DANIELXD96

Review Prospects for applications of genomic tools in registration testing and seed certification of ryegrass varieties Abstract The ryegrass (Lolium) species, perennial ryegrass (Lolium perenne L.) and Italian ryegrass (Lolium multiflorum Lam.), are the two most important pasture grass species for global temperate regions and are also useful for amenity purposes. Due to an outbreeding reproductive habit, high levels of genetic heterogeneity are present within cultivated varieties. Acontinual increase in delivery of new cultivars to the marketplace, along with limited capacity to discriminate on the basis of morphological characteristics in a growout test, has caused difficulties for distinctness, uniformity and stability (DUS) testing in the current plant variety protection (PVP) system. A range of genomic tools and resources have been generated for ryegrasses, which provide new opportunities and challenges to the current PVP system. In this review, the currently available genomic tools and resources are described, along with prospects for applications to the PVP system and testing for seed certification and purity accreditation .Key words: Lolium — plant variety rights — distinctness —uniformity and stability — molecular genetic marker — seed purity Properties of Ryegrasses The ryegrass genus (Lolium) belongs to the Poaceae family. Having diverged from the group of broad leaved fescues (Festucaspecies) c. 2 million years ago, it has undergone recent differentiation into distinct species (Charmet and Balfourier 1994, Char-met et al. 1997). Both natural and artificial hybrids have beenreported between ryegrass and European species of Festuca (Jen-kin 1933). Due to the close affinity, the two groups are collectively referred to as the Festuca–Lolium or Lolium–Festuca complex (Kopecky et al. 2009, Hand et al. 2010). The Lolium genus contains 10 species (Terrell 1968, Scholz et al. 2000,Scholz and Scholz 2005), which are native to Europe, Asia and Northern Africa, as well as being cultivated and naturalized through out temperate regions of the world (Hubbard 1954).Lolium perenne (perennial ryegrass) and Lolium multiflorum(Italian ryegrass) are economically important for pastures and have also been used for turf and lawn purposes. The two speciesare believed to share a relatively recent common ancestor (Bor-rill 1976, Balfourier et al. 1998), with an overlapping range of phenotypic characters (Bulinska-Radomska and Lester 1985,Bennett et al. 2000). Interspecific hybridization can occur naturally, the resulting taxon being named L. x boucheanum Kunth (synonymous with L. x hybridum Hausskn.). The hybrid taxon is of value in breeding as a means to combine the favourable attributes of Italian ryegrass (rapid establishment and winterand early spring growth) with those of perennial ryegrass (suchas persistence). The generation of an increasing number of cultivars from the two species, along with their hybrids, creates acontinuum of related plant genetics and trait expression, whichpresents challenges for the definition of varietal identity. Ryegrass species are naturally diploid, with a chromosome constitution of 2n = 2x = 14, and tetraploid ryegrass genotypes can be artificially produced by treatment with colchicine (Ahloo-walia 1967) and cultivated in order to deliver improved herbage production. Furthermore, both perennial and Italian ryegrass areoutbreeding species with a gametophytic system of self incompatibility controlled by two gene loci, each of which exhibits multiple alleles (Cornish et al. 1979). The outbreeding nature results in high genetic heterogeneity within populations, which must be taken into consideration in the plant variety protection(PVP) system. Genomic Tools and Resources for Ryegrass Species Lolium genome sequence and annotation and comparative genomics The haploid (1C) nuclear genome size of perennial ryegrass has been estimated as c. 2.6 Gb (Kopecky et al. 2010), and the value for Italian ryegrass is expected to be similar. Studer et al. (2012)created a high density single nucleotide polymorphism (SNP)genetic map for perennial ryegrass, including placement of 732 expressed genes. Subsequently, a perennial ryegrass GenomeZipper was generated and a further 3315 transcript DNA sequences were anchored to the map through analysis of comparative genomics with barley, Brachypodium distachyon, rice and sorghum(Pfeifer et al. 2013). De novo assembly of the transcriptome for perennial ryegrass as well as four other species from the Lolium–Festuca complex (Festuca pratensis, L. multiflorum, L. m. var.westerwoldicum, and L. temulentum) has been generated (Farrellet al. 2014, Shinozuka et al. 2014, Czaban et al. 2015). A draft enome sequence, based on assignment of sequence contigs to the framework defined by the