INVITED REVIEW
Recent Progress in Genomic Analysis of Ornamental Plants, with a Focus on Carnation
2015 Volume 84 Issue 1 Pages 3-13
Details
2015 Volume 84 Issue 1 Pages 3-13
Genomic analysis and marker-assisted selection have long been familiar terms. Nevertheless, compared with that on other horticultural crops, genome-related research on ornamentals has been delayed because of the polyploid nature and/or highly heterozygous genetic background of many such species. With the advent of next-generation sequencing (NGS) technology in recent years, however, the situation is changing. The acquisition of comprehensive transcriptome sequences using NGS technology has been conducted in major ornamentals, and whole-genome sequences have been generated for carnation. This review discusses recent progress in the genomic analysis of carnation, including the construction of an SSR-based reference genetic linkage map, QTL analysis of carnation bacterial wilt (CBW) resistance, and the development of tightly linked markers for CBW resistance and flower type. The current state of NGS technology-based genomic research is also summarized for other major ornamentals.
Successful plant breeding, which refers to genetic crop improvement and the production of new cultivars, has long been dependent on the experience and intuition of breeders. With advances in our knowledge of plants at the molecular level, however, DNA marker-assisted breeding approaches have been developed. Marker-assisted selection (MAS), a method that uses DNA markers closely linked to relevant traits, holds great promise as a means to raise the efficiency of the breeding line selection process. MAS enables accurate selection unaffected by environmental factors and is even possible at the seedling stage as long as DNA can be extracted. Various types of PCR-based marker have been developed for use in numerous ornamentals, including random-amplified polymorphic DNA (RAPD), amplified fragment-length polymorphism (AFLP), microsatellite (also termed simple sequence repeat; SSR), and single-nucleotide polymorphism (SNP) markers (Rout and Mohapatra, 2006). Applications of these markers have mainly focused on questions of genotyping, cultivar identification, and genetic diversity. In contrast, comparatively little work has been performed to analyze genome structure and gene function, and very few reports have dealt with the potential use of these markers for MAS or even cloning of genes (Debener, 2012). Because new ornamental cultivars are commonly produced by hybridization between elite cultivars and propagated vegetatively, the genetic background of most ornamentals is highly heterozygous, with polyploidy also being observed in some species. This situation complicates detailed genetic analysis using crossing populations; as a consequence, the development of sophisticated breeding strategies in ornamentals has lagged behind those for most agricultural crops. Furthermore, the economic importance of individual crops is relatively small compared with that of so-called cash crops in agriculture, with public funding and industrial support for applied research projects therefore being limited (Debener, 2012).
With the development of next-generation sequencing (NGS) technology, the situation has changed, and this new technology is now facilitating crop improvement (Deschamps and Campbell, 2010). The advent of NGS platforms has made it possible to read billions of base pairs of sequences in a single or a limited number of runs, thereby expediting marker development from genome and expressed sequence tag (EST) sequences.
In this review, I focus on our recent progress in carnation genome research. In addition, I discuss recent developments in genomic analysis resources and NGS applications to other major ornamentals.
Carnation (Dianthus caryophyllus L.) is one of the most popular flowers, not only in Japan but also worldwide. More than 600 million carnation flowers were consumed in Japan in 2012 (Ministry of Agriculture, Forestry and Fisheries, Japan). However, annual carnation production in Japan has gradually decreased. More than half of the carnations used in Japan are now imported, largely from Colombia and China, where production is carried out on a large scale in a low-cost manner and the climate is well suited to carnation production. To promote carnation production in Japan, scientists at the NARO Institute of Floricultural Science have worked to produce new cultivars with superior characteristics to those of imported cultivars. For example, Onozaki et al. (2006a) have produced ‘Miracle Rouge’ and ‘Miracle Symphony’, which have about three times the flower vase life of normal cultivars. As another example, Yagi et al. (2010) produced the new cultivar ‘Karen Rouge’, which has carnation bacterial wilt (CBW) resistance derived from D. capitatus Balbis ex DC. ssp. andrzejowskianus Zapal. ‘Karen Rouge’ is one of the few ornamental flower cultivars bred by MAS. The breeding process used for ‘Karen Rouge’ as well as related genomic research was reviewed by Yagi (2013). In the current review, I summarize additional recent progress in carnation genomic analysis.
