Given that both methods capture genetic relatedness among accessions, a significant relationship between these two methods is expected.Principal component analysis revealed significant population structure defined by geography and use, but strong signals of genetic structure also exist at the level of pedigree relatedness. Previous work determined that 75% of the accessions evaluated here are related to at least one other accession by a first-degree relationship, and over half of the accessions are interrelated and form a single, complex pedigree network.Both the strong population-level and pedigree-level signals of genetic structure in our sample present challenges in genetic mapping as these are significant confounding factors when performing GWAS. Moreover, the rapid LD decay previously described for this and other diverse populations of V. vinifera suggests that millions of SNPs are required for well-powered GWAS in grapes.Despite our relatively low marker density and the challenges presented by strong genetic structure, we performed GWAS for all 33 phenotypes. For most traits, we found no convincing GWAS signals . However, we reasoned that we may find SNPs associated with key traits that experienced strong selection during domestication and breeding because selection results in extended LD surrounding the targeted loci, thereby requiring a lower SNP density than that required to map-unselected traits. We hypothesized that, by combining association mapping with selective sweep mapping ,macetas con drenaje we may identify loci associated with traits targeted during grape domestication and breeding. A key transition in grapevine domestication was the switch from dioecy to hermaphroditism: all wild Vitis species, including the ancestor of V. vinifera, are dioecious, and nearly all V. vinifera are hermaphroditic.

Hermaphroditism was likely the first, and arguably the most important, transition from wild vines to cultivated grapes: it enables self-pollination and subsequent clonal propagation of elite cultivars without the need for pollinators.Dioecy is found at low frequency in our sample: only 50 of the 550 accessions with flower sex data were labeled as dioecious. Despite this low frequency, we identified SNPs significantly associated with flower sex on chromosome 2 . The most significantly associated SNP overlaps with the 1.5 Mb region repeatedly identified via linkage mapping.This SNP is also found within the fine-mapped 143 kb region believed to harbor the causal flower sex locus.We therefore demonstrate that, even with only 50 accessions carrying the ancestral dioecy phenotype, we successfully map the flower sex locus at relatively high resolution using GWAS relative to traditional linkage mapping approaches. A genome-wide Fst scan comparing dioecious to hermaphroditic accessions also revealed that the SNP most strongly associated with flower sex had the highest Fst value genomewide, consistent with the effect of selection for hermaphroditism at this locus . If grape domestication resulted in a rapid increase in the frequency of the hermaphroditism allele, one would expect extended haplotype homozygosity, and thus extremely high xpEHH values, in and around the flower sex locus. While none of the xpEHH values at the flower sex locus fall within the top 1% most extreme values genome-wide, we do observe a suggestive peak with xpEHH values within the top 2.6% of genome-wide xpEHH values . xpEHH values in the bottom 1% of the genome-wide distribution are found directly adjacent to the flower sex locus identified here. We have no explanation for why a potential signature of selection could exist for dioecy in such close proximity to the flower sex locus. There are SNPs with extreme xpEHH and Fst values, indicating potential selection for hermaphroditism, at the distal end of chromosome 1 ranging from positions 366 to 467 kb . This genomic region overlaps with the region previously associated with flower sex in a bi-parental mapping population using the same Vitis9KSNP microarray employed for this study.

However, thisregion is several Mb from the locus highlighted in Figure 6a that color within a diffuse peak on chromosome 2 between 10 and has been repeatedly associated with flower sex. We hypothesize 17 Mb . Although the genomic region containing that this distal signal of selection is due to inaccurate localization significant GWAS hits for color overlaps the VvmybA1 gene, the of the array’s SNPs in the reference genome since, when this trait most significantly associated SNP found here is 3.6 Mb from the is mapped using genotyping-by-sequencing in the same bi- known causal mutation. Our inability to map the known color parental population, the flower sex colocalizes with the known locus with precision is consistent with results from rice and flower sex locus according to the reference genome.It is unclear Arabidopsis where markers with the strongest association why such mismapping occurs with the Vitis9KSNP array data, but signals were not found directly at known causative loci. Moreover, unexpected hybridization of non-targeted paralogous regions this result is unsurprising given the relatively low marker density may possibly contribute to these observations. of the SNP array employed here. Skin color in grapes is largely controlled by a single locus on While the diffuse association signal for grape color spanning chromosome 2, where a retrotransposon insertion in the MYBA1 nearly 7 Mb indicates that we have poor mapping resolution for gene results in a loss in pigmentation by disrupting anthocyanin this phenotype, it also suggests the presence of long-range LD biosynthesis.Although rare, white-skinned grapes have been potentially caused by selection. Saccharomyces cerevisiae is an important experimental model organism in addition to its commercial significance as the predominant yeast species during wine fermentation. Modern strains of S. cerevisiae are thought to have arisen in Asia given the diversity of strains and reproductive isolation observed in a study of S. cerevisiae isolates from human-associated and non-human-associated environments in China .

