Population structure and genetic bottleneck in sweet cherry estimated with SSRs and the gametophytic self-incompatibility locus

Abstract

Background: Domestication and breeding involve the selection of particular phenotypes, limiting the genomic diversity of the population and creating a bottleneck. These effects can be precisely estimated when the location of domestication is established. Few analyses have focused on understanding the genetic consequences of domestication and breeding in fruit trees. In this study, we aimed to analyse genetic structure and changes in the diversity in sweet cherry Prunus avium L.

Results: Three subgroups were detected in sweet cherry, with one group of landraces genetically very close to the analysed wild cherry population. A limited number of SSR markers displayed deviations from the frequencies expected under neutrality. After the removal of these markers from the analysis, a very limited bottleneck was detected between wild cherries and sweet cherry landraces, with a much more pronounced bottleneck between sweet cherry landraces and modern sweet cherry varieties. The loss of diversity between wild cherries and sweet cherry landraces at the S-locus was more significant than that for microsatellites. Particularly high levels of differentiation were observed for some S-alleles.

Conclusions: Several domestication events may have happened in sweet cherry or/and intense gene flow from local wild cherry was probably maintained along the evolutionary history of the species. A marked bottleneck due to breeding was detected, with all markers, in the modern sweet cherry gene pool. The microsatellites did not detect the bottleneck due to domestication in the analysed sample. The vegetative propagation specific to some fruit trees may account for the differences in diversity observed at the S-locus. Our study provides insights into domestication events of cherry, however, requires confirmation on a larger sampling scheme for both sweet cherry landraces and wild cherry.

Figure 1

Structure bar plot results obtained on the whole set of data at K = 2. 1 is population identification for modern varieties, 2 is population identification for landraces and 3 is population identification for wild cherries.

Figure 2

Structure bar plot results obtained on the whole set of data at K = 3. 1 is population identification for modern varieties, 2 is population identification for landraces and 3 is population identification for wild cherries.

Figure 3

Structure bar plot results obtained on the wild cherry at K = 4

Figure 4

Distribution of F_ST values as a function of expected heterozygosity (H) using the F_ST between modern 1 and landrace 1 groups (0.059099). The envelope of values corresponding to neutral expectations with the infinite allele model was constructed as described by [53]. Dotted lines with plain circles represent the 0.5(1 - 0.95) and 0.5(1 + 0.95) quantiles. Dotted lines and circles represent the median of values. Triangles represent the observations not significant at 5%. Circles represent the observations significant at 5%.

Figure 5

Distribution of F_ST values as a function of expected heterozygosity (H) using the F_ST between modern 2 and landrace 2 groups (0.048377). The envelope of values corresponding to neutral expectations with the infinite allele model was constructed as described by [53]. Dotted lines with plain circles represent the 0.5(1 - 0.95) and 0.5(1 + 0.95) quantiles. Dotted lines and circles represent the median of values. Triangles represent the observations not significant at 5%. Circles represent the observations significant at 5%.

Figure 6

Distribution of F_ST values as a function of expected heterozygosity (H) using the F_ST between wild cherries and modern sweet cherry varieties (0.021650). The envelope of values corresponding to neutral expectations with the infinite allele model was constructed as described by [53]. Dotted lines with plain circles represent the 0.5(1 - 0.95) and 0.5(1 + 0.95) quantiles. Dotted lines and circles represent the median of values. Triangles represent the observations not significant at 5%. Circles represent the observations significant at 5%.