Grapevine germplasm variety identification
Abstract
The aim of this study was to validate the name and DNA profile of all accessions in the CSIRO and SARDI grapevine collections as well as unique accessions in four other Australian grapevine collections. Genotyping was carried out using a combination of DNA typing techniques over two stages and the results compared to databases of major European collections. In total, 1514 accessions were DNA typed, successfully confirming the identity of 363 unique wine, table/dried and multi-purpose grapevine varieties and 875 accessions currently within Australian germplasm collections. Some cases of misnaming were identified. The study also allowed variety Prime names to be assigned to valuable Australian historical material. The validated varieties form the nucleus for the establishment of the National Grapevine Collection and pave the way for the collections to be reopened to industry. The work also provides the foundation of a public register of planting material in Australia and has led to the creation of a database that could form the basis of a commercial DNA typing service for grapevine varieties.
Summary
Grapevine varieties in the major Australian grapevine germplasm collections have traditionally been identified and named by ampelography (visual inspection of the grapevine, particularly leaves). The Albariño/Savagnin blanc misidentification of 2008 highlighted the need to validate the names of varieties held in Australian collections and resulted in the closure of the CSIRO and SARDI collections, which hold up to 80% of the unique material in Australia. The primary objectives of this project were to validate the name and DNA profile of varieties using international references and to develop a single nucleotide polymorphism (SNP) profile database containing the validated varieties that could be used by an Australian service provider to offer an improved grapevine identification service to industry.
Initially, the project planned to work with three major European grapevine collections using a set of 48 SNP markers (Cabezas et al. 2011). Once an agreed list of varieties had been developed, DNA would be sampled from the varieties held in participating Australian collections and from reference vines of the same varieties held in the European collections. The DNA samples from the European collections would then be sent to Australia and along with the DNA samples from the Australian collections, all DNA samples would then be genotyped by the same provider using the same set of 48 SNP markers. However, the logistics and practicality of working with three European collection-holders proved difficult and following discussion with Wine Australia it was decided to work with one major European collection holder, the Spanish organisation, Consejo Superior De Investigaciones Cientificas (CSIC) through Instituto de Ciencias de la Vid y del Vino (ICVV).
In Stage 1 of the project a total of 1514 accessions were genotyped - 1037 from the CSIRO collection, 455 from the SARDI collection and a further 22 from OTHER collections including Adelaide Hills Vine Improvement, New South Wales Department of Primary Industries, West Australian Agriculture Authority and Yalumba Nursery. CSIRO arranged sampling of the accessions. DNA extraction and SNP genotyping was carried out by the Australian Genome Research Facility Limited (AGRF). The DNA profiles were then sent to the Spanish collaborator. The Spanish collaborator compared the SNP profiles of Australian accessions with accessions in their own collection and then nominated six different naming outcomes for each accession that was supplied:
Outcome 1: The SNP pattern of the Australian sample fully matched a genotype in the ICVV-SNP database with a variety Prime name. The variety Prime name was assigned using upper case letters or, where the Prime name is a global name for a cluster of varieties derived from a single original embryo, using lower case letters.
Outcome 2: The SNP pattern of the Australian sample fully matched a genotype in the ICVV-SNP database with a Provisional name. A Provisional name was assigned in ICVV-SNP database when there had been insufficient independent evidence to enable establishment of a variety Prime name for the genotype. Thus, the identification of the Australian accession was considered inconclusive.
Outcome 3: The SNP pattern of the Australian sample almost fully matched a genotype in the ICVV-SNP database, so the name of the closest genotype was assigned. If the closest genotype was a Prime name, that Prime name was assigned. If the closest genotype was a Provisional name, that Provisional name was assigned. In both cases the identification of the Australian accession was considered inconclusive.
Outcomes 4-6: No name could be assigned because either the SNP pattern of the Australian sample did not match with any ICVV-SNP database genotype or matched a genotype in the database which matches multiple varieties or the SNP genotyping quality was too low.
Out of a total of 1514 accessions analysed, 1339 (88%) could be assigned a VIVC variety name (i.e. Outcomes 1-3) while 175 (12%) could not be assigned a variety name (Outcomes 4-6). Of the 1339 accessions that were assigned a name, 751 (56%) were assigned an Outcome 1 variety Prime name, 204 (15%) were assigned an Outcome 2 Provisional name and 384 (29%) were assigned an Outcome 3 Prime or Provisional name. It should be noted that the set of 48 SNPs that were used in this project was unable to distinguish clones and had insufficient resolution to reliably distinguish non-vinifera Vitis species and varieties with non-vinifera Vitis species in their pedigree. As most, if not all rootstocks used commercially in Australia, have American non-vinifera Vitis species in their background, it was not possible to assign names to these rootstocks using this SNP panel.
