Population structure and association analyses of the core collection of hexaploid accessions conserved ex situ in the Japanese gene bank NBRP-Wheat

ABSTRACT

In this study, we investigated the genetic diversity and population structure of the core collection of hexaploid wheat accessions in the Japanese wheat gene bank NBRP-Wheat. The core collection, consisting of 188 accessions of Triticum aestivum, T. spelta, T. compactum, T. sphaerococcum, T. macha and T. vavilovii, was intensively genotyped by DArTseq markers and consisted of 20,186 SNPs and 60,077 present and absent variations (PAVs). Polymorphic markers were distributed in all chromosomes, with a tendency for smaller numbers on the D-genome chromosomes. We examined the population structure by Bayesian clustering and principal component analysis with a general linear model. Overall, the core collection was divided into seven clusters. Non-admixture accessions in each cluster indicated that the clusters reflect the geographic distribution of the accessions. Both structure analyses strongly suggested that the cluster consisting of T. spelta and T. macha is out-grouped from other hexaploid wheat accessions. We performed genome-wide association analysis pilot studies for nine quantitative and seven qualitative traits and found marker-trait associations for all traits but one, indicating that the current core collection will be useful for detecting uncharacterized QTLs associated with phenotypes of interest.

INTRODUCTION

Along with maize and rice, wheat (Triticum spp.) is one of the most important staple crops that have supported human civilization. Hexaploid wheat (T. aestivum L.) originated from hybridization between domesticated tetraploid wheat (T. turgidum L.) and Aegilops tauschii Coss (Kihara, 1944; McFadden and Sears, 1944). Accumulated archeological and paleontological evidence indicates that the hybridization event occurred either south or west of the Caspian Sea approximately 8,000 years ago (Nesbitt and Samuel, 1996; Salamini et al., 2002; Giles and Brown, 2006). Today, hexaploid wheat is cultivated widely in subarctic, temperate and subtropical regions of both hemispheres, from 67°N in Norway, Finland and Russia to 45°S in Argentina, and accounts for more than 30% of the global cereal cultivation area (220 million ha in 2016, http://faostat.fao.org/). The production of wheat has increased since the release of modern high-yielding varieties and application of chemical fertilizers in the 1950s. The great success in raising wheat production has been referred to as a “green revolution”. Today, more than 700 million tonnes of wheat is harvested annually (749 million tonnes in 2016, http://faostat.fao.org/). However, the increase in wheat production has plateaued in recent decades and there is an urgent need to promote sustainable wheat production to meet the demands of the increasing world population amid global climate change (McCouch et al., 2013). Diverse genetic resources are essential for breeding new cultivars. The landraces cultivated globally are especially important as a genetic resource because they have adapted to various environments (Harlan, 1975; McCouch et al., 2013). The National Bioresources Project-Wheat (NBRP-Wheat), launched by the Japanese government in 2002, collects, stores and supplies wild species, landraces and experimental strains of wheat and related species (currently providing ca. 12,000 accessions). The collections of wild species and landraces derived from several expeditions since Dr. Hitoshi Kihara’s first expedition in 1955 are currently available rarely at the collection sites.

Association mapping is a method of detecting quantitative trait loci (QTLs). This approach is based on correlations between genotypes and phenotypes, and is additionally affected by the structure of linkage disequilibrium (LD) within a test population (Buckler and Thornsberry, 2002; Zondervan and Cardon, 2004). In addition to QTL analysis, which is based on phenotypic and genotypic polymorphisms within a bi-parental mapping population, association mapping is also applicable to a much larger gene pool and a more diverse genetic structure (Remington et al., 2001). Whole-genome association mapping has been applied to common wheat for several agronomic traits (Breseghello and Sorrells, 2006; Crossa et al., 2007; Neumann et al., 2011; Le Gouis et al., 2012; Edae et al., 2014; Wang et al., 2014; Zanke et al., 2014, 2015, 2017; Li et al., 2016). However, these studies focused on a common wheat (T. aestivum) population but not on other hexaploid Triticum species.

In this study, we provide intensive genotype information for the NBRP hexaploid wheat core collection to elucidate its genetic diversity and population structure, which is a prerequisite for association studies. We performed genotyping on the core collection lines using a DArT marker system (Jaccoud et al., 2001). The core collection includes modern cultivars frequently used for breeding in Japan, and landraces cover the geographical distribution and taxonomic groups of hexaploid wheat. Therefore, the core collection is a comprehensive hexaploid wheat collection in terms of systematics of the genus Triticum, i.e., it represents not only common wheat but also other hexaploid wheat species. We aimed to test the applicability of the core collection for association analyses of quantitative and qualitative traits.