1. Identification of tightly linked markers in the new CBW-resistant line 85-11CBW, caused by Burkholderia caryophylli (Burkholder) Yabuuchi, Kasako, Oyaizu, Yano, Hotta, Hashimoto, Ezaki, and Arakawa, is one of the most severe diseases affecting Japanese carnation production. To improve selection efficiency during the breeding of plants with CBW resistance using D. capitatus ssp. andrzejowskianus, Yagi et al. (2006) constructed the first genetic linkage map for carnation. This map comprised 137 RAPD and 9 SSR loci within 16 linkage groups (LGs). Quantitative trait locus (QTL) analysis identified one major (Cbw1) and two minor (Cbw2 and Cbw3) QTLs for CBW resistance in lines derived from D. capitatus ssp. andrzejowskianus. During breeding for improved flower longevity, Yagi et al. (2012) unexpectedly identified line 85-11, which exhibited a significant level of CBW resistance not derived from D. capitatus ssp. andrzejowskianus. To develop markers linked to the CBW resistance of line 85-11, Yagi et al. (2012) constructed an SSR-based genetic linkage map using F2 populations between line 85-11 and a susceptible cultivar (‘Pretty Favvare’). SSR markers are highly polymorphic, co-dominant, and transferable as anchor points for comparing linkage maps. A new carnation map was constructed that included 178 SSRs in 16 LGs covering 843.6 cM (Yagi et al., 2012). QTL analysis for CBW resistance revealed only one QTL in LG 85P_4 (Cbw4), with two tightly linked SSR markers identified adjacent to the Cbw4 locus: CES2643 and CES1161. The difference in mean disease incidence among the three groups categorized according to genotypes at the CES2643 locus in the F2 mapping population (i.e., those homozygous for the 85-11 allele, heterozygous, or homozygous for the susceptible cultivar allele) demonstrated the existence of a clear correlation between marker genotypes and disease incidence (Fig. 1). These SSR markers, which are tightly linked to CBW resistance, can be directly used in MAS to facilitate the introduction of desirable genes into certain cultivars.
Frequency distribution of mean disease incidence in carnation F2 mapping populations (85P) categorized according to genotypes at the SSR locus CES2643.
Comparative analysis of CBW resistance loci between LG 85P_4 (Cbw4) and LG NP_4 (Cbw1) using SSR markers revealed nearly identical positions in both LGs (Fig. 2). STS-WG44, which has been tightly linked to Cbw1 using sequence-tagged site (STS) markers (Onozaki et al., 2004), was mapped onto LG 85P_4. Interestingly, the map position of the major QTL for CBW resistance was similar between the normal carnation cultivar (Cbw4) and the wild Dianthus species (Cbw1). To determine whether Cbw4 and Cbw1 are homologous genes, further analysis using techniques such as a high density of markers around Cbw4 or Cbw1, or cloning of each gene, will be needed. A bacterial artificial chromosome (BAC) library of ‘Karen Rouge’ is currently being constructed, and cloning of the gene corresponding to Cbw1 will resolve this issue.
Genetic map of the genomic region containing major QTLs for carnation bacterial wilt resistance (Cbw4 and Cbw1). Marker names and map distances (cM) are shown on opposite sides of each linkage group. Cap1 is one of the QTLs governing carnation anthocyanin pigmentation. Markers located on both maps are underlined and connected by lines.
