dangshansuli’ Asian pear genome (Additional file 1: Note 1). This four-step process led us to ultimately select ‘Dangshansuli’ as the reference genome. A total of 510 SNP loci were randomly selected, and Primer 3 was used to design primers for PCR-based sequence verification. Following PCR amplification, fragments were Sanger sequenced by Invitrogen Inc. (USA).
Population genetics analysis
Genetic distances determined in analyses of 113 accessions and 57 wild accessions were calculated by sampling with replacement SNPs (200 times) using the p-distance method , and neighbor-joining trees were constructed using the neighbor program in the EMBOSS toolbox . Trees were then merged, and Figtree (http://tree.bio.ed.ac.uk/software/figtree/) was used to adjust the neighbor-joining tree. A principal component analysis (PCA) was conducted using the eigen function in R base to obtain an eigenvector. The top four eigenvectors of samples were plotted using the ggplot2 R package . FRAPPE  was used to infer the population structure among samples, wherein the maximum iteration time was set to 10,000, and the number of population groups (K) was varied from 2 to 5. To determine the most appropriate population structure’s classification for all 113 accessions, FRAPPE analysis  was performed 20 times on 1000 randomly selected SNPs at 4dTv (four-fold degenerate site) for each K value from 1 to 10 according to Evanno G et al. .
To estimate the gene flow between Asian and European pears, both wild and cultivated pears, 4.7 M SNPs were selected with the following criteria: the missing rate was < 0.9 in both Asian and European pears. Based on these SNPs, we used Treemix version 1.13 to investigate the gene flow between groups/subgroups, with the settings: “-se -bootstrap -k 500 -m”, wherein the number (−m) varied from 1 to 5.
Nucleotide diversity analyses were conducted, including the average pairwise divergence within a population (θπ) , the Watterson’s estimator (θw) , and Tajima’s D . A sliding window of 10 kb, along with a step of 5 kb, was used to estimate the θπ, θw, and Tajima’s D values. For each window, these values were calculated using an in-house Perl script with the Bio::PopGen package. Pairwise FST values  were computed in the same windows to measure the population differentiation between groups. We also calculated the nucleotide diversity for various types of genomic regions (mRNA, CDS, introns, UTRs, and intergenic regions).
LD and LD blocks
Correlation coefficients (r2) of alleles were calculated using Haploview  to measure LD values in each of the four pear populations (i.e., Asian cultivated, Asian wild, European cultivated, and European wild). The parameters were set as follows: -maxdistance 200, -dprime, -minGeno 0.6, -minMAF 0.05, and -hwcutoff 0.01. LD decays were then plotted using a custom R script for each of the four pear populations. The parameter “-blockoutput GAB” was added to the Haploview program to detect LD blocks for each of the four pear populations.
Identification of identical-by-descent segments between Asian and European pear
Using pairwise accessions, identical-by-descent (IBD) regions were identified in contiguous 10-kb windows with no overlaps. The number of SNPs in each window should be more than 10. Similarity scores were calculated using the p-distance method  in each of the windows. Windows with percent similarity scores higher than 95% were deemed as IBD windows. The percentage of IBD windows along the entire genome was calculated for every pair of accessions.
Based on the geographical origins of accessions used in this study, wild accessions were separated into the following five wild geographic groups: Asian group II accessions, Asian group III accessions, Central Asian accessions, Western Asian accessions, and European mainland accessions. The average percentage of IBD values in the sliding windows (AverIBD) were calculated for each of the five wild geographic groups using the following formula:
where n corresponds to the number of accessions for a geographical group, m corresponds to the number of all Asian wild pear accessions, P ij is the percentage of IBDs in the genome for a pair of accessions (each from a different groups), and k is the count of P ij between a geographical group and all Asian wild pear accessions.
Self-incompatibility gene analysis
By mapping all pear sequencing reads to identified S-RNase alleles using bwa version 0.7.12-r1039) , cleaned reads specific for S-RNase genes were identified and used to call SNPs using the GATK package Haplotype Caller . Based on these called SNPs, θw and θπ, which indicate the nucleotide diversity, of S-RNase genes were calculated for different pear groups. The evolution rate of S-RNase and other genes under balancing selection were calculated using the d/2 T formula, where d is the nucleotide diversity and T is the duration of the time since divergence from the most recent common ancestor.