GenomeZipper, has recently been described (Byrne et al. 2015). The complete chloroplast genome(cpDNA) as well as the mitochondrial genome sequence of perennial ryegrass is available (Diekmann et al. 2009, Islam et al. 2013). A collection of insertion/deletion polymorphisms, SNPs and SSRs have been identified from these sequences. Germplasm resources and genetic diversity Germplasm resources for ryegrass species have been collected by major plant genetic resource centres and regional research centres such as the Germplasm Resources Information Net-work (GRIN), USDA; the Institut f€ur Planzengenetik undKulturpflanzenforschung (IPK) Gatersleben, Germany; theNordic Genetic Resource Centre, Sweden; and the Australian Pastures Genebank in the South Australian Research and Development Institute. All of these collections are publicly accessible for research purpose under a standard material transfer agreement.Genetic diversity studies of ryegrasses, using mainly morphological characteristics, were initially focused on natural populations, for the establishment of core collections of valuable genebank materials and in order to assist understanding of processes of diversification in response to ecological factors (Hayward1985, Prendes et al. 1989, Charmet et al. 1993, Balfourier and Charmet 1994, Fernando et al. 1997, Oliveira et al. 1997, Bal-fourier et al. 1998). Following the development of moleculargenetic marker systems, these technologies were applied to studies of the diversity of both nondomesticated and cultivated rye-grass germplasm (Stammers et al. 1995, Huff 1997, Warpehaet al. 1998, Roldan-Ruiz et al. 2000, 2001, Cresswell et al.2001, Guthridge et al. 2001, Ghariani et al. 2003, Kubik et al.2001, Bolaric et al. 2005a,b,c, van Treuren et al. 2005, Wanget al. 2009) in order to understand genetic relationships. Recently, multiplexed SNP genotyping tools were developed for perennial ryegrass and were able to differentiate perennial, Italian and hybrid cultivars, including the capacity to discriminate between diploid and tetraploid cultivars (Wang et al. 2014).Genomewide allele frequency finger prints (GWAFFs) on a population basis have also been used to distinguish between eight perennial ryegrass cultivars, without a requirement for genotyping of single plants (Byrne et al. 2013). QTLs and gene specific markers Several genetic linkage maps of perennial ryegrass have beenconstructed (Hayward et al. 1998, Bert et al. 1999, Jones et al.2002, Armstead et al. 2004, Jones et al. 2002, Armstead et al.2002, Cogan et al. 2005, Studer et al. 2012). These maps have provided the basis for identification of over 500 quantitative traitloci (QTLs) for a range of characteristics (reviewed by Shi-nozuka et al. 2012). Candidate gene based association mapping studies for traits such as flowering time, herbage quality and drought tolerance have been reported for perennial ryegrass(Skøt et al. 2005, 2007, 2011, Pembleton et al. 2013, Yu et al.2013). In addition, a low density genome wide association study(GWAS) for detection of loci controlling multiple agronomictraits was performed on Italian ryegrass cultivars (Wang et al.2015a). A number of genes for specific phenotypic traits have been isolated and characterized, including those controlling stress tolerance and time of flowering (vernalization and photoperiod response) (Jensen et al. 2001, Armstead et al. 2002, Martinet al. 2004, Jensen et al. 2005, Andersen et al. 2006, Petersenet al. 2006, Xiong and Fei 2006, Ciannamea et al. 2006a,b,2007, Han and Suleiman 2008, Sandve et al. 2008, Zhang et al. 2010, Asp et al. 2011, Fiil et al. 2011, Cao et al. 2015,Patel et al. 2015). However, none of the relevant markers have been reported as a reliable predictor of any phenotypic traitand as such are unlikely to be of value in predicting performance of a cultivar or distinguishing cultivars. Status and Issues Arising for Current Distinctness,Uniformity and Stability Testing for Ryegrass The International Union for the Protection of New Varieties ofPlants (UPOV) has set distinctness, uniformity and stability (DUS) testing guidelines for cultivated ryegrass species (http://www.upov.int/edocs/tgdocs/en/tg004.pdf), which require assessment of 23 characteristics in a growout test in both spaced plant and row plot trials. Apart from the ploidy level, all are morphological characteristics, only five of which require evaluation at the vegetative growth stage, while the others are related to flowering and so can only be measured at the reproductive development stage. Trait numbers 9 and 11 constitute a single trait (time of inflorescence emergence), relevant to different ryegrass species which either lack or possess vernalization requirements. Trait numbers 16 (flag leaf: length/widthratio) and 21 (inflorescence: density) are secondary