As mentioned above, an RAPD-based genetic linkage map (NP map) derived from D. capitatus ssp. andrzejowskianus and an SSR-based genetic linkage map (85P map) from line 85-11 had been constructed. Because the number of markers was insufficient, however, the number of LGs did not coincide with the number of chromosomes (x = 15). The production of a high-density genetic map would be useful for breeding purposes and genetic research. Yagi et al. (2013) accordingly refined the SSR-based genetic linkage map to include 412 SSR loci covering 978.3 cM, adding 234 new SSR markers derived from genomic SSR libraries and transcriptome analysis (RNA-seq; Tanase et al., 2012) via NGS technology. By adding 192 SSR, 8 RAPD, and 2 STS loci, they refined the RAPD-based genetic linkage map to comprise 15 LGs consisting of 348 loci covering 978.3 cM. The two maps had 125 SSR loci in common, with most marker positions conserved between them. Although the 85P map contained 17 LGs, the LGs comprising a small number of markers were integrated by connection with common SSR markers. The number of LGs in both maps thus corresponded to the haploid chromosome number (x = 15). Many SSR markers have therefore been mapped in carnation, a non-model crop. The improved genetic linkage maps and SSR markers currently developed will serve as a genetic reference for carnation and other members of the genus Dianthus, and should be useful for mapping QTLs associated with various traits and for improving carnation breeding programs.
3. Mapping of flower color and typeVarious flower colors and patterns exist in carnation. The main end products of flower pigments in carnation are anthocyanin glucosides derived from pelargonidin and cyanidin, and flavonol derivatives based on kaempferol and quercetin (Nakayama et al., 2000). Metabolism of these and other flavonoids has been well studied in carnation flowers (Itoh et al., 2002; Mato et al., 2000; Matsuba et al., 2010). Using a population segregated for CBW resistance, Yagi et al. (2013) identified two QTLs governing anthocyanin content in flower petals: carnation anthocyanin pigmentation loci 1 (Cap1) and 2 (Cap2) on LGs NP_4 and 10, respectively. Cbw1 was located on the same LG as Cap1, at a distance of 15.4 cM (Fig. 2). In the course of breeding for ‘Karen Rouge’, many resistant lines in the initial breeding populations were observed to have purplish flowers. This phenomenon could be attributed to linkage between the major resistance gene and the QTL for anthocyanin pigmentation. Mehlquist and Geissman (1947) showed that a basic factor (S) controls anthocyanin concentrations in carnation. Larsen et al. (2003) suggested that the mutation of glutathione S-transferase (GST), which is involved in the transportation of anthocyanins to the vacuole, is responsible for the pale anthocyanin coloration in carnation. Using the results of genome analysis of a mutable flower line bearing deep pink sectors on pale pink petals, Sasaki et al. (2012) determined that DcGSTF2 encoding GST-like protein in carnation is responsible for flower color intensity in carnation. The QTL for anthocyanin content may be related to such genes regulating the anthocyanin biosynthesis pathway or encoding GST. Future mapping of genes involved in anthocyanin biosynthesis may allow the determination of whether any of these genes correspond to the identified QTL.
With respect to carnation flower type, relatively little is known about the genetics of doubleness in carnation (Conners, 1913; Saunders, 1917). Nearly a century ago, Saunders (1917) suggested that the carnation flower phenotype is a monogenic trait and designated the locus involved as “D” (where the recessive homozygote allele [dd] is the single flower type, the heterozygote [Dd] is double, and the dominant homozygote [DD] is super double). Scovel et al. (1998) provided supporting evidence for this flower-type genetics and developed markers linked to the d allele. Onozaki et al. (2006b) identified the “single” locus originating from D. capitatus ssp. andrzejowskianus that controls the single flower type and developed dominant, linked RAPD markers. A simultaneous increase in petals and other floral organs is commonly observed in floricultural plants, including carnations, petunias, and garden balsam (Nishijima, 2012). Each floral organ in double flowers is fully functional. Through mutant analysis, molecular cloning, and functional characterization of the corresponding genes in model species such as Arabidopsis thaliana and Antirrhinum majus, several research groups have helped establish a model that shows the combinatory actions of four classes of homeotic gene (A, B, C, and E) in flower organ determination and development (Bendahmane et al., 2013). In carnation, Yagi et al. (2014b) located the D85 locus for flower type (double or single) on LG 85P_15-2 and identified four co-segregating SSR markers (Fig. 3). Among the four markers, Yagi et al. (2014b) concluded that CES1982 and CES0212 were tightly linked to the D85 locus, and that the 176-bp allele at CES1982 and the 269-bp allele at CES0212 were tightly linked to the dominant D allele responsible for the double flower phenotype. In rose, the locus controlling the simple versus double corolla phenotype, referred to as Blfo or d6, has been identified through genetic mapping studies (Crespel et al., 2002; Debener and Mattiesch, 1999). The responsible genes have not been reported, however. To identify the gene corresponding to the D allele, we are producing a BAC library for carnations and have sequenced the whole genome (Yagi et al., 2014a).