Divergence time of Asian and European pears
Various plant species, including Vitis vinifera (common European grape) , Carica papaya (papaya) , Fragaria vesca (woodland strawberry) , Prunus persica (peach) , Malus x domestica (cultivated apple) , Arabidopsis thaliana , and Populus trichocarpa (black cottonwood) , were used to estimate the divergence time of cultivated Chinese pears (P. bretschneideri)  from cultivated European pears (Pyrus communis) . A total of 420 single-copy gene families in all nine species were identified. Based on 4dTv (four-fold degenerate sites) in these 420 single-copy gene families, a phylogenetic tree was constructed using PhyML (v3.0) . Based on this phylogenetic tree and known divergence time range between Populus trichocarpa and Arabidopsis thaliana (100–120 MYA), we used MCMCTREE (PAML version 4.l4) to estimate the divergence time between cultivated Asian and cultivated European pears [68, 69].
Selective signals in pear
First, we removed the admixed genotypes (Additional file 1: Table S8) based on population structure analysis (Fig. 2a). Then, for Asian pears, cultivated accessions of P. pyrifolia and P. bretschneideri were used to detect selective sweeps. This strategy was used due to the admixture nature of P. sinkiangensis and availability of a limited number of P. ussuriensis accessions (Additional file 1: Table S8). Finally, 89 pear accessions including 19 wild Asian, 22 cultivated Asian, 24 wild European, and 24 cultivated European accessions (Additional file 10) were used for selective sweep analysis.
SNPs with missing rates of less than 0.5 in both Asian and European pears were deemed as common SNPs for selection sweep analysis. To identify regions with signals for selective sweeps in cultivated pears, θπ, Tajima’s D, reduction of diversity (ROD = 1 − θπcul/θπwild), and FST parameters were calculated in non-overlapping windows of 10 kb along the entire pear genome, based on common SNPs. Regions (10-kb window) with signals for selective sweeps were identified using the following criteria: among the top 5% of FST, ROD > 0.5, and bottom 10% of Tajima’s D distribution. Regions with balancing selection were identified using the bottom 5% of FST, and top 5% of Tajima’s D, and the top 10% of θπ.
RNA sequencing and sequence mapping
RNA was extracted from fruit flesh for a total of 24 samples (eight species × three stages). RNA sequencing libraries were constructed using the Illumina standard mRNA-Seq Prep Kit (TruSeq RNA and DNA Sample Preparation Kits version 2). Single end RNA-Seq data were generated with length of 49 bp. Reads were filtered and trimmed and then mapped onto ‘Dangshansuli’ (Pyrus bretschneideri) coding sequences using SOAPaligner .
Sugar content measurements
High performance liquid chromatography (HPLC) was used to measure pear fruit sugars, including sucrose, glucose, fructose, and sorbitol. Sugars were extracted from pear flesh by grinding, and then were dissolved and filtered through a SEP-C18 cartridge (Waters, WAT021515) and Sep-Pak filter. Sugars were processed using a Waters 1525 system (Waters, Shanghai, China); the column was 6.5 mm × 300 mm, inner diameter, 10 μm particle size (Waters), with a Sugar-pak 1 Guard-Pak Holder and Insert (Waters) cartridge for the guard column. Column temperature was set to 85 °C, and 35 °C was the reference cell temperature.
Association of selective sweep regions with QTLs
Selective sweep regions were associated with QTLs identified in pears, including previously published QTLs and new QTLs related to fruit quality that we identified in the course of the present study. Two F1 populations were used for these new QTL mapping analyses, including an F1 population containing 102 individuals derived from crossing ‘Bayuehong’ × ‘Dangshansuli’ (phenotyping of different fruit traits, including sugar content, acid content, stone cell content, and fruit size, were conducted in 2014 and 2015) and an F1 pear population of 176 individuals from a cross between ‘Niikata’ × ‘Hongxiangsu’ using an 8× re-sequencing strategy (phenotyping of fruit-related traits such as sugar content, acidity content, stone cell content), which were investigated in 2015 and 2016.