Distinct linages were observed for isolates from primeval and secondary forests . However this study considered few isolates from wine environments and found fewer isolates of S. cerevisiae from fruit sources and more from tree bark, rotting wood and soil samples than from fruit samples. The authors concluded that grape and orchard isolates were similar to those of the wine European lineage. Our goal was to evaluate in greater depth the diversity of natural vineyard isolates from two wine regions in China. Several studies have reported on the genetic diversity of S. cerevisiae strains in different wine-producing regions. These studies revealed that geographic region , climatic conditions , vintage , grape varieties and must characteristics , inoculation of starter yeasts , and SO2 addition affected the diversity of S. cerevisiae observed. In many cases genetically distinct strains of S. cerevisiae were isolated from the same fermentation during wine fermentation . The diversity of S. cerevisiae strains present in fermentations has been shown to play an important role in the characteristics of the final product . Numerous molecular methods have been developed to study the ecology and population dynamics of S. cerevisiae strains . Interdelta sequencing typing uses the variation of the number and position of the delta element, a repeated sequence that flanks the Ty1/Ty2 retrotransposon , that allows interpreting strain similarities and evolutionary or adaptive distance . A succession of different S. cerevisiae strains are established during native as well as inoculated fermentations that could have positive or negative effects on the course of fermentation and wine quality . Vezinhet et al. analyzed the evolution of S. cerevisiae strains isolated from spontaneous fermentations during six consecutive years. These authors concluded that the wide distribution of some strains in the studied areas and their presence over years,macetas 7 litros constitute evidence for the occurrence of specific indigenous strains representative of an enological region. China is an important wine-producing country and while some studies have investigated indigenous yeast species and population dynamics during wine fermentation within local viticulture regions ; few studies have focused on the breadth of the diversity of S. cerevisiae wine genetic resources of China. A study of human- and non-human-associated strains of China found novel distinct lineages only distantly related to the wine strain linages . The genetic diversity and relatedness of indigenous wine S. cerevisiae resources have not been extensively compared with that of wine strains isolated from other geographic regions. Ningxia and Xinjiang provinces, where the strains in this study were isolated, are two of the oldest wine producing regions in China. Shanshan, Xinjiang in northwestern China belongs to a temperate continental climate, with an average temperature of 12˚C. It is situated 92°22′E, 42°87′N with an average altitude of 3986 m.

Qing Tongxia, Ningxia in north central China also belongs to a temperate zone with an arid and semi-arid climate. It is situated 105°21′ to 105°21′ E, 37°36′ to 38°15′N with an average altitude of 1118 m. A comparison of the genetic diversity of S. cerevisiae resources in different viticulture regions of China with isolates from other diverse geographical regions is of importance to the study of global S. cerevisiae ecology. In the present study, interdelta sequence typing with improved primers was used as genetic marker for the distinction of S. cerevisiae strains. Dendrograms were constructed based on similarity among different patterns of bands and the genetic relationships of all strains were evaluated. The strains used in the study were either isolated from fermentations of different grape varieties in the Ningxia and Xinjiang Provinces in China or obtained from the Department of Viticulture and Enology Culture Collection at the University of California, Davis. The aims of the present work were to evaluate the genetic diversity and relatedness amongSaccharomyces strains of different geographic origin, to establish a strain collection to preserve the S. cerevisiae genetic resources of China, and to identify strains useful for further development for commercial wine production in China. Fifty-four isolates collected from fifteen spontaneous fermentations of grapes grown in China and one commonly used commercial yeast, Lavin RC212, were used in this study and obtained from the collections of the College of Enology, Northwest A&F University, Yangling, Shaanxi, China. This set of strains was selected on the basis of interdelta sequence profiles from a total of 349 isolates collected from fifteen spontaneous fermentations of grapes grown in Shanshan, Xinjiang and Qing Tongxia, Ningxia. Fifty-nine yeast colonies were isolated from six spontaneous fermentations of different commonly used grape varieties: Red Globe, Small-berry Thompson Seedless, Big-berry Thompson Seedless, Merlot, Mixed red and Mixed white in Xinjiang. Two hundred ninety isolates were obtained from nine spontaneous fermentations of the grape varieties Cabernet Gernischt, Cabernet Sauvignon, Cinsault, Merlot, Pinot Noir, Riesling, Sauvignon Blanc, Semillon, and Yan73 in Ningxia . The grape must fermentations were allowed to proceed spontaneously at 25~28˚C for 7~11 days until dry. Fermentations were sampled at early, mid and the final stage of fermentation, and serial ten-fold dilutions were inoculated onto WLN and incubated for five days at 28˚C.These yeasts were differentiated and classified according to colony morphology and color. S. cerevisiae isolates were purified and then maintained in 20% glycerol at -80 °C until further analysis, resulting in the selected set of 349 isolates for the interdelta sequence analysis. The composition of the different grape musts is reported in Supplemental Table 1 for Ningxia and in Supplemental Table 2 for the fermentations from Xinjiang. The fifty-four yeast strains selected from this larger population of isolates represented the major strain clusters of interdelta sequence profiles identified in the earlier preliminary study. The origins of the 54 isolates used in this study are shown in Table 1. Identification of S. cerevisiae was confirmed by PCR-RFLP of the 5.8S-ITS rDNA using restriction enzymes HaeIII, HpaII, and ScrFI as described by Li et al. . Strains were maintained in frozen stocks at -80˚C before use. Note that a similar strain numbering system was independently used by Wang et al. in their study but the strains are unrelated. We retained our numbering system since that is the designation given to the strains in the Northwest A&F University strain collection. Other strains were obtained from the Wine Yeast and Bacteria Collection of the Department of Viticulture and Enology at the University of California, Davis. The data from all fifty-two Saccharomyces isolates listed in Table 1 of Liu et al. were included in this study as the method of interdelta sequence analysis was identical. These yeast strains were collected from California, France, Italy, Northern Europe, and Spain .The interdelta sequence patterns obtained after gel electrophoresis were used for the construction of a presence/absence matrix, taking into account the total number of different bands observed. The interdelta sequence patterns were obtained following electrophoresis.