MATERIALS AND METHODS

Plant materials and growth conditions

Plant materials used in this study are the collection of 188 accessions of hexaploid wheat (Table 1), comprised of T. aestivum (T. aestivum ssp. aestivum L., 160 accessions), T. spelta (T. aestivum ssp. spelta (L.) Thell., 12 accessions), T. compactum (T. aestivum ssp. compactum (Host) MacKey, six accessions), T. macha (T. aestivum ssp. macha (Dekapr. & Menabde) MacKey, four accessions), T. sphaerococcum (T. aestivum ssp. sphaerococcum (Percival) MacKey, three accessions), T. vavilovii (T. aestivum var. vavilovii Jakubz., two accessions) and synthetic hexaploid wheat (one accession, W7984). The synthetic wheat W7984 (Altar84/Aegilops tauschii (219) CIGM86.940) is the parental accession of the mapping population distributed by the International Triticeae Mapping Initiative (ITMI). In this report, we used nomenclature according to the National BioResource Project (NBRP)-Wheat catalogue. Detailed information on accessions is available on the NBRP-Wheat database KOMUGI (http://www.shigen.nig.ac.jp/wheat/komugi/). The core collection includes both landraces and modern cultivars, and they were selected from the NBRP-Wheat hexaploid wheat collection based on their taxonomic and geographical distribution. In this paper, we refer to the accessions with evident breeding pedigrees as cultivars. Seeds were sown in January, 2011 and the seedlings were transplanted to pots (one plant per pot). The pots were placed in a glasshouse without air conditioning at Yoshida Campus, Kyoto University, Japan (N35.031, E135.79).

Table 1. NBRP-Wheat hexaploid wheat core collection lines and cluster IDs based on population structure analysis

