Genetic structure and linkage disequilibrium in four Vitis species

Corresponding author: doligez@supagro.inra.frInternational audienceGenome-wide association (GWA) genetics is potentially of great interest in grapevine to much more rapidly and efficiently find the main loci (QTLs) and alleles involved in the variation of traits of interest to breeding, compared to searches based on individual segregating progenies. Knowledge of the structure of genetic diversity and linkage disequilibrium (LD) is a necessary prerequisite to design association genetics studies. Vitis vinifera genetic diversity and structure are already well characterized, but LD in this species has not been thoroughly estimated yet. Other Vitis species are a large potential source of resistance to biotic and abiotic stresses for introgression into V. vinifera, but relatively few genetic resources are available in germplasm repositories and the genetic diversity of these species is hardly known. The objectives of the DLVitis project (ANR 2009-2011) are to assess genetic structure and linkage disequilibrium in V. vinifera and 3 American species (V. riparia, V. cinerea, V. aestivalis). For V. vinifera, 3 samples of 93 cultivars were defined, based on available SSR data, by minimizing structure and kinship and maximizing genetic distances. An additional sample of 93 accessions of wild V. vinifera (subsp. silvestris) was also defined. For each American species, 150-200 seeds were collected in natura last autumn. Genetic diversity and structure will be assessed this year using SSRs, and a sample of 93 accessions will be defined for each species. Four genomic regions of 2 Mb each were chosen to estimate LD between specific SNPs in ca. 1 gene out of 3 in each sample of 93 individuals described above. Novel estimates of LD taking structure and kinship into account were developped and will be presented. The first results obtained in V. vinifera in 2 genomic regions showed differences in the rate of LD decay between samples and regions, suggesting past occurrence of differential selection. The consequences of these results for the design of GWA studies will be discussed