MapQTL6.0 (https://www.kyazma.nl/index.php/MapQTL/) was used for linkage map construction using a regression mapping algorithm and the Kosambi function. MapQTL6.0 was also used to perform interval mapping and to conduct MQM and Kruskal-Wallis tests to evaluate candidate QTLs. Finally, markers with p < 0.005 and interval mapping LOD values higher than 3.5 were identified as QTLs. In comparison with the overlap of selective sweep regions and QTLs, we used enrichment tests with a sliding window size of 10 kb with 100,000 repetitions throughout the genome to find the overlap regions with candidate QTL regions.
Potter D, Eriksson T, Evans RC, Oh S, Smedmark JEE, Morgan DR, et al. Phylogeny and classification of Rosaceae. Plant Syst Evol. 2007;266:5–43.
Bell RL, Quamme HA, Layne REC, Skirvin RM. Pears. In: Janick J, Moore JN, editors. Fruit breeding. New York: John Wiley and Sons; 1996. p. 441–514.
Westwood MN, Bjornstad HO. Some fruit characteristics of interspecific hybrids and extent of self-sterility in Pyrus. Bull Torrey Botanical Club. 1971;98:22–4.
Westwood M, Challice J. Morphology and surface topography of pollen and anthers of Pyrus species. J Am Soc Hort Sci. 1978;103:28–37.
Jang JT, Tanabe K, Tamura F, Banno K. Identification of Pyrus species by leaf peroxidase isozyme phenotypes. J Jpn Soc Hortic Sci. 1992;61:273–86.
Rubtsov GA. Geographical distribution of the genus Pyrus and trends and factors in its evolution. Am Nat. 1944;78:358–66.
Kikuchi A. Speciation and taxonomy of Chinese pears. Collected Records Hortic Res. 1946;3:1–8.
Pieniazek SA. Fruit production in China. In: Proceedings of the XVII International Horticulture Congress, 23–26 July, East Cansing. Michigan; 1967. p. 427–56.
Hedrick UP, Howe G, Taylor OM, Francis EH, Tukey HB. The pears of New York. J Pomol Hort Sci. 1924;3:153–5.
Fan L, Zhang MY, Liu QZ, Li LT, Song Y, Wang LF, et al. Transferability of newly developed pear SSR markers to other Rosaceae species. Plant Mol Biol Rep. 2013;31:1271–82.
Song Y, Fan L, Chen H, Zhang M, Ma Q, Zhang S, et al. Identifying genetic diversity and a preliminary core collection of Pyrus pyrifolia cultivars by a genome-wide set of SSR markers. Sci Hortic Amsterdam. 2014;167:5–16.
Zheng X, Cai D, Potter D, Postman J, Liu J, Teng Y. Phylogeny and evolutionary histories of Pyrus L. revealed by phylogenetic trees and networks based on data from multiple DNA sequences. Mol Phylogen Evol. 2014;80:54–65.
Llegren H. Genome sequencing and population genomics in non-model organisms. Trends Ecol Evol. 2014;29:51–63.
Wu J, Wang Z, Shi Z, Zhang S, Ming R, Zhu S, et al. The genome of the pear (Pyrus bretschneideri Rehd.). 23:396–408. Genome Res. 2013;23:396–408. https://www.ncbi.nlm.nih.gov/genome/12793
Chagné D, Crowhurst RN, Pindo M, Thrimawithana A, Deng C, Ireland H, et al. The draft genome sequence of European pear (Pyrus communis L.‘Bartlett’). PLoS One. 2014;9:e92644. Genome Databse for Rosaceae, URL https://www.rosaceae.org/species/pyrus/pyrus_communis
Lam HM, Xu X, Liu X, Chen W, Yang G, Wong F-L, et al. Resequencing of 31 wild and cultivated soybean genomes identifies patterns of genetic diversity and selection. Nat Genet. 2010;42:1053–9.