Comparison of genetic maps of the genomic region controlling flower type in carnation line 85-11 (D85) and single-flowering Dianthus capitatus ssp. andrzejowskianus (single). Marker names and map distances (cM) are shown on opposite sides of linkage groups. Markers located on both maps are underlined and connected by lines.
A large quantity of EST data has been generated for horticultural crops including vegetables and fruit trees as well as model plants such as Oryza sativa L. and Arabidopsis thaliana. Because ESTs only reflect genes that are expressed, their corresponding analyses are effective and relatively inexpensive. Only 669 carnation ESTs were available on the NCBI website as of early June 2012 (Tanase et al., 2012). Tanase et al. (2012) sequenced the transcripts from various vegetative tissues, flowers at different developmental stages, and ethylene-treated flowers from ‘Francesco’ by 454 sequencing. Clustering and assembly of the generated sequences resulted in 300,740 unigenes consisting of 37,844 contigs and 262,896 singletons. Transcripts were identified for almost every gene involved in flower chlorophyll and carotenoid metabolism and anthocyanin biosynthesis. Transcripts were also identified for every step in the ethylene biosynthesis pathway. A search for di-, tri-, tetra-, and pentanucleotide repeats identified 17,362 potential SSRs in 14,291 unigenes. Finally, 4,177 SSR primer pairs were designed from these unigenes. A subset of these primer pairs was used for linkage map construction (Yagi et al., 2013).
Ohmiya et al. (2013) prepared a custom oligonucleotide array based on the carnation EST database generated by Tanase et al. (2012) and compared carotenogenic gene expression levels in pale-green and white petals. Because plants in the order Caryophyllales, including carnation, do not accumulate carotenoids in petals, there are no carnation cultivars with deep-yellow flowers. Ohmiya et al. (2013) suggested that the low levels of carotenoids in carnation petals are caused not by enzymatic degradation but rather by low rates of carotenoid biosynthesis caused by the suppression of the genes encoding phytoene synthase and lycopene ε-cyclase. Such transcriptomic studies shed light on the complexity of gene expression and regulatory networks at various developmental stages, including flower opening and senescence, as well as response to biotic/abiotic stress.