traits,which are derived from several primary traits, namely flag leaflength and width, inflorescence length and spikelet number,respectively.A total of 19 620 varieties of the Lolium genus were listed inthe UPOV plant variety database (www.upov.int/pluto/en/), of which 4577 with plant breeder's rights (PBR), as of 5/01/2016.A total of 118 varieties of the Lolium genus were listed in the PBR database in Australia (http://pericles.ipaustralia.gov.au/pbr_db/) as of 5/01/2016. Of the 118 varieties, 85 are provided with detailed descriptions of trial data. It is apparent that some of the traits such as Trait number 6 (plant: width after vernalization), 10 (tendency to form inflorescences without vernalization) and 13 (plant: width at inflorescence emergence) are seldom employed. Alternatively, some traits that are not included on the specified list have been reported as being used for DUS testing,including score of anthocyanin coloration of the vegetative leafs heath, number of nodes in the stem, stem number per plant,number of florets per spikelet, seeds number per spikelet, 1000-seeds weight, seasonal yield score, rust resistance score and others. Visual scoring of traits has frequently been performed through the use of numerical scales with different minimum–maximum ranges (e.g. 1–3, 1–5or1–9), but sometimes in opposing directions (e.g. ‘1’ scored as ‘prostrate’ in one trial but as ‘erect’ in another, when assessing the extent of growth habit).Measurements have also been taken in different units. For example, Trait number 21 (inflorescence: density) has been reportedin terms of inflorescence length/number of spikelets (accordingto the UPOV guideline), but also as number of spikelets arrayed over a particular distance on the spike (such as every 5 cm or10 cm). Such anomalies create difficulties for comparisons of varietal performance across trials.Selection of comparator varieties for the growout test has been mainly made on the basis of ploidy level (either diploidor tetraploid) and time of inflorescence emergence. As the number of varieties increases, discrimination on the basis of morphophysiological characteristics has become increasingly difficult, due to a requirement to include larger numbers of similar varieties in the trial, thus increasing cost and complexity. Possible Applications of Genomic Tools in DUS Testing Suitability of three models proposed by UPOV for cultivated ryegrasses Research and discussion into the implementation of molecular technologies for DUS testing in a broad range of agricultural plant species has been proceeding for a while. The UPOV technical committee proposed three potential general application models, namely (i) characteristic specific molecular genetic markers; (ii) calibrated molecular marker based distances in the management of variety collections; and (iii) the use of molecular genetic marker-based characteristics (http://www.upov.int/edocs/infdocs/en/upov_inf_18_1.pdf). Model 1 represents the case in which molecular genetic markers are directly predictive for a particular character, such as barley seasonal type (winter or spring) and row number (2 or 6 rows), which can be predicted perfectly by molecular markers(Cockram et al. 2012). A special example of model 1 for ryegrass is ploidy level. Diploid and tetraploid individuals may be discriminated through the use of standard cytological methods,or by electrophoretic separation of phosphoglucoisomerase (PGI)isozyme variants (UPOV). It has also been demonstrated that ploidy level can be determined through SNP genotyping (Wanget al. 2014). However, ploidy level represents an isolated instance, which is a whole genome characteristic instead of a phenotypic trait. Although many QTLs and some genes forspecific traits have been identified, to the best of the authors’knowledge, there are no instances to date of phenotypic traits that are amenable to prediction by diagnostic genetic markers forryegrass species and as such model 1 is highly unlikely to be applicable.The proposed application for model 2 is as a means to reduce the number of reference cultivars used in the growout test. The model will work if the same decisions (on whether to include or exclude a variety) can be reached based on the genetic distances derived from either genetic marker or phenotypic characteristic based data. So far, the correlation between the two have not been particularly encouraging for adoption of this model, based on the studies of various crop species (Tommasini et al. 2003,Kwon et al. 2005, Gunjaca et al. 2008, Smykal et al. 2008,Jones et al. 2013). Similarly, the correlation between distances estimated from AFLP marker and trait specific data for various cultivars of ryegrass species was low (r = 0.10, van Treurenet al. 2005). On the basis of these results, future