In Stage 2, a decision was made, following discussions with Wine Australia, to confirm the identity of a selection of the most important wine, table/dried and multi-purpose varieties from the germplasm collections that had been assigned Prime or Provisional names, which could form the basis of an Australian national germplasm collection. Leaf samples of these selected accessions were re-collected from the field in October 2020.
For accessions that had been assigned an Outcome 1 variety Prime name, SNP analysis was repeated, and the results analysed in-house using the SNP profile database that had been developed as part of Stage 1. In total, 583 accessions that had been assigned a variety Prime name from Stage 1, were re-sampled and re-analysed using this methodology and all were found to match the SNP profile obtained previously in Stage 1. This demonstrates the excellent reproducibility of the SNP DNA typing method.
For accessions that had been assigned Outcome 2 Provisional or Outcome 3 Prime or Provisional variety names, an alternative method of DNA typing based on simple sequence repeat (SSR) analysis was employed in an attempt to validate the variety names assigned by SNP analysis in Stage 1. Because of the much greater expense involved in SSR analysis, compared to SNP analysis, it was not feasible to analyse all 588 Outcome 2 and Outcome 3 accessions assigned a variety name by SSR analysis in Stage 1. Therefore, where there were multiple accessions with identical SNP patterns, a representative genotype was selected for SSR analysis. A total of 185 DNA samples of scion accessions were sent to the Institut Francais de la Vigne et du Vin (IFV) in France where they were subjected to SSR analysis. We also took the opportunity to send samples of the 20 most commonly used rootstocks for SSR analysis due to the issues observed with trying to genotype rootstock accessions using our 48 SNP panel.
From the 185 scion DNA samples analysed using SSRs, 150 (81%) were found to match samples in the IFV SSR database and were assigned a variety name. The remaining 35 samples did not match any SSR profiles in the IFV database and were classified as “unknowns”. Of the 150 that were assigned a variety name, 137 were found to match the Prime or Provisional name that had previously been assigned using SNP analysis in Stage 1. However, 12 of the 185 samples analysed by SSR analysis were assigned variety names that did not match the Outcome 2 or Outcome 3 Prime or Provisional variety names assigned using SNP analysis in Stage 1. In contrast to the results obtained from SNP analysis, SSR analysis confirmed the identity of the 19 of the 20 rootstocks analysed.
In a number of cases, the variety Prime name assigned by dual SNP analysis or combined SNP/SSR analysis differed from the name given to this accession in the Australian germplasm collection. These ranged from minor spelling differences to, in some cases, name differences. For many varieties, the name of the variety in the Australian collection was a recognized VIVC synonym of the assigned Prime name for that variety e.g., Shiraz (Australian collection) vs Syrah (VIVC) or Traminer (Australian collection) vs Savagnin (VIVC). Investigations showed that in the majority of these instances where cuttings had been distributed to industry, this was either under the correct variety name or the varieties are not being used for wine production in Australia.
The single exception was the variety labelled as Petit Manseng in the CSIRO germplasm collection which was confirmed in 2019 to be MANSENG GROS BLANC (syn. Gros Manseng) and records indicated that this variety had been distributed from the CSIRO collection under the incorrect name. Furthermore, the National Vintage Survey reports that on average 44 tonnes per year of grapes labelled as Petit Manseng were crushed between 2015-19. As this has implications for the varietal labelling of the resultant wine, affected producers were notified by Wine Australia at the time through its regulatory function.
Discussions were also initiated with AGRF to investigate the possibility of them setting up a commercial service for grapevine DNA typing based on the SNP data collected from this project. To facilitate this process, we also carried out a pilot study looking at the feasibility of DNA extraction from buds and wood from canes, rather than leaf material, because leaf material may not always be available when a grape grower wishes to undertake DNA typing of germplasm in his vineyard. The results of the pilot study on 14 grapevine genotypes indicated that, although the failure rate for correctly reading individual SNPs on the panel was slightly higher with DNA extracted from buds and wood, than leaves, it did not significantly impact on varietal identification of the genotypes.
Finally, of special interest is the fact that this study has been able to assign variety Prime names to 11 accessions obtained from historical Australian collections and held in the CSIRO and SARDI collections for many years. Some of these accessions pre-date the phylloxera crisis in Europe in the late 1800s. These include six accessions from the Best’s (Best Western, Victoria) collection, four accessions from the Coldstream (Victoria) collection and one accession from the Auldana (South Australia) collection.
Overall, the duplicate SNP/SNP analysis and the combined SNP/SSR analysis allowed assignment of a VIVC Prime/global Prime name to 362 unique genotypes and 875 accessions currently within Australian germplasm collections. These varieties form the nucleus for the establishment of the National Grapevine Collection (NGC) and a public register of planting material in Australia, and pave the way for the reopening of the collections to industry.