5. Whole-genome sequencing of carnationRecent advances in high-throughput sequencing technology have already benefited whole-genome sequencing projects in non-model organisms. To our knowledge, no whole-genome sequencing analysis has been reported for ornamentals other than ‘Francesco’, the leading red carnation cultivar in Japan. Transcriptome sequences of ‘Francesco’ were previously collected by Tanase et al. (2012). Most carnation cultivars are diploid with a chromosome number of 2n = 2x = 30 (Gatt et al., 1998; Yagi et al., 2007). The genome size of carnation (~685 Mb)—as estimated by Figueira et al. (1992), Nimura et al. (2003), and Agulló-Antón et al. (2013)—is very small compared with those of other ornamental flowers listed in the Plant C-values database (http://data.kew.org/cvalues/), such as Rosa hybrida (1.1 Gb), Antirrhinum majus (1.5 Gb), Petunia hybrida (1.6 Gb), Chrysanthemum morifolium (9.4 Gb), and Tulipa gesneriana (26 Gb). We determined the whole genome of ‘Francesco’ using a combination of new-generation multiplex sequencing platforms (Yagi et al., 2014a). The genome size of carnation ‘Francesco’, as estimated by k-mer analysis of HiSeq 1000 sequence data, was 622 Mb, about 90% of the previous estimate. The total length of non-redundant sequences was 569 Mb, consisting of 45,088 scaffolds, which covered 91% of the 622-Mb carnation genome estimated by k-mer analysis. The N50 values of contigs and scaffolds were 16,644 bp and 60,737 bp, respectively, and the longest scaffold was 1.29 Mb. Comparison of the independently determined sequences of the two BAC clones with the assembled genomic sequences showed perfect alignment with correct order and coverage, demonstrating that the coverage and quality of the assembled genomic sequences were high. We also revealed the correlation between the genomic sequences and their positions on the carnation reference genetic linkage map comprising 412 SSR loci. Single corresponding scaffolds could be identified for 378 (91.7%) of the 412 SSR loci and the remaining SSR loci were assigned to multiple scaffolds containing identical or highly similar sequences. Consequently, 268 scaffolds could be located on the genetic linkage map. The total length of the mapped scaffolds was 51.4 Mb, equivalent to 8.3% of the estimated genome size.
For protein-encoding genes, 43,266 complete and partial gene structures, excluding those in transposable elements, were deduced. Gene coverage was approximately 98%, as deduced from the coverage of the core eukaryotic genes. Intensive characterization of the assigned carnation genes and comparison with those of other plant species revealed characteristic features of the carnation genome. Information about the obtained genomic sequences is freely available online at Carnation DB (http://carnation.kazusa.or.jp/) (Kazusa DNA Research Institute).
The information and material resources generated by whole-genome sequencing of carnation will enhance both fundamental and applied research on carnations and related plants. Surprisingly, the day after the publication of the carnation genome, genome sequences were reported for sugar beet (Beta vulgaris), which belongs to the Amaranthaceae family within Caryophyllales (Dohm et al., 2014). The flower pigments of species belonging to families in Caryophyllales, except for Caryophyllaceae and Molluginaceae, are betalains; these pigments have never been detected along with anthocyanins in the same species (Sasaki et al., 2009). The mutual exclusiveness of anthocyanins and betalains in the order Caryophyllales has sparked considerable taxonomic debate (Clement and Mabry, 1996) and represents an interesting unresolved issue concerning flower pigmentation (Tanaka et al., 2008). Comparison of the sequences of carnation, belonging to one of the two families bearing anthocyanins in Caryophyllales, with those of betalain-containing sugar beet should help resolve this issue.
Chrysanthemum (C. morifolium) is one of the world’s most important ornamental crops, especially in Japan. Most cultivated chrysanthemum varieties are hexaploid (2n = 6x = 54), with somatic chromosome numbers ranging from 2n = 47 to 63 both between and within plants (Anderson, 2006; Dowrick, 1953). The genome of C. morifolium is estimated to be approximately 9.4 Gb (http://data.kew.org/cvalues/). Genetic improvement of chrysanthemum is hampered mainly by its genomic complexity, high levels of heterozygosity, and the occurrence of both inbreeding depression and self-incompatibility. Because of its large, complex genome and complicated genetic background, very few genomic and genetic resources are currently available for chrysanthemum (Xu et al., 2013). Genetic analysis using mapping populations has been conducted mainly in China. Zhang et al. (2010) reported a preliminary genetic linkage map of chrysanthemum cultivars using RAPD, ISSR, and AFLP markers. QTL analysis has been conducted for inflorescence-related traits, flowering architecture, and flowering time (Zhang et al., 2011a, b, 2012, 2013a). Using Illumina sequencing technology, Wang et al. (2013) collected whole-transcriptome data from the diploid species C. nankingense (Nakai) Tzvel and designed 1,788 EST-SSR primer pairs. Also using Illumina technology, Xu et al. (2013) performed large-scale transcriptome sequencing of chrysanthemum plants under dehydration stress. A Japanese research team recently detected genes encoding florigens and antiflorigens, the key regulators of photoperiodic flowering in chrysanthemum, using the wild diploid chrysanthemum C. seticuspe (Higuchi et al., 2013; Oda et al., 2012). Information from wild diploid species should help resolve the genetic complexity of chrysanthemum and contribute to genetic analysis of important traits, map-based gene cloning, and efficient breeding of polyploid cultivars.