opportunities for application of model 2 appear to be limited.The proposition to treat molecular genetic markers as characteristics in their own right (model 3) is an interesting conceptand is probably more suitable for implementation than the two preceding models. In principle, no fuller description of a variety may be obtained than an entire DNA sequence (or, as is the case for outbreeding species such as ryegrasses, collection of DNA sequences). However, for the data set to be truly informative,information on phenotype is still necessary. Therefore, a rational outlook for a future PVP system would be a twoarmed system, one of which consists of the current phenotype, the other of newly obtained genotypic characterization. While the second armis being developed, the first may remain undisrupted and fully functional. The major advantage of a genotypic data set over a phenotypic counterpart derived from a growout test is that genotyping a collection of reference cultivars is not necessary each time when a new candidate variety is submitted for assessment.The two armed system would permit genetic relationships between varieties to be revealed, allowing DUS testing in terms of genotypic data. Application 1: Revealing genetic relationships between varieties Research has shown that molecular genetic marker based analysis is superior to phenotype based methods in terms of revealing levels of genetic relatedness between current and historically used plant cultivars (Garland et al. 1999, Lombard et al. 2000,Tommasini et al. 2003, Iba~nez et al. 2009, Pourabed et al.2015). This would be particularly useful for ryegrass cultivars, for which pedigree information has not always been clearly oraccurately recorded. As a specific example for ryegrasses, a SNP genotyping tool with high multiplex ratio (380 loci) was used to genotype 48–192 individuals from 27 cultivars of perennial, Italian and hybrid ryegrass, allowing construction of a neighborjoining (NJ) tree and assessment of relationships which proved consistent with knowledge of breeding history (Wang et al.2014). Later, based on the use of amplicon based genotyping-by sequencing (GBS) approach on bulked seed samples, the NJ tree was extended to 63 cultivars of ryegrass, data being obtained at much lower cost (Pembleton et al. 2016). In both of the studies, two major cultivar specific groups were identified,corresponding to the two species (perennial and Italian ryegrass).Hybrid ryegrass varieties were found to be distributed along the tree branch that connected between the two groups, showing varying levels of affinity to either the perennial or Italian rye grass cultivar clusters depending on the relative proportions of contributing genome from each species. For this reason, inter specific hybrid cultivars are more usefully described in terms of relative affinity (expressed as a percentage or proportion) to perennial or Italian ryegrass, than as single uniform type. Inclusion of varieties of cultivated fescue species into the dendrogram would also probably be desirable, due to the close relationship between the Lolium and Festuca genera, and the capacity to produce intergeneric Festulolium hybrids, as well as the potential presence of introgressed Festuca chromosomal segments within Lolium genomes. Given success in determination of the structure of present day varieties, cultivars that are developed in future may be added to the catalogue, allowing immediate visualization of genetic relationships. Information from the dendrogram canal so provide guidance for future breeding activities to maximize genetic diversity, make effective choices and introduce new genetic resources. Application 2: Distinctness test The most effective system for differentiation and identification of varieties must be practical, precise, reliable and robust. A SNP-based genotyping system easily meets these requirements and would be expected to exceed the corresponding performance of a growout test. The marker number may be much higher than the number of traits that are measured in the growout test, while genotypes are less environmental dependent than phenotypic traits, and genotypic assays have the potential to include more individuals and to be implemented at early stages of plant growth over a much reduced time frame with lower costs. It is possible that some of the assayed markers will serve as unique identification keys for individual varieties (Pourabed et al.2015). However, cultivars may be separated more commonly onthe basis of a combination of allele frequencies derived from a set of markers. The capacity to use molecular genetic markers as a tool for germplasm differentiation has been demonstrated for abroad range of plant species, including ryegrasses (Guthridgeet al. 2001, Kubik et al. 2001, Momotaz et al. 2004, Wang