Acc_ID*	Species	Cultivar (strain no.)	Country (region)	Cluster_ID
6×001	T. aestivum	(KU-166)	China (Inner Mongolia)	E
6×002**	T. aestivum	Norin 10 (KU-265)	Japan	A
6×003**	T. aestivum	Baart (KU-309)	USA	D
6×004**	T. aestivum	Selkerk (KU-333)	Canada	E
6×005**	T. aestivum	Wichita (KU-336)	USA	E
6×006**	T. aestivum	April Bearded (KU-366)	UK	F
6×007**	T. aestivum	Renown (KU-370)	UK	F
6×008**	T. aestivum	Wilhelmina (KU-371)	UK	F
6×009**	T. aestivum	Little Joss (KU-372)	UK	F
6×010**	T. aestivum	Victor (KU-373)	UK	F
6×011**	T. aestivum	Yeoman (KU-374)	UK	F
6×012	T. aestivum	(KU-405)	Uzbekistan	C
6×013	T. aestivum	(KU-479)	China (Sichuan)	B
6×014	T. aestivum	(KU-481)	China (Shaanxi)	E
6×015	T. aestivum	(KU-483)	Tanzania	E
6×016	T. aestivum	(KU-497)	India	E
6×017	T. aestivum	(KU-504)	China	B
6×018	T. aestivum	Gifu Wase Komugi (KU-601)	Japan	A
6×019	T. aestivum	(KU-1002)	Spain	E
6×020	T. aestivum	(KU-1005)	Spain	F
6×021	T. aestivum	(KU-1011)	Spain	D
6×022	T. aestivum	(KU-1049)	Spain	F
6×023	T. aestivum	(KU-1143)	Spain	F
6×024	T. aestivum	Akita Zairai (KU-1215)	Japan	F
6×025	T. aestivum	Fukoku (KU-1230)	Japan	A
6×026	T. aestivum	Murasaki Aka (KU-1279)	Japan	F
6×027	T. aestivum	(KU-1302)	Greece	F
6×028	T. aestivum	(KU-1347)	Greece	D
6×029	T. aestivum	(KU-1392)	Romania	F
6×030	T. aestivum	(KU-1394)	Romania	F
6×031	T. aestivum	(KU-1421)	Romania	F
6×032	T. aestivum	(KU-1424)	Romania	F
6×033	T. aestivum	(KU-1521)	Azerbaijan	F
6×034	T. aestivum	(KU-1527)	Armenia	E
6×035	T. aestivum	(KU-1644)	Armenia	D
6×036	T. aestivum	(KU-1668)	Georgia	F
6×037	T. aestivum	(KU-1697)	Georgia	F
6×038	T. aestivum	(KU-1797)	Georgia	F
6×039	T. aestivum	(KU-3006)	Pakistan	C
6×040	T. aestivum	(KU-3010)	Pakistan	C
6×041	T. aestivum	(KU-3037)	Pakistan	C
6×042	T. aestivum	(KU-3045)	Afghanistan	C
6×043	T. aestivum	(KU-3054)	Afghanistan	C
6×044	T. aestivum	(KU-3062)	Afghanistan	C
6×045	T. aestivum	(KU-3083)	Afghanistan	C
6×046	T. aestivum	(KU-3089)	Afghanistan	C
6×047	T. aestivum	(KU-3097)	Iran	C
6×048	T. aestivum	(KU-3098)	Iran	C
6×049	T. aestivum	(KU-3121)	Iran	C
6×050	T. aestivum	(KU-3126)	Iran	C
6×051	T. aestivum	(KU-3136)	Iran	C
6×052	T. aestivum	(KU-3162)	Iran	C
6×053	T. aestivum	(KU-3184)	Iran	D
6×054	T. aestivum	(KU-3189)	Iran	D
6×055	T. aestivum	(KU-3202)	Iran	C
6×056	T. aestivum	(KU-3232)	Iran	D
6×057	T. aestivum	(KU-3236)	Iran	D
6×058	T. aestivum	(KU-3274)	Iran	C
6×059	T. aestivum	(KU-3289)	Iran	C
6×060	T. aestivum	(KU-3299)	Pakistan	B
6×061	T. aestivum	(KU-3351)	Pakistan	B
6×062	T. aestivum	(KU-3752)	Egypt	B
6×063	T. aestivum	(KU-3777)	Jordan	D
6×064	T. aestivum	(KU-3778)	Lebanon	E
6×065	T. aestivum	(KU-3780)	Syria	E
6×066	T. aestivum	(KU-3784)	Turkey	E
6×067	T. aestivum	(KU-3789)	Turkey	F
6×068	T. aestivum	(KU-3801)	Turkey	E
6×069	T. aestivum	(KU-3806)	Turkey	E
6×070	T. aestivum	(KU-3818)	Turkey	D
6×071	T. aestivum	(KU-3834)	Turkey	D
6×072	T. aestivum	(KU-3851)	Turkey	D
6×073	T. aestivum	(KU-3857)	Turkey	F
6×074	T. aestivum	(KU-3860)	Turkey	D
6×075	T. aestivum	(KU-3868)	Italy	F
6×076	T. aestivum	(KU-4703)	Nepal	E
6×077	T. aestivum	(KU-4714)	Nepal	B
6×078	T. aestivum	(KU-4734)	Nepal	B
6×079	T. aestivum	(KU-4759)	Nepal	B
6×080	T. aestivum	(KU-4769)	Nepal	B
6×081	T. aestivum	(KU-4783)	Nepal	B
6×082	T. aestivum	(KU-7001)	Bhutan	B
6×083	T. aestivum	(KU-7041)	Bhutan	B
6×084	T. aestivum	(KU-7113)	Bhutan	B
6×085	T. aestivum	(KU-7180)	Bhutan	B
6×086	T. aestivum	(KU-7356)	Ethiopia	E
6×087	T. aestivum	(KU-7379)	Ethiopia	F
6×088	T. aestivum	(KU-7406)	Ethiopia	F
6×089	T. aestivum	(KU-7437)	Afghanistan	C
6×090	T. aestivum	(KU-7459)	Afghanistan	B
6×091	T. aestivum	(KU-7480)	Afghanistan	C
6×092	T. aestivum	(KU-7624)	Afghanistan	C
6×093	T. aestivum	(KU-7653)	Afghanistan	C
6×094	T. aestivum	(KU-7669)	Afghanistan	C
6×095	T. aestivum	(KU-9431)	Ethiopia	D
6×096	T. aestivum	(KU-9460)	Ethiopia	D
6×097	T. aestivum	(KU-9797)	Ethiopia	D
6×098	T. aestivum	(KU-9820)	Ethiopia	D
6×099	T. aestivum	(KU-9867)	Ethiopia	D
6×100	T. aestivum	(KU-10001)	Iraq	D
6×101	T. aestivum	(KU-10154)	Iraq	D
6×103	T. aestivum	(KU-10393)	Iran	D
6×105	T. aestivum	(KU-10439)	Iran	C
6×106	T. aestivum	(KU-10480)	Iran	D
6×107	T. aestivum	(KU-10510)	Iran	D
6×108	T. aestivum	(KU-11201)	Afghanistan	D
6×109	T. aestivum	(KU-11214)	Afghanistan	D
6×110	T. aestivum	(KU-11240A)	Afghanistan	C
6×111	T. aestivum	(KU-11351)	Romania	D
6×112	T. aestivum	(KU-11702)	Greece	F
6×113	T. aestivum	(KU-11809)	Greece	D
6×114	T. aestivum	(KU-11829)	Greece	D
6×115	T. aestivum	(KU-13501)	China (Sichuan)	F
6×116	T. aestivum	(KU-13506)	China (Tibet)	F
6×117	T. aestivum	(KU-13546)	China (Tibet)	B
6×118	T. aestivum	(KU-13631)	China (Tibet)	B
6×119	T. aestivum	(KU-13662)	China (Tibet)	B
6×120	T. aestivum	(KU-13708)	China (Tibet)	B
6×121	T. aestivum	(KU-13807)	China (Tibet)	B
6×122	T. aestivum	(KU-13891)	China (Tibet)	B
6×123	T. spelta	(KU-1020)	Spain	G
6×124	T. spelta	(KU-1062)	Spain	G
6×125	T. spelta	(KU-1137)	Spain	G
6×126	T. spelta	(KU-3377)	Iran	C
6×127	T. spelta	(KU-3401)	Germany	G
6×128	T. spelta	(KU-3413)	Germany	G
6×129	T. spelta	(KU-3416)	Germany	G
6×130	T. spelta	(KU-3417)	Germany	G
6×131	T. spelta	(KU-3421)	Germany	G
6×132	T. spelta	(KU-3443)	Germany	G
6×133	T. spelta	(KU-3444)	Germany	G
6×134	T. spelta	(KU-3445)	Germany	G
6×135	T. compactum	(KU-152)	China (Inner Mongolia)	F
6×136	T. compactum	(KU-1208)	Japan	A
6×138	T. compactum	(KU-3063)	Afghanistan	C
6×139	T. compactum	(KU-3242)	Iran	D
6×140	T. compactum	(KU-7350)	Turkey	C
6×141	T. compactum	(KU-9873)	Ethiopia	–
6×142	T. macha	(KU-197)	Turkey	G
6×143	T. macha	(KU-1812)	Georgia	G
6×144	T. macha	(KU-1814)	Georgia	G
6×145	T. macha	(KU-1817)	Georgia	G
6×146	T. sphaerococcum	(KU-161)	–	B
6×147	T. sphaerococcum	(KU-162-2)	Pakistan	B
6×148	T. sphaerococcum	(KU-3004)	Pakistan	B
6×149	T. vavilovii	(KU-192)	–	C
6×150	T. vavilovii	(KU-3848)	Turkey	C
6×151**	T. aestivum	Opata 85	Mexico	E
6×152**	Synthetic	W7984	–	G
6×153	T. aestivum	Chinese Spring (LPGKU2269)	China	B
6×154	T. spelta	var. duhamerianum (KT019-001)	Germany	G
6×155**	T. aestivum	Timstein	Australia	E
6×156**	T. aestivum	Hope	USA	E
6×157**	T. aestivum	Cheyenne	USA	F
6×158**	T. aestivum	Hokkai 240	Japan	F
6×159**	T. aestivum	Kanto 107	Japan	A
6×160**	T. aestivum	Norin 26 (LPGKU2270)	Japan	A
6×161**	T. aestivum	Norin 61 (LPGKU2271)	Japan	A
6×163**	T. aestivum	Chihokukomugi	Japan	E
6×164**	T. aestivum	Nanbukomugi	Japan	–
6×165**	T. aestivum	Zenkojikomugi	Japan	A
6×166**	T. aestivum	Sumai 3	China	A
6×167**	T. aestivum	Nobeokabouzu	Japan	A
6×168**	T. aestivum	Tamaizumi	Japan	A
6×169**	T. aestivum	Kitanokaori	Japan	F
6×170**	T. aestivum	Minaminokaori	Japan	A
6×171**	T. aestivum	Haruyokoi	Japan	E
6×172**	T. aestivum	KS831957	USA	F
6×173**	T. aestivum	Hanamanten	Japan	A
6×175**	T. aestivum	Kinuiroha	Japan	A
6×176**	T. aestivum	U24	China	E
6×177**	T. aestivum	Seikai 165	Japan	A
6×178**	T. aestivum	Ayahikari	Japan	A
6×179**	T. aestivum	Minaminokomugi	Japan	A
6×180**	T. aestivum	Chogokuwase	Japan	A
6×181**	T. aestivum	Akimakigata-Abukumawase	Japan	A
6×182**	T. aestivum	Nishikazekomugi	Japan	A
6×183**	T. aestivum	Chikugoizumi	Japan	A
6×184**	T. aestivum	Iwainodaichi	Japan	A
6×185**	T. aestivum	Shiroganekomugi	Japan	A
6×186**	T. aestivum	Seikai 193	Japan	A
6×187**	T. aestivum	Kitakamikomugi	Japan	F
6×188**	T. aestivum	Nebarigoshi	Japan	A
6×189**	T. aestivum	Shunyou	Japan	A
6×190**	T. aestivum	Fujimikomugi	Japan	A
6×191**	T. aestivum	Gamenya	Australia	E
6×192**	T. aestivum	Akadaruma	Japan	A
6×193**	T. aestivum	Bobwhite	Mexico	F

