Oenologists Review November 2015

Since its domestication, the vine has been improved through a combination of hybridizations and selections of somatic mutations resulting in the existence of 5,000 to 10,000 cultivated varieties (or cultivars, or grape varieties). The open pollinations that generated the old grape varieties, referred to here as traditional, are obviously not documented. The first artificial crosses of Vitis species were carried out in the eastern United States in the early 19th century, with the mother plant often being the only parent chosen. From the late 19th century, thousands of interspecific crosses of Vitis were made to create cultivars, including rootstocks resistant to phylloxera and other pathogens introduced in Europe. In Vitis vinifera, the first documented crosses (with selection of both maternal and paternal parents) appear to be those made by L. Bouschet de Bernard in 1828 in the south of France, in order to enhance the color of red wines. From that time on, several private breeders undertook hybridization work to improve agronomically important traits. The varieties resulting from these recent works will be referred to here as modern. After World War II, most vine improvement programs were conducted by agricultural institutes. The supposed genealogy of modern grape varieties is known thanks to breeders' data, although these information may contain some inaccuracies. In addition, since the pioneering work of Thomas et al. in 1994, numerous studies based on the heritability of highly polymorphic nuclear microsatellite markers (or nSSRs) have been conducted to specify the parentage of several grape varieties. These studies have confirmed some breeders' data, refuted others, but also discovered the parentage of traditional grape varieties. The majority of studies have focused on resolving the genealogy of grape varieties belonging to a particular geographical area or historical period, generally analyzing fewer than a hundred cultivars. The objective of our study was to analyze a large sample of 2,344 cultivated varieties of Vitis vinifera to discover or confirm their parentage. The aim was to better understand the general characteristics of vine improvement throughout history and to provide information to assist current breeders in selecting appropriate parents in new breeding programs.Material and methodsWe studied the data of 20 nSSRs obtained by Laucou et al. (2011) for 2,344 cultivars of Vitis vinifera preserved in the INRA collection at the Domaine de Vassal (34340 Marseillan-Plage). Parentage analysis was performed using the Famoz software. Logarithm of odds (LOD) scores were assigned to each possible parentage. LOD scores represent the probability that an individual is the parent of a descendant divided by the probability that these two individuals are not related. The threshold for relevant LOD scores was determined empirically through the validation of known pedigrees. The final assignment of parentage took into account historical data when available in ampelographic literature, as well as certain phenotypic characteristics (berry skin color, seedlessness, muscat flavor) capable of revealing errors in doubtful cases. Grape Genealogy Hybridization Selection Mutation Cultivars Gouais blanc Breeder Vitis Vinifera Parent