et al.2009, 2014, Pembleton et al. 2016). In most cases, cultivars have been effectively discriminated, although the resolution was limited for those varieties that have been derived from land racesor ecotypes. This may not be an issue because breeding strategies based on mere selection of ecotype are no longer adopted. Furthermore, it has been demonstrated that resolution increasedas the number of SNP loci increased (Byrne et al. 2013), offering the potential to ultimately resolve such population types. Application 3: Uniformity test Currently, a cross-pollinated variety is considered to be uniform for a measured morphological characteristic if the variance for that characteristic does not significantly exceed the level of variation observed in comparable varieties (referred to as the standard deviation approach). When varieties that are currently lodged in the Australian PBR database are examined, the uniformity criterion is rarely mentioned. From the test dataset which provides the standard deviations of measured characteristics, some candidate varieties showed higher levels of variation than comparable varieties.This observation raises two issues with the current system. Firstly, the assessment of uniformity is dependent on the selection of comparable varieties. The numbers of such varieties are different for each candidate, as are the precise nature of the uniformities. Secondly, the standard for assessment of uniformity appears to be relaxed. As mentioned above, even when the variability of a candidate exceeded those of the comparable varieties, PBR was still granted. If this trend was to continue, the limit of tolerance will gradually increase and erosion of uniformity will be inevitable. Applications of genotypic data for uniformity test have been discussed in in breeding crop species (Cooke et al. 2003, Tom-masini et al. 2003, Singh et al. 2004, Velez and Ibanĕz 2012,Wang et al. 2015b). When genetic marker based analysis did not reveal any plant to plant variation within a variety, the variety was consequently deemed uniform (Singh et al. 2004). A much more complicated situation will arise for ryegrass species, due to the outbreeding reproductive system, and a large portion of the relevant molecular variance resides within a cultivar rather than between cultivars. Any test for uniformity will certainly involve the analysis of individual plants rather than on a population basis, given the requirement to determine within population variability.The assignment test of individual plants to their source cultivars based on molecular markers (Kubik et al. 2001, Wang et al.2009) could identify ‘off types’. This type of assignment test could be used as a uniformity measure, comparable to the ‘off types’ approach in the grown-out test. More detailed research is required in this area, to optimize the selection of marker loci, chosen genotyping protocol, and systems for the reliable measurement of uniformity (such as an assignment test,AMOVA,ordetermination of the ratio of within to between cultivar molecular variance), as well as setting of the corresponding threshold values. Application 4: Stability test The statement that a variety is stable implies that it remains true to its description after repeated reproduction or propagation. Stability of cultivated ryegrass varieties has rarely been reported.This is probably because a variety, which has been shown to be uniform, may also be considered stable (UPOV Test Guideline).Additionally, there has been little research into the stability of outcrossing plant species based on the use of molecular genetic markers. Theoretically, it would be much easier and more practical to test for stability through the use of molecular genetic markers than by a growout test, in which environmental factors would be expected to exert a larger effect. Such analysis is complicated in comparison with that of in breeding species, for which homozygous allele status provides evidence for stability (Wanget al. 2015b) and inertia towards change from generation to generation. In the context of the genetic system of an outbreeding species, heterozygosity is expected to be the norm for most geneloci, and allelic fixation to be less frequent. Assessment of stability, therefore, must be mainly based on the measurements of allele frequency. If these frequencies remain constant across multiple generations, a variety may be deemed stable. According to the Hardy–Weinberg law, equilibrium of allele frequencies within a population can be reached by single cycles of random mating in the absence of selection, non random mating, differential migration or differential mutation. Therefore, in grass breeding, following selection of parental genotypes and polycross to produce Syn 1 seeds, the resulting populations are advanced through