*  Acc_ID stands for Accession ID.

**  Modern cultivar.

Phenotyping

We scored nine quantitative traits for every accession: heading time (HT), flowering time (FT), maturation time (MT), number of spikelets per spike (SN), spike length (SL), spike density (SD), culm length (CL), 100-grain weight (HGW) and number of tillers (TN); and seven qualitative traits: wax-like deposits on culm (WC), ochrea (WO) and spike (WS), anthocyanin pigmentation on coleoptile (AC) and seedling (AS), glume hair (GH), and awnedness (AWN). HT, FT and MT are expressed as the number of days after November 1^st. SD is obtained by dividing SN by SL. CL is the value of the longest culm of each individual. Every quantitative trait was measured for three individuals per accession and their means were used for analyses. The quantitative traits were rated on a three-point scale. Wax-like deposits and anthocyanin pigmentation on each part were evaluated by naked-eye observations. We also scored glume hair and awnedness on three-point scales: hairless, short-hairy or hairy; and awnless, short-awn or long-awn, respectively.

DNA extraction and genotyping

Total DNA was extracted from young leaves of individual plants using a DNeasy Plant Maxi Kit (Qiagen Tokyo, Japan) according to the manufacturer’s protocol. DArTseq marker genotyping, a genotype-by-sequencing method, was carried out by Diversity Arrays Technology, Yarralumla, Australia. The DArTseq marker system generates two types of data: single-nucleotide polymorphisms (SNPs) and presence and absence variations (PAVs). We removed SNPs with call rates less than 0.8 or with hetero fractions less than 0.05 in the ensuing analyses. The marker system provides a pair of 69-bp sequence (tag) information; DNA polymorphisms therein were recognized as alleles. To select high-confidence markers and to reveal on which chromosome the markers were located, we performed a BLAST search (Zhang et al., 2000) for tag sequences on the IWGSC RefSeq ver. 1.0 (https://wheat-urgi.versailles.inra.fr/Seq-Repository) (International Wheat Genome Sequencing Consortium [IWGSC], 2018). We adopted only polymorphic markers with one of the tag sequences matched perfectly to a single position on a chromosome.

Statistical analyses

Polymorphism information content (PIC) values were calculated for each marker according to Roldán-Ruis et al. (2000), as follows: PIC_i = 2f_i (1 – f_i), where PIC_i is the PIC of marker ‘i’, f_i is the frequency of an allele and 1 – f_i is the frequency of another allele. Missing data were excluded from calculations.