Zhou Z, Jiang Y, Wang Z, Gou Z, Lyu J, Li W, et al. Resequencing 302 wild and cultivated accessions identifies genes related to domestication and improvement in soybean. Nat Biotechnol. 2015;33:408–14.
Hufford MB, Xu X, Van Heerwaarden J, Pyhäjärvi T, Chia J-M, Cartwright RA, et al. Comparative population genomics of maize domestication and improvement. Nat Genet. 2012;44:808–11.
Cao K, Zheng Z, Wang L, Liu X, Zhu G, Fang W, et al. Comparative population genomics reveals the domestication history of the peach, Prunus persica, and human influences on perennial fruit crops. Genome Biol. 2014;15:415.
Wang W, Feng B, Xiao J, Xia Z, Zhou X, Li P, et al. Cassava genome from a wild ancestor to cultivated varieties. Nat Commun. 2014;5:5110.
Lijavetzky D, Antonio Cabezas J, Ibanez A, Rodriguez V, Martinez-Zapater JM. High throughput SNP discovery and genotyping in grapevine (Vitis vinifera L.) by combining a re-sequencing approach and SNPlex technology. BMC Genomics. 2007;8:424.
Hazzouri KM, Flowers JM, Visser HJ, Khierallah HSM, Rosas U, Pham GM, et al. Whole genome re-sequencing of date palms yields insights into diversification of a fruit tree crop. Nat Commun. 2015;6:8824.
Xu X, Liu X, Ge S, Jensen JD, Hu F, Li X, et al. Resequencing 50 accessions of cultivated and wild rice yields markers for identifying agronomically important genes. Nat Biotechnol. 2012;30:105–11.
Doebley JF, Gaut BS, Smith BD. The molecular genetics of crop domestication. Cell. 2006;127:1309–21.
Duan N, Bai Y, Sun H, Wang N, Ma Y, Li M, et al. Genome re-sequencing reveals the history of apple and supports a two-stage model for fruit enlargement. Nat Commun. 2017;8:249.
Slatkin M. Linkage disequilibrium-understanding the evolutionary past and mapping the medical future. Nat Rev Genet. 2008;9:477–85.
Bao L, Chen K, Zhang D, Cao Y, Yamamoto T, Teng Y. Genetic diversity and similarity of pear (Pyrus L.) cultivars native to East Asia revealed by SSR (simple sequence repeat) markers. Genet Resour Crop Evol. 2007;54:959–71.
Liu Q, Song Y, Liu L, Zhang M, Sun J, Zhang S, et al. Genetic diversity and population structure of pear (Pyrus spp.) collections revealed by a set of core genome-wide SSR markers. Tree Genet Genom. 2015;11:128.
Dondini L, Sansavini S. European pear. In: Badenes ML, Byrne DH, editors. Fruit breeding. New York: Springer; 2012. p. 369–413.
Boulnois L. Silk road: monks, warriors & merchants on the Silk Road. New York: WW Norton & Co Inc; 2004.
Prevas J. Envy of the gods: Alexander the Great's ill-fated journey across Asia. Boston: Da Capo Press; 2005.
Wu J, Wang YT, Xu JB, Korban SS, Fei ZJ, Tao ST, et al. Diversification and independent domestication of Asian and European pears. BioProject. 2018; https://www.ncbi.nlm.nih.gov/search/?term=PRJNA381668
Wu J, Wang YT, Xu JB, Korban SS, Fei ZJ, Tao ST, et al. Diversification and independent domestication of Asian and European pears. Pear Genome Project. 2018; http://peargenome.njau.edu.cn/default.asp?d=4&m=2
Huang X, Kurata N, Wang Z-X, Wang A, Zhao Q, Zhao Y, et al. A map of rice genome variation reveals the origin of cultivated rice. Nature. 2012;490:497–501.
Dai F, Nevo E, Wu D, Comadran J, Zhou M, Qiu L, et al. Tibet is one of the centers of domestication of cultivated barley. Proc Natl Acad Sci U S A. 2012;109:16969–73.