multiple generations of random mating to approach link age equilibrium (Vogel and Pedersen 1993) and to generate sufficient seed for agronomic assessment. In reality, non random mating may occur due to variation in flowering time and fertility among individuals and their relative spatial locations. As a con-sequence, the precise value of the tolerance levels (between pre-determined maximum and minimum values for allele frequency)for each marker locus will be of key importance for stability assessment, requiring further empirical research, which may beassisted by simulation studies. Practical considerations for implementation of the two armed PVP system The logical steps leading to implementation of this strategy would seem to be as follows: (i) assembly of a comprehensive cultivar catalogue of both current and historical varieties of ryegrasses, including interspecific hybrids, as well as possibly Festulolium intergeneric hybrids; (ii) genotypic analysis of allcurrent and particularly valued historical varieties to construct a comprehensive cultivar relationship dendrogram; (iii) to establisha public database for storage and sharing of data, with established links to the PBR database; and (iv) genotypic analysis ofany candidate variety and positioning within the dendrogram, so that the genetic relationship can be readily evaluated in terms of similarity with or differences from existing varieties. To implement this strategy in practice, detailed information on marker type and assay, marker number and numbers of individu-als sampled per cultivar may need to be determined or standardized. The type of genetic marker and nature of the assay would be relatively uncontroversial, given the abundance of SNP markersthat are provided by GBS methods. This is also a key consideration for the number of markers employed, which will not be limited in number when generated by GBS techniques, although a minimum number may need to be determined. The number of individuals per cultivar could presumably adopt the value that is conventionally used in the growout test, which is 60, or other agreed values, in particular for uniformity testing. However, toobtain GWAFFs for a cultivar for distinctness and stability testing, a bulk sample may be used, and hence, the number of individuals could far exceed 60. Although variations of marker number and number of individuals may contribute to small changes in allele frequency based on the genotyping methods, the relationships that are revealed would be expected to be robust. In support of this view, SNP genotyping based on Golden Gate oligopool assay(OPA) genotyping system (Wang et al. 2014) and GBS assays(Pembleton et al. 2016) revealed similar relationships.Future delivery to the marketplace of ryegrass varieties is likely to feature the increasing of the use of F1 hybrids, whichare expected to provide substantial agronomic benefits through capture of heterosis. Recently, a method capable of producing c. 83% F1 hybrid progeny between combining parental populations has been described, which relies on restriction of allelic diversity at the S and Z gametophytic self incompatibility loci within parental populations (Pembleton et al. 2015). The F1 hybrid varieties generated by this procedure will provide novel attributes for the DUS testing systems. A high degree of distinctiveness may be anticipated, given the combinations of attributes from complementary and potentially contrasted parental populations.The uniformity will depend on the variability of the parental populations, and the stability of the hybrid varieties will also depend on the stability of the parental populations. Therefore, the DUS testing of parental populations is necessary. Possible Applications of Genomic Tools in Seeds Certification and Purity Test Certified seed provides users with the confidence that the expected advantageous qualities of a given cultivar can be reliably delivered. The Grass and Legume Seeds Scheme of the Organisation for Economic Cooperation and Development(OECD) sets specific rules and regulations for the production ofseeds, so as the National Seeds Certification Manual (http://pir.sa.gov. au/__data/a ssets/pdf_file/0003/14 8134/Seed_Certifica-tion_Manual.pdf). Aust ralia Seeds Auth ority Ltd provide d anational l ist of plant v arieties eligible for seed certification inAustralia (as of 21 October 2014), which i ncludes 37 Italian ryegras s, 25 perennial ryegrass, 6 hybrid and 2 Lolium rigidum(annual ryegrass) cultivars. Judicious use of genomic tools can assist in the following aspects of seeds certification process. Application 1: Varietal purity in seeds crops during field inspection and after seeds production At least one field inspection is required during seed production to determine the number of plants that are not true