We estimated the population structure by Bayesian clustering using the program ADMIXTURE 1.3 (Alexander et al., 2009). To estimate the optimal value of K, which describes the number of subpopulations making up the total population, we calculated a 10-fold cross-validation procedure for each K value from 1 to 15. Principal component analysis (PCA) was also performed to estimate population structure using the pipeline implemented in the software TASSEL 5.0 (Bradbury et al., 2007). Missing data were imputed to the mean score for each locus (Jolliffe, 2002; Reich et al., 2008).

Inference on associations between genotypes and phenotypes was made using the pipeline implemented in TASSEL 5.0. We removed rare polymorphic markers that appeared in less than 5% of the accessions (PIC < 0.095) from analyses and used principal components (PC1 to 5) obtained from PCA to infer population structures. We performed analyses with an eigenvalue decomposition of the covariance matrix. We used a general linear model (GLM) for estimation of regression between genotypes and phenotypes. The permutation analysis was run 1,000 times.

RESULTS

Genotyping of NBRP-Wheat hexaploid wheat core collection by the DArT marker system

The DArTseq genotyping system was applied for 188 accessions. We could not obtain genotyping data from one common wheat accession (6×164: T. aestivum ‘Nanbukomugi’), and obtained less than 80% data from one T. compactum accession (6×141: KU-9873). We removed these two accessions from the ensuing analyses. We obtained 34,234 SNP data sets by DArTseq genotyping. We removed low-confidence SNPs with call rates less than 0.8 or with heterozygous fractions greater than 0.05 from the SNP data sets. The mode of the frequency of missing data was 0.00–0.05 (Supplementary Fig. S1A). A total of 20,186 SNPs were selected and subjected to BLAST searches against IWGSC RefSeq ver. 1.0. A total of 13,394 tag sequences matched to single regions on the reference sequences. We obtained 5,596 SNPs located on the A genome, 5,820 on the B genome, 1,768 on the D genome, and 210 on one pseudomolecule composed of unanchored scaffolds (chrUn, Table 2). The number of SNPs located on each chromosome ranged from 91 on chromosome 4D to 1,198 on chromosome 7A (Table 2). There were fewer SNPs located on homoeologous group-4 chromosomes than on other chromosomes in each genome (Table 2). The average PIC values for each chromosome ranged from 0.220 (chromosome 4D) to 0.252 (chromosome 3A) (Table 2).

Table 2. Summary of markers obtained in this study

Chromosome	SNP			PAV
	Number	PIC	kbp/SNP*	Number	PIC	kbp/PAV*
chr1A	749	0.245	790	876	0.240	676
chr2A	820	0.239	951	1,018	0.247	766
chr3A	803	0.252	932	986	0.214	759
chr4A	602	0.242	1,231	905	0.248	821
chr5A	845	0.243	839	967	0.229	733
chr6A	579	0.234	1,066	742	0.250	831
chr7A	1,198	0.234	615	1,193	0.245	616
A genome	5,596	0.241	–	6,687	0.239	–
chr1B	818	0.234	839	1,070	0.255	641
chr2B	1,043	0.234	768	1,304	0.256	614
chr3B	996	0.234	833	1,243	0.264	666
chr4B	411	0.233	1,637	641	0.205	1,049
chr5B	895	0.235	796	1,134	0.262	628
chr6B	790	0.239	912	1,078	0.266	668
chr7B	867	0.235	865	1,014	0.257	739
B genome	5,820	0.235	–	7,484	0.255	–
chr1D	288	0.223	1,713	738	0.170	669
chr2D	469	0.230	1,384	1,181	0.199	551
chr3D	259	0.239	2,373	933	0.145	658
chr4D	91	0.220	5,589	473	0.116	1,075
chr5D	178	0.222	3,166	825	0.140	685
chr6D	206	0.223	2,296	698	0.161	677
chr7D	277	0.241	2,298	1,051	0.153	606
D genome	1,768	0.230	–	5,899	0.159	–
chrUn**	210	0.239	–	453	0.267	–
All	13,394	0.237	–	20,523	0.223	–

SNP, PAV, and PIC are abbreviations for single-nucleotide polymorphism, presence and absence variation, and polymorphism information content, respectively.

*  Average distances between markers.

**  A pseudomolecule composed of unanchored scaffolds (IWGSC RefSeq ver. 1.0).

A data set of 60,077 PAVs was obtained by DArTseq genotyping and the data call rates for every PAV were more than 0.8. Overall, the mode of frequency of missing data was 0.000–0.025. For the genomically tagged PAVs, the mode of frequency of missing data was 0.125–0.150 (Supplementary Fig. S1B). BLAST searches against IWGSC RefSeq ver. 1.0 resulted in the matching of a total of 20,523 tag sequences to unique reference sequences. We obtained 6,687, 7,484, 5,899 and 453 PAV tags located on genomes A, B, D and one pseudomolecule composed of unanchored scaffolds, respectively (Table 2). The PAVs assigned to each chromosome ranged from 473 on chromosome 4D to 1,304 on chromosome 2B (Table 2). The average PIC values of PAVs for each chromosome ranged from 0.116 on chromosome 4D to 0.266 on chromosome 6B (Table 2).

