Search

Background: Whole-genome studies of vine cultivars have brought novel knowledge about the diversity, geographical relatedness, historical origin and dissemination, phenotype associations and genetic markers. Method: We applied SOM (self-organizing maps) portrayal, a neural network-based machine learning method, to re-analyze the genome-wide Single Nucleotide Polymorphism (SNP) data of nearly eight hundred grapevine cultivars. The method generates genome-specific data landscapes. Their topology reflects the geographical distribution of cultivars, indicates paths of cultivar dissemination in history and genome-phenotype associations about grape utilization. Results: The landscape of vine genomes resembles the geographic map of the Mediterranean world, reflecting two major dissemination paths from South Caucasus along a northern route via Balkan towards Western Europe and along a southern route via Palestine and Maghreb towards Iberian Peninsula. The Mediterranean and Black Sea, as well as the Pyrenees, constitute barriers for genetic exchange. On the coarsest level of stratification, cultivars divide into three major groups: Western Europe and Italian grapes, Iberian grapes and vine cultivars from Near East and Maghreb regions. Genetic landmarks were associated with agronomic traits, referring to their utilization as table and wine grapes. Pseudotime analysis describes the dissemination of grapevines in an East to West direction in different waves of cultivation. Conclusion: In analogy to the tasks of the wine waiter in gastronomy, the sommelier, our &lsquo, SOMmelier&rsquo, approach supports understanding the diversity of grapevine genomes in the context of their geographic and historical background, using SOM portrayal. It offers an option to supplement vine cultivar passports by genome fingerprint portraits.

INTRODUCTION – Common limits for phenotyping are costs in terms of equipment, specialized researchers, and time. Moreover, phenotyping requires specific methods in relation to the studied traits, and each method is generally designed to highlight very specific characteristics of the singular trait. Thus, the obtained results are not easily comparable between different institutions. However the data sharing is clearly the most powerful approach to limit the costs and efforts optimizing the obtained results in term of number of studied accessions. It should be stressed that funds are another great limit in research, especially in East countries, where the most interesting cultivar variability is preserved. 28 AIMS AND SCOPES – The main purpose of this work is the establishment of a phenotyping network able to produce comparable data among the Pan-European grapevine collections. MATERIALS AND METHODS – Two protocols were proposed: one for phenology and, the other, for enocarpological traits. Data were recorded in two consecutive vintages: 2012 and 2013. Concerning phenology, the extended BBCH scale (universal and flexible) was adapted to our objects by moving it into a chrono- phenological scale. Then, a cubic curve was selected for modelling the phenological courses. Thus, computing for each accession specific coefficients, it was possible to obtain the relative phenograms. Considering the different site climatic conditions, it was possible to spatialize the results, obtaining reference phenological maps. Concerning eno- carpological traits, the attention was focused on characteristics with a clear impact on the production quality, such as berry morphology (weight, diameters, number of seeds …) and composition (sugar, titratable acidity, anthocyanin and phenolic contents …). The protocol was organized to concatenate the data acquisition, with the objective to generate new derived variables. The methods were selected to be easily reproduced in all the involved collections, without the necessity of expensive facilities. RESULTS AND DISCUSSIONS – The first year of dataset was used to evaluate the methods. Two publications are already available confirming the protocols quality. Concerning phenology, both the variability between accessions grown into the same collection and the phenological shifts between the reference varieties cultivated in different areas were highlighted by the protocol. Matching climatic and phenological information of reference cultivars, phenophases maps were obtained. The performances obtained by the eno-carpological protocol were satisfactory both in a global point of view and in a singular collection evaluation. The project obtained an impressive participation, considering the volunteer contribution without any material funding to cover the work costs. Twenty-four institutions collaborate to the data collection. 1724 accessions were evaluated: 931 (583 in 2012 and 348 in 2013) for phenology and 793 (424 in 2012 and 369 in 2013) for eno-carpological descriptions. CONCLUSIONS AND POSSIBLE APPLICATIONS – The protocol sharing allows the data comparison among different grapevine collections. The possible future application for this dataset could be related to scientific studies of genetic associations. Moreover, the improved knowledge of minor and neglected cultivars could have a direct impact on the industrial production, giving information concerning new varieties for future introduction in viticultural areas.