Kakui H, Kato M, Ushijima K, Kitaguchi M, Kato S, Sassa H. Sequence divergence and loss-of-function phenotypes of S locus F-box brothers genes are consistent with non-self recognition by multiple pollen determinants in self-incompatibility of Japanese pear (Pyrus pyrifolia). Plant J. 2011;68:1028–38.
Heng W, Wu HQ, Huang S, Zhang SJ, Wu J, Fang CQ, et al. Identification of S-genotypes and novel S-RNases in native Chinese pear. J Hortic Sci Biotechnol. 2008;83:629–34.
Zisovich AH, Stern RA, Sapir G, Shafir S, Goldway M. The RHV region of S-RNase in the European pear (Pyrus communis) is not required for the determination of specific pollen rejection. Sex Plant Reprod. 2004;17:151–6.
Schopfer CR, Nasrallah ME, Nasrallah JB. The male determinant of self-incompatibility in Brassica. Science. 1999;286:1697–700.
Swanson WJ, Vacquier VD. The rapid evolution of reproductive proteins. Nat Rev Genet. 2002;3:137–44.
Tajima F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics. 1989;123:585–95.
Frary A, Nesbitt TC, Frary A, Grandillo S, Van Der Knaap E, Cong B, et al. fw2.2: a quantitative trait locus key to the evolution of tomato fruit size. Science. 2000;289:85–8.
Fatland BL, Ke J, Anderson MD, Mentzen WI, Cui LW, Allred CC, et al. Molecular characterization of a heteromeric ATP-citrate lyase that generates cytosolic acetyl-coenzyme A in Arabidopsis. Plant Physiol. 2002;130:740–56.
Yao GF, Yang ZJ, Zhang SL, Cao YF, Liu J, Wu J. Characteristics of components and contents of organic acid in pear fruit from different cultivated species. Sci Agric Sin. 2014;41:755–64.
Yamamoto T, Terakami S, Takada N, Nishio S, Onoue N, Nishitani C, et al. Identification of QTLs controlling harvest time and fruit skin color in Japanese pear (Pyrus pyrifolia Nakai). Breeding Sci. 2014;64:351–61.
Kumar S, Kirk C, Deng C, Wiedow C, Knaebel M, Brewer L. Genotyping-by-sequencing of pear (Pyrus spp.) accessions unravels novel patterns of genetic diversity and selection footprints. Hortic Res. 2017;4:17015.
Wu J, Li LT, Li M, Khan MA, Li XG, Chen H, et al. High-density genetic linkage map construction and identification of fruit-related QTLs in pear using SNP and SSR markers. J Exp Bot. 2014;65:5771–581.
Wang Y, Tang H, JD DB, Tan X, Li J, Wang X, et al. MCScanX: a toolkit for detection and evolutionary analysis of gene synteny and collinearity. Nucleic Acids Res. 2012;40:e49.
Li R, Yu C, Li Y, Lam TW, Yiu SM, Kristiansen K, et al. SOAP2: an improved ultrafast tool for short read alignment. Bioinformatics. 2009;25:1966–7.
Li R, Li Y, Fang X, Yang H, Wang J, Kristiansen K, et al. SNP detection for massively parallel whole-genome resequencing. Genome Res. 2009;19:1124–32.
Rice P, Longden I, Bleasby A. EMBOSS: the European Molecular Biology Open Software Suite. Trends Genet. 2000;16:276–7.
Ginestet C. ggplot2: elegant graphics for data analysis. J R Stat Soc A Sta. 2011;174:245.
Tang H, Peng J, Wang P, Risch NJ. Estimation of individual admixture: analytical and study design considerations. Genet Epidemiol. 2005;28:289–301.
Evanno G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14:2611–20.
Watterson GA. On the number of segregating sites in genetical models without recombination. Theor Popul Biol. 1975;7:256–76.
Weir BS, Cockerham CC. Estimating F-statistics for the analysis of population structure. Evolution. 1984;38:1358–70.
Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21:263–5.
Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–60.
McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20:1297–303.