A total of 2,326 SNPs and 3,084 PAVs matched to at least one gene model sequence (Table 3) (IWGSC RefSeq annotation ver. 1.1 https://wheat-urgi.versailles.inra.fr/Seq-Repository/Annotations). Due to the presence of multiple tag sequences in a gene model, we found a final total of 4,524 gene models that contained at least one SNP or PAV (Table 3).

Table 3. Number of markers in the gene model sequences

Chromosome	No. SNPs within gene models	No. gene models including SNPs	No. PAVs within gene models	No. gene models including PAVs	No. markers within gene models	No. gene models including markers
chr1A	126	114	129	115	255	204
chr2A	145	132	185	174	330	272
chr3A	107	99	119	116	226	204
chr4A	100	90	101	94	201	170
chr5A	130	115	133	130	263	224
chr6A	131	115	127	118	258	201
chr7A	165	146	134	126	299	236
A genome	904	811	928	873	1,832	1,511
chr1B	133	111	166	155	299	239
chr2B	193	174	213	198	406	334
chr3B	190	166	174	161	364	293
chr4B	92	82	98	91	190	162
chr5B	148	124	166	155	314	255
chr6B	156	142	168	160	324	266
chr7B	130	117	134	127	264	219
B genome	1,042	916	1,119	1,047	2,161	1,768
chr1D	48	44	124	118	172	154
chr2D	115	106	185	173	300	254
chr3D	45	41	161	151	206	180
chr4D	20	20	90	83	110	98
chr5D	25	24	134	127	159	143
chr6D	39	38	135	129	174	157
chr7D	44	41	149	140	193	173
D genome	336	314	978	921	1,314	1,159
chrUn*	44	38	59	56	103	86
All	2,326	2,079	3,084	2,897	5,410	4,524

SNP, and PAV are abbreviations for single-nucleotide polymorphism, and presence and absence variation, respectively.

*  A pseudomolecule composed of unanchored scaffolds (IWGSC RefSeq ver. 1.0).

Population structure analysis of the core collection

To investigate the population structure of the hexaploid wheat core collection, ADMIXTURE software was used for all the marker data. The optimal cluster number based on the cross-validation value is seven (Supplementary Fig. S2). Here, we name the seven clusters Pop-A to -G in the order of largest to smallest Q values (the ancestry fractions, Fig. 1 and Table 1).

Fig. 1.

Population structures based on DArTseq genotyping data. Analyses for K = 2 to 7 for the NBRP-Wheat hexaploid wheat core collection lines. The seven clusters, according to the largest Q value, are shown to the left of the bar plots. Boxes to the right indicate the non-aestivum exotic wheat accessions.

Pop-A consisted of 28 Eastern Asian T. aestivum (27 Japanese and one Chinese) accessions and one T. compactum accession (KU-1208) (Table 1). The 25 Pop-B accessions originated from southwest China, Himalayan regions, and Southern Asian countries. The Chinese Spring standard cultivar was classified into this cluster. Notably, all T. sphaerococcum accessions belong to this cluster (Table 1). Pop-C contains 30 accessions including all T. vavilovii (KU-192 and KU-3848), two T. compactum (KU-3063 and KU-7350), one T. spelta (KU-3377) and 25 T. aestivum. They are derived from Western Asian regions (Pakistan, Afghanistan, Uzbekistan, Iran and Turkey) (Table 1). Pop-D, consisting of 28 T. aestivum and one T. compactum (KU-3242 from Iran) accessions, is characterized by richness of admixture accessions, with more than half being admixture accessions with Pop-C or Pop-F (Fig. 1). These admixture accessions are derived from Western Asian and European regions, respectively. The accessions with higher ancestry fractions in this cluster (more than 0.8) are from Turkey, Iran and Ethiopia. Pop-E contains 22 accessions, the majority of which are also admixture accessions. We noted that they are modern cultivars from various regions (Fig. 1 and Table 1). The accessions with higher ancestry fractions in this cluster (more than 0.8) are from Turkey, Syria and Lebanon. All but one of the accessions in Pop-F (35 accessions) are T. aestivum, the exception being a T. compactum accession derived from China (KU-152). Many modern European cultivars are included in this cluster and some accessions in this cluster are admixture accessions with Pop-D and Pop-E (Fig. 1). All 17 accessions classified into Pop-G are non-aestivum (T. spelta derived from Europe (Germany and Spain), T. macha and synthetic hexaploid) (Table 1). Both T. macha and synthetic wheat accessions were admixtures with Pop-D (Fig. 1).

Principal component analysis of the core collection

We used PCA as an alternative method for investigating the population structure of the hexaploid wheat core collection. The first, second and third principal components (PC1 to PC3) explained respectively 8.1, 5.6 and 4.1% of their variation. Accessions plotted on the PC1-PC2 plane formed a triangular shape (Fig. 2A). Accessions classified as Pop-A and Pop-C clustered separately at two vertices (Pop-A at the bottom and Pop-C at the top right in Fig. 2A) of the triangle. Accessions identified as Pop-B were plotted between Pop-A and Pop-C (Fig. 2A). Accessions classified as Pop-F and Pop-G clustered together at the third vertex of the triangle (Fig. 2A). Accessions of Pop-D and Pop-E were plotted along the line between Pop-C and Pop-F (Fig. 2A).

