Genetic structure in cultivated grapevines is linked to geography and human selection

Abstract

Background: Grapevine (Vitis vinifera subsp. vinifera) is one of the most important and ancient horticultural plants in the world. Domesticated about 8-10,000 years ago in the Eurasian region, grapevine evolved from its wild relative (V. vinifera subsp. sylvestris) into very diverse and heterozygous cultivated forms. In this work we study grapevine genetic structure in a large sample of cultivated varieties, to interpret the wide diversity at morphological and molecular levels and link it to cultivars utilization, putative geographic origin and historical events.

Results: We analyzed the genetic structure of cultivated grapevine using a dataset of 2,096 multi-locus genotypes defined by 20 microsatellite markers. We used the Bayesian approach implemented in the STRUCTURE program and a hierarchical clustering procedure based on Ward's method to assign individuals to sub-groups. The analysis revealed three main genetic groups defined by human use and geographic origin: a) wine cultivars from western regions, b) wine cultivars from the Balkans and East Europe, and c) a group mainly composed of table grape cultivars from Eastern Mediterranean, Caucasus, Middle and Far East countries. A second structure level revealed two additional groups, a geographic group from the Iberian Peninsula and Maghreb, and a group comprising table grapes of recent origins from Italy and Central Europe. A large number of admixed genotypes were also identified. Structure clusters regrouped together a large proportion of family-related genotypes. In addition, Ward's method revealed a third level of structure, corresponding either to limited geographic areas, to particular grape use or to family groups created through artificial selection and breeding.

Conclusions: This study provides evidence that the cultivated compartment of Vitis vinifera L. is genetically structured. Genetic relatedness of cultivars has been shaped mostly by human uses, in combination with a geographical effect. The finding of a large portion of admixed genotypes may be the trace of both large human-mediated exchanges between grape-growing regions throughout history and recent breeding.

Figure 1

Similarity index among STRUCTURE runs. Similarity index among runs for each STRUCTURE K-level, and its confidence intervals (10 runs for each K).

Figure 2

Characterization of the STRUCTURE groups. Characterization of the STRUCTURE groups according to geography and use.

Figure 3

Genetic composition of the geographic groups. Genetic composition of the geographic groups for the Ks = 3 of STRUCTURE. For the detailed country list, see Table 1. The histograms represent the percentage of non-admixed (green) versus admixed (orange) genotypes. For the non-admixed cultivars, the pies represent the proportion of each cluster in each region: Table / East (yellow); Wine / Balkans and East-Europe (violet); Wine / West and Central Europe (blue). As 100% of the Italian genotypes are admixed, the ITAP pie is empty (grey).

Figure 4

Principal component analysis on SSR data. Principal component analysis on SSR data of genotypes belonging to Ks = 5. Colors of the groups correspond to the colors in Figure 2. Axes 1 and 2 explain 30.3% and 21.4% of the total variance respectively (black vertical bars in the eigenvalues histogram).

Figure 5

Dendrogram based on Ward’s clustering. Dendrogram based on Ward’s clustering. For levels three and five of clustering, the comparison among the Ward and STRUCTURE groups is summarized with the double code labels (W-x.x / S-x.x), and the percentage of shared individuals between them.