to the variety, as well as the number of plants of other species. For certified Lolium seeds, the maximum number of plants of the same species being not true to the variety and of other species is set to be1 within 10 square metres. The inspection is dependent on visualassessment from an experienced inspector. An assignment test based on molecular genetic markers can assist in determining the‘off-types’ as discussed in previous section and would be more accurate and objective. Similarly, the assignment test may also apply after seed production to test the purity of seed lots as an effective quality control measure. Application 2: Detection of mislabelling Pembleton et al. (2016) demonstrated that the GBS approach can be useful for detecting mislabelled seed bags. GBS-OPA-based genotyping of different seed batches from the varieties ‘Trojan’,‘Bronsyn’ and ‘Warrior’ exhibited distinct and separate locations within the cultivar relationship dendrogram. However, two batches of seeds that were both labelled as belonging to variety‘Crusader’ were genotyped, and one was identified as being probably derived instead from cultivar ‘Warrior’, based on the coincidence within the dendrogram. Application 3: Detection of fungal endophyte in seeds lot It is also worth noting that commercial ryegrass seeds often contain fungal endophytes due to beneficial effects on performance. For this reason, molecular genetic marker assays that combinetests both for identity of the host grass genotype and presence/identity of the endophyte genotype would be particularly valuable. Identification of SNP variation between the haploid genomes of candidate endophytes belonging to the taxaEpichlo€e. festucae var. lolii, LpTG-2 and LpTG-3 has permitted the design of a combined assay for survey of both grass and endophyte genotypes (Kaur et al. 2015). Summary and Future Directions A range of genetic and genomic tools have been developed foruse with cultivated ryegrass species and have proved effective for revealing genetic relationships between different varieties, to differentiate species, to provide better descriptions of interspecific hybrids and for certification and purity testing of seeds. Such tools also offer more accurate measures than pedigrees, which are often incomplete or unavailable for ryegrass varieties. The relative absence of trait specific markers and observed low correlations between genetic distance and phenotypic distance imply that a third model may be more suitable for the modification ofthe PVP system, in which new genotypic data sets are constructed while the previous phenotype based system remains functional. The new two armed system will allow relatedness, DUS to be measured and redefined to a new and improved level of resolution. Assembly of a genetic database of current cultivars and linkage to the PBR database is likely to provide a useful starting point. In the future, a genetic test (e.g. GWAFFs) maybe required for any candidate variety in addition to the grownout test, allowing the database to be gradually expanded to full operational capacity. Unless PVP systems evolve beyond the current status, the implementation of novel technologies is likely to generate difficulties. Successful production of F1 hybrids (Pembleton et al. 2015), in association with accelerated genetic gains within parental pools (estimated from both simulation and empirical-based studies) due to the use of genomic selection, will drive increased productivity, and so create increased awareness of plant performance characteristics in the field of cultivar registration. It is also possible that increased rates of genetic gain will deliver cultivars with shorter lifespans within the marketplace, placing further strain on the existing system. Additional ‘disrup-tive’ genomic technologies will have further impacts and may lead to a requirement for intellectual property (IP) protection toaccompany PVP systems, as is already the case for some major grain crops. One example would be the widespread adoption of genome editing technologies (Bortesi and Fischer 2015). It is clear that the nature of ryegrass breeding will be changed by the application of genomic technologies and that the requirements for an efficient and accurate system to protect PBR must adapt to meet these changed circumstances. Acknowledgements The authors acknowledge support from the Victorian Department of Eco-nomic Development, Jobs, Transport and Resources. Research on molecular breeding of ryegrass species has been funded by Dairy Australia, the Geoffrey Gardiner Dairy Foundation and Meat and Livestock Australia through the Molecular Plant Breeding and Dairy Futures Cooperative Research Centres. The authors would like to thank Professors Kevin F.Smith and German Spangenberg for critical reading of the manuscript.