Fig. 2.

Scatter plots based on principal component analysis. Colors of each plot represent the seven clusters based on the population structure analysis illustrated in Fig. 1. (A) PC1 vs PC2. (B) PC1 vs PC3. (C) PC2 vs PC3.

Accessions identified as Pop-G were divided into two sub-clusters according to PC1 and PC3 (Fig. 2B). These two sub-clusters correspond to a taxonomical discrimination between T. spelta and T. macha. Synthetic hexaploid wheat accession W7984 belongs to the sub-cluster with T. macha. On the other hand, scatter plots on the PC1-PC3 as well as on the PC2-PC3 planes show that Pop-G accessions are distinguishable from other accessions by PC3 (Figs. 2B, 2C).

Phenotype diversity and association mapping of the hexaploid wheat core collection

We measured nine quantitative traits, namely heading time (HT), flowering time (FT), maturation time (MT), number of spikelets per spike (SN), spike length (SL), spike density (SD), culm length (CL), 100-grain weight (HGW) and number of tillers (TN); and seven qualitative traits, namely wax-like deposits on culm (WC), ochrea (WO) and spike (WS), anthocyanin pigmentation on coleoptile (AC) and seedling (AS), glume hair (GH), and awnedness (AWN) (Tables 4 and 5). We performed marker-trait association analyses for these traits using both SNP and PAV data sets. We detected a total of 183 significant marker-trait associations (MTAs) for 127 SNPs and 140 for 96 PAVs at a significance level of 1% after Bonferroni multiple test correction. Fourteen SNP and eight PAV tag sequences found were within gene model sequences (Supplementary Table S1).

Table 4. Summary of quantitative traits of the NBRP-Wheat hexaploid wheat core collection lines

Trait	Max.	Min.	Mean	S.D.
HT (days)	230.0	155.7	187.3	14.2
FT (days)	232.0	164.3	190.5	12.9
MT (days)	248.0	209.7	224.4	8.2
CL (cm)	145.5	54.9	100.8	20.7
TN	34.0	7.0	16.6	4.9
SL (mm)	178.6	51.7	108.8	26.5
SN	29.0	10.7	18.2	3.3
HGW (g)	5.56	1.53	3.55	0.73
SD (SN/SL)	0.426	0.105	0.174	0.044

HT: heading time; FT: flowering time; MT: maturation time; CL: culm length; TN: number of tillers; SL: spike length; SN: number of spikelets per spike; HGW: 100-grain weight; SD: spike density.

Table 5. Summary of qualitative traits of NBRP-Wheat hexaploid wheat core collection lines

Trait	–	+	++
WC	8	95	87
WO	5	51	134
WS	14	114	62
AC	130	47	12
AS	151	34	4
GH	164	14	12
AWN	5	51	134

WC, WO and WS: wax-like deposits on culm, ochrea and spike, respectively; AC and AS: anthocyanin pigmentation on coleoptile and seedling, respectively; GH: glume hair; AWN: awnedness.

–: Non-waxed, non-pigmentated, hairless or awnless.

+: Few waxed, few pigmented, short-hairy or short-awn.

++: Waxed, pigmented, hairy or long-awn.

The 68 MTAs (38 SNPs and 30 PAVs) with HT were distributed on 20 chromosomes except for 4D (Fig. 3A and Table 6). Five SNPs and five PAVs are within gene model sequences (Supplementary Table S1). We detected 66 significant MTAs (37 SNPs and 29 PAVs) with FT on all chromosomes except 4D and 6A (Supplementary Fig. S3A and Table 6). Four SNPs and three PAVs reside in gene model sequences (Supplementary Table S1). We detected 50 MTAs (32 SNPs and 18 PAVs) with MT on 17 chromosomes (Supplementary Fig. S3B and Table 6). Six SNPs and one PAV were within gene model sequences (Supplementary Table S1). MTA analyses with the three flowering-related traits (HT, FT and MT) resulted in detection of a similar set of markers: 26 markers (17 SNPs and nine PAVs) were commonly associated with all three flowering-related traits, 29 markers (13 SNPs and 16 PAVs) with the HT and FT traits, one SNP with the HT and MT traits, and three markers (two SNPs and one PAV) with the FT and MT traits (Supplementary Table S1). As expected, the phenotypic scores of the three flowering-related traits are highly correlated with each other (Supplementary Fig. S4).

Fig. 3.

Manhattan plots of marker-trait associations. Blue, red and green circles indicate markers on the A, B and D genomes, respectively. Purple circles indicate markers on a pseudomolecule composed of unanchored scaffolds (IWGSC RefSeq ver. 1.0). In each chromosome, mapped markers are sorted according to genetic positions (from the tip of the short arm to the tip of the long arm, according to IWGSC RefSeq ver. 1.0). The vertical axis shows the negative logarithm of the association P-value. (A) Association with heading time (HT). (B) Association with culm length (CL). (C) Association with awnedness (AWN). (D) Association with anthocyanin pigmentation on coleoptile (AC).