In terms of genomic analysis, rose (R. hybrida) is the most advanced ornamental species. Ploidy levels of rose species range from 2x to 8x, with the majority of wild species being diploid and most cultivars being tetraploid (2n = 4x = 28) (Debener and Linde, 2009). Although high ploidy levels and heterozygosity cause genome complexity, many linkage maps have been reported for diploid (Crespel et al., 2002; Debener and Mattiesch, 1999; Dugo et al., 2005; Hosseini Moghaddam et al., 2012; Spiller et al., 2010; Yan et al., 2005) and tetraploid rose (Hibrand-Saint Oyant et al., 2008; Kawamura et al., 2011; Rajapakse et al., 2001; Zhang et al., 2006). An integrated consensus map for diploid rose has also been reported (Spiller et al., 2011). As a member of the Rosaceae family, rose is related to important fruit crops including strawberry, apple, peach, and cherry. Draft genomic sequences have been completed for the major rosaceous fruit crops apple (Velasco et al., 2010), strawberry (Hirakawa et al., 2014; Shulaev et al., 2011), and peach (Verde et al., 2013). Rosaceae genomic, genetic, and breeding data and analysis tools to facilitate basic, translational, and applied Rosaceae research are available in the Genome Database for Rosaceae (GDR) at http://www.rosaceae.org/. This information has been directly applied to the development of markers and comparative genomic analysis based on macro- and microsynteny. The close genetic relationship between Fragaria and Rosa allowed an autotetraploid linkage map to be constructed for rose using the strawberry genome sequence (Gar et al., 2011). In recent years, the R. chinensis cultivar ‘Old Blush’ has been chosen as a model for the development of genomic and genetic transformation tools (Dubois et al., 2012). This diploid recurrent-flowering rose is a common ancestor of many commercial modern roses and has contributed to recurrent-flowering and tea-scent traits (Dubois et al., 2012). Using a combination of Illumina and 454 sequencing technologies, Dubois et al. (2012) generated information on Rosa sp. transcripts based on this cultivar. Kim et al. (2012) collected transcripts from three rose cultivars and R. rugosa Thunb., and Yan et al. (2014) acquired transcript data from another R. chinensis cultivar. Pei et al. (2013) created a rose floral transcriptome using 454 technology and exploited the transcriptome to identify potential key regulators of ethylene-influenced cell expansion. BAC library construction is a powerful tool for the cloning of target genes from adjacent markers and the development of more tightly linked markers to target traits. Construction of BAC libraries for rose and sequencing of their clones have been conducted (Biber et al., 2010; Kaufmann et al., 2003). Terefe-Ayana et al. (2012) sequenced BAC clones, including the region of the rose black spot resistance gene, using 454 technology. As almost all genomic analysis tools, such as linkage maps, RNA-seq data, and BAC clones, have already been established in rose, the publication of rose genome sequences is expected in the near future.
