Oat is a cereal crop of global importance used for food, feed, and forage. It is adapted to cool climates and is cultivated predominantly in temperate regions or in winter seasons. Most cultivated varieties of oat belong to Avena sativa
L., an allohexaploid species with 2n = 6× = 42. Other species in the genus Avena
have ploidy levels ranging from diploid to hexaploid [1
], and some of these species have been used as sources of new traits for cultivated oat.
In oat, as in other crop species, there is growing recognition of the need to identify patterns of global genetic diversity, and to use this information in concert with tools for genomic discovery and molecular breeding. Genetic diversity (and associated population structure) depends largely on historical patterns of deliberate and passive efforts to create improved crop varieties [3
]. For pragmatic reasons, most oat breeders tend to favour crosses among locally adapted varieties, which may erode genetic diversity within a breeding program and create geographically-dependent population structures. Patterns of diversity also develop as a result of breeding objectives that are targeted toward specific adaptations and commercial uses. Both spring and winter forms of A. sativa
exist, but the characteristics that define winter oat (requirement for vernalization and tolerance to freezing) are expressed to varying degrees, and many winter oat varieties can also be grown as spring-seeded annuals. Another distinction is made between varieties with groats (oat kernels) that thresh free from their hulls (hulless, or naked oat) vs.
those with hulls that adhere to the seed (covered oat). Further distinctions are made based on hull colour and grain composition, and these characteristics can be relevant to commercial use or adaptation. However, most types of hexaploid oat are fully cross-fertile, and cross-hybridizations are made among different categories to varying degrees by different breeders.
A study of AFLP markers in a core set of cultivated oat germplasm [5
] indicated that most genetic relatedness in cultivated oat is associated with geographical origin and with the presence of a distinct, red-seeded, byzantina
-type that has sometimes been considered as a separate species or subspecies. Although the hulless character results in a distinct market class, this trait is affected primarily by a single locus [6
], and frequent inter-mating among covered and hulless types has apparently reduced this as a factor in population structure. A distinction between spring and winter types was not made in the above study [5
], but parallel development of spring and winter types within the same breeding program is rare. Cross-hybridization of spring and winter types is not common, due to the genetic complexity of these different adaptations, and also due to the technical difficulty in hybridizing varieties with different flowering times.
Modern genomics research in oat was inaugurated in 1992 with the publication of the first RFLP map in diploid Avena
]. This was followed by original and updated versions of hexaploid maps based on the 'Kanota' × 'Ogle' (KxO) recombinant inbred line (RIL) population [8
] and by the addition of new sets of mapped markers [10
]. Many additional maps, both partial and complete, have been published in hexaploid oat, as reviewed by Rines et al
] and compiled in an online database [18
]. However, most maps contain very few markers that are shared among other populations. The KxO map contains the most complete set of markers and has been the primary reference for comparative map analysis in oat. However, the KxO population presents some challenges as a reference population; notably, the population is relatively small, and contains at least one major translocation [19
] that causes clustering and pseudo-linkage of markers from two different chromosomes [9
]. Unlike wheat, where a combination of consensus mapping and physical mapping has resolved linkage groups that correspond to 21 chromosomes [20
], efforts in oat have not yet produced a true consensus map in which all linkage groups are assigned to the expected 21 oat chromosomes.
The current arsenal of molecular markers in oat is based on technologies that include SCAR, SSR, AFLP, and RFLP. Unfortunately, this diversity of technologies creates difficulties in performing comparative genomics within the oat community. Technologies based on PCR are the easiest to implement, but they require a multitude of different primers, and conditions for amplification may need to be re-optimized in different laboratories. Efforts to increase the availability of SSR markers in oat are ongoing, but only a limited number of these markers have been published, and only a subset of these are polymorphic in any given oat population [10
The above factors highlight the urgent need for a set of molecular markers that provide complete genome coverage, that are based on a homogeneous technology, and that can be scored readily in new germplasm by any member of the global oat research community. Such a resource would accelerate the development of new maps, and would allow the integration of existing maps into a single consensus map. It would also allow oat researchers to conduct routine marker analysis for breeding applications, for mapping novel traits, for studying genetic diversity, and for other diagnostic applications. Future advances in oat research, including sequencing and functional genomics, will depend on the availability of a robust consensus map containing reliable markers that can be scored on a high-throughput basis. Furthermore, there is mounting evidence that whole genome association studies can yield informative results in an inbreeding species such as wheat [4
], and this strategy has shown good potential in oat [22
]. This work would benefit greatly from increased map coverage, and from markers that can be scored efficiently and consistently.
Recently, a novel technique for the development and application of microarray-based molecular markers has been described [23
]. The patent for this technique, known as diversity array technology (DArT), is licensed freely under an open-source model [24
]. DArT has been applied successfully in several crops including barley [25
] and wheat [26
], and information on the current status of technology development is available online [27
]. Briefly, the DArT technique is based on isolating a random set of cloned DNA fragments from a complexity-reduced, pooled DNA sample. These clones are arrayed on a solid phase slide, where they selectively hybridize to complexity-reduced, PCR-amplified, genomic samples. Differential hybridization is usually a result of single-nucleotide polymorphisms (SNP) that affect the presence of restriction sites (and, therefore, certain amplified fragments) in the complexity reduction. A major advantage of DArT is that it provides a consistent high-throughput method whereby a complete set of markers with full genome coverage can be surveyed in parallel across many genomic samples. Because this technology is based on a set of cloned DNA fragments, these fragments can be sequenced and made accessible to the international research community. Furthermore, since the technology is freely available, the assay can be performed by an experienced provider at reasonable cost. However, the prerequisite for DArT analysis is the development and validation of a diagnostic DArT array.
The objectives of this study were: (1) to develop and describe a set of DArT markers giving complete genome coverage in hexaploid oat, (2) to revise and improve a hexaploid oat linkage map through incorporation of these new markers, and (3) to use these markers to analyse genetic diversity in a global collection of oat germplasm. Because this is the first report describing DArT marker analysis in oat, we also provide a detailed set of additional reference material to support future analyses utilizing these markers. Throughout this study, we refer to oat germplasm accessions collectively as 'varieties', regardless of whether they are cultivated varieties (cultivars), breeding lines, or experimental varieties.