Jaillon O, Aury JM, Noel B, Policriti A, Clepet C, Casagrande A, et al. The grapevine genome sequence suggests ancestral hexaploidization in major angiosperm phyla. Nature. 2007;449:463–7. NCBI Genome, URL https://www.ncbi.nlm.nih.gov/genome/401
Ming R, Hou S, Feng Y, Yu Q, Dionne-Laporte A, Saw JH, et al. The draft genome of the transgenic tropical fruit tree papaya (Carica papaya Linnaeus). Nature. 2008;452:991–6. NCBI Genome, URL https://www.ncbi.nlm.nih.gov/genome/513
Shulaev V, Sargent DJ, Crowhurst RN, Mockler TC, Folkerts O, Delcher AL, et al. The genome of woodland strawberry (Fragaria vesca). Nat Genet. 2011;43:109–16. Genome Databse for Rosaceae, URL https://www.rosaceae.org/organism/Fragaria/vesca
Verde I, Abbott AG, Scalabrin S, Jung S, Shu S, Marroni F, et al. The high-quality draft genome of peach (Prunus persica) identifies unique patterns of genetic diversity, domestication and genome evolution. Nat Genet. 2013;45:487–94. Genome Databse for Rosaceae, URL https://www.rosaceae.org/organism/Prunus/persica
Velasco R, Zharkikh A, Affourtit J, Dhingra A, Cestaro A, Kalyanaraman A, et al. The genome of the domesticated apple (Malus× domestica Borkh.). Nat Genet. 2010;42:833–9. Genome Databse for Rosaceae, URL https://www.rosaceae.org/organism/Malus/x-domestica
The Arabidopsis Genome Initiative. Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature. 2000;408:796–815. NCBI Genome, URL https://www.ncbi.nlm.nih.gov/genome/794
Tuskan GA, DiFazio S, Jansson S, Bohlmann J, Grigoriev I, Hellsten U, et al. The genome of black cottonwood, Populus trichocarpa (Torr. & Gray). Science. 2006;313:1596–604. NCBI Genome, URL https://www.ncbi.nlm.nih.gov/genome/1598
Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol. 2010;59:307–21.
Rannala B, Yang Z. Inferring speciation times under an episodic molecular clock. Syst Biol. 2007;56:453–66.
Yang ZH, Rannala B. Bayesian estimation of species divergence times under a molecular clock using multiple fossil calibrations with soft bounds. Mol Biol Evol. 2006;23:212–26.
This work was funded by the National Natural Science Foundation of China (31725024, 31471839, 31672111, and 31772276), the Earmarked Fund for China Agriculture Research System (CARS-28), the Science Foundation for Distinguished Young Scholars in Jiangsu Province (BK20150025), and the “333 High Level Talents Project” of Jiangsu Province (BRA2016367), and the US National Science Foundation (IOS-1539831).
Availability of data and materials
All of the raw reads generated in this study have been deposited in the NCBI database under BioProject accession PRJNA381668 . The variation datasets are available through the pear genome database website  or are freely available upon request. Supplementary information and source data files are available with the online version of this article. European wild pear accessions used in this study have been kindly provided to us by the USDA-ARS National Clonal Repository in Corvallis, Oregon following our request, and Asian wild pear accessions were collected by Nanjing Agricultural University, all those are publically available.
The authors declare no competing financial interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary data and analysis. (DOC 12492 kb)
Summary of sequencing and mapping of 113 pear accessions. (XLSX 51 kb)
SNPs and genotype verification by Sanger sequencing. (XLSX 76 kb)
S-genotyping for different accessions. (XLSX 20 kb)
List of genes in selective sweeps in Asian and European groups during domestication from wild to cultivated pears. (XLSX 188 kb)
List of 47 genes in common selective sweeps of both Asian and European groups during domestication from wild to cultivated pears. (XLSX 39 kb)
List of genes in selective sweep regions involved in important traits during domestication from wild to cultivated pears in Asian group. (XLSX 60 kb)
List of QTL regions and association SNPs. (XLSX 48 kb)
List of domesticated genes within QTL regions and association SNPs. (XLSX 62 kb)
List of genes in selective sweep regions of P. ussuriensis during domestication from wild to cultivated pears. (XLSX 44 kb)