The monocot perennial herbs lily and tulip (Liliaceae family) are favorite ornamentals both in Japan and worldwide. Flowering from seed takes about 5 years in tulip and 2–3 years in lily. These long growth cycles hamper efficient breeding. Lily and tulip are also very interesting from an evolutionary point of view, as both species have huge genomes (1C = 25 Gb in tulip and 36 Gb in lily; Shahin et al., 2012b). In lily, linkage maps have been generated using RAPD, AFLP, ISSR, diversity arrays technology, and nucleotide binding site markers, and QTL analysis has been conducted for flower color (Abe et al., 2002; Nakano et al., 2005) and disease resistance (Jansen, 1996; Shahin et al., 2011). Large-scale EST sequencing was recently conducted using NGS technology for both species from the Netherlands (Shahin et al., 2012b). The resulting data were applied for SNP discovery in lily and tulip (Shahin et al., 2012a; Tang et al., 2013).
Eustoma grandiflorum (Raf.) Shinn. (lisianthus) belongs to the Gentianaceae family and originated in North America (Anderson, 2006). Unlike other ornamentals, Eustoma is one of the few plant products for which Japanese breeding companies have taken the lead in regard to new cultivar developments and seed propagation. Genetic information, such as DNA markers and linkage maps, has not yet been reported. Kawabata et al. (2012) conducted large-scale EST sequencing using 454 sequencing and designed a 60-mer oligo-array representing the EST assemblies to analyze transcriptome changes in the petals of E. grandiflorum during flower opening.
The Orchidaceae constitute the largest family of flowering plants, with the number of species possibly exceeding 25,000 (Hsiao et al., 2011). Orchids are important export ornamentals in Taiwan. Hence, genomic analysis using NGS technology was conducted relatively early in these ornamentals. EST databases for Phalaenopsis (Fu et al., 2011; Hsiao et al., 2011; Su et al., 2011; Tsai et al., 2013) and Oncidium (Chang et al., 2011) have been obtained using NGS technology. An integrated database, Orchidstra, has recently been constructed (http://orchidstra.abrc.sinica.edu.tw/; Su et al., 2013a). The Orchidstra database contains transcriptome information for five orchid species and one commercial hybrid as well as microRNA (Chao et al., 2014) and microarray-based expression profiling data for a potential orchid model species, P. aphrodite (Su et al., 2013b). Transcriptome analysis of Cymbidium using Illumina sequencing has been reported from China (Li et al., 2013; Zhang et al., 2013b). These massive collections of genomic information will contribute to the understanding of orchids and their great biodiversity, vast geographical distribution, and variations in morphology, physiology, habitat, and ecosystem interactions.
As reviewed above, most of the massive transcriptome datasets generated from major ornamentals have been obtained using NGS technology (Table 1). Sequencing of BAC clones (Terefe-Ayana et al., 2012) and SSR-enriched libraries (Nakatsuka et al., 2012) was also conducted using NGS technology (Table 1). Although whole-genome sequence analysis has been completed at present only for carnation, genomes of other ornamentals should be available in the near future.
Genomic research studies using next-generation sequencing technology in ornamentals.
After the identification of reference genome sequences, comparison of whole-genomic regions between cultivars or between lines in a particular species is achievable. This development will be a revolutionary step in the genomic analysis of ornamentals, which has been delayed for the past few years compared with that in other crops. Many ornamentals are marked by high ploidy levels (chrysanthemum and rose) or huge genome sizes (lily and tulip). Consequently, restriction site-associated DNA sequencing (Baird et al., 2008; Miller et al., 2007) or genotyping-by-sequencing (Elshire et al., 2011) using NGS technology will be useful to reduce genome complexity (Debener, 2012). To use genomic information effectively for the development of useful markers and the identification of responsible genes, more precise and more rapid evaluation of traits will be needed. Advances in instrumentation such as sequencers and associated computers are rapidly occurring, and various types of software package for analysis of the resulting data are also being developed. The application of techniques and the operation of specialized equipment are very difficult for individual researchers and institutions. In every discipline, collaboration between researchers with different areas of expertise is essential. Such cooperation among experts using NGS is especially important with respect to ornamentals, as available personnel and financial resources were initially limited and efforts in this field are thus still at an early stage.