Table 6. Summary of marker-trait associations (MTAs) detected in this study and of contigs, including significant marker tags

Chromosome	Quantitative trait									Qualitative trait					Total	Within gene model
	HT	FT	MT	CL	TN	SL	SN	HGW	SD	WO	AC	AS	GH	AWN
chr1A	3|1	3|1	2|0	1|0	1|0							2|0	2|1		14|3	3|0
chr1B	3|0	3|1	2|1	1|0								1|1			10|3
chr1D	1|1	1|0	1|0	2|0								1|4			6|5	0|1
chr2A	2|1	1|1	2|2	1|0	1|0				1|1			1|1	1|0	1|0	11|6	1|1
chr2B	3|1	3|0	3|0	1|1	0|2				1|0	0|1		2|0	1|0		14|5	1|0
chr2D	2|1	2|1	2|0	3|0	0|1	0|1	1|0					0|1			10|5	4|0
chr3A	2|3	2|3	1|2	1|0									0|1		6|9
chr3B	2|3	1|4	2|2	2|1	0|1		1|1					1|0			9|12	1|0
chr3D	1|0	1|0				1|0							1|0		4|0
chr4A	1|6	1|5	1|1	1|12		1|0	2|3		1|0						8|27
chr4B	2|1	1|1		1|1			1|0		1|0						6|3
chr4D					1|0				1|0				0|1	1|0	3|1
chr5A	1|2	2|3	0|2	5|1			0|1		1|0			3|3	1|0	0|1	13|13	3|0
chr5B	2|0	4|1	2|0	0|2	1|0	1|0	1|1						1|0		12|4	0|1
chr5D	1|2	1|1	0|1												2|4	0|1
chr6A	1|0		0|4	2|0					2|0						5|4	2|0
chr6B	3|2	5|2	6|1	2|0				1|0				1|0	1|0		19|5	4|5
chr6D	1|0	1|0													2|0
chr7A	0|2	0|2	3|1		1|0		0|2				0|1	1|1	1|0		6|9
chr7B	4|3	3|2	5|0	3|3			0|1					0|1	1|0		16|10	3|2
chr7D	1|1	0|1	0|1		1|3	0|1				0|1		0|1			2|9
chrUn*	2|0	2|0				0|1	0|1						1|1		5|3	0|1
All	38|30	37|29	32|18	26|21	6|7	3|3	6|10	1|0	8|1	0|2	0|1	13|13	11|4	2|1	183|140
within gene model	5|5	4|3	6|1	2|1			0|1		3|1			1|0	1|0			22|12

HT: heading time; FT: flowering time; MT: maturation time; CL: culm length; TN: number of tillers; SL: spike length; SN: number of spikelets per spike; HGW: 100-grain weight; SD: spike density; WC, WO and WS: wax-like deposits on culm, ochrea and spike, respectively; AC and AS: anthocyanin pigmentation on coleoptile and seedling, respectively; GH: glume hair; AWN: awnedness.

Number of (SNPs) | (PAVs) detected with significant MTAs (a significance level of 1% after Bonferroni multiple test correction).

*  A pseudomolecule composed of unanchored scaffolds (IWGSC RefSeq ver. 1.0).

A total of 47 MTAs with CL were found on all the chromosomes expect 3D, 4D, 5D, 6D, 7A and 7D (Fig. 3B and Table 6). Three markers (two SNPs and one PAV) were within gene model sequences (Supplementary Table S1). Notably, three markers significantly associated with CL were also associated with HT and FT, and two with HT (Supplementary Table S1). We detected MTAs with SN with sixteen markers (six SNPs and 10 PAVs) on chromosomes 2D, 3B, 4A, 4B, 5A, 5B, 7A and 7B (Supplementary Fig. S3E and Table 6). Four markers (one SNP and three PAVs) were commonly associated with SN as well as HT, FT and MT (Supplementary Table S1). Moreover, one PAV marker associated with SN on chromosome 4A was also detected in the MTA analysis of HT and FT (Supplementary Table S1).

The results of other MTAs for the traits TN, SL, SD, AWN, AC, AS, WO, WC, WS and GH are depicted in Supplementary Figs. S3 and S5, and summarized in Table 6. We found significant MTAs for the traits TN, SL, SD, AWN, AC, AS, WO and GH, but not for WC or WS. The number of significant MTAs for these traits varied from zero (WC and WS) to 26 (AS, Supplementary Table S1).

DISCUSSION

Diversity of the NBRP-Wheat hexaploid wheat core collection as revealed by DArT markers

The 187 accessions collected from various geographical regions were genotyped by the DArTseq marker system (34,234 SNPs and 60,077 PAVs). For analyses, we selected 13,394 SNPs and 20,523 PAVs as the data sets that could be anchored to the IWGSC RefSeq ver. 1.0 by BLAST searches. The number of markers was highest in the B genome (5,820 SNPs and 7,484 PAVs) and lowest in the D genome (1,768 SNPs and 5,899 PAVs) (Table 2). These differences in the number of markers among genomes are coincident with those observed in previous studies (Akbari et al., 2006; Semagn et al., 2006; Zhang et al., 2011; Nielsen et al., 2014). The mean PIC value was also lowest in the D genome (SNPs: 0.230; PAVs: 0.159) (Table 2). These results show that the diversity is lower in the D genome than in the A and B genomes. The fewer polymorphisms present in the D genome may reflect the genetic bottleneck that occurred through hexaploidization and that brought only a limited amount of genetic diversity to Ae. tauschii (Caldwell et al., 2004; Chao et al., 2009; Mizuno et al., 2010).

Population structure and genetic relationships among common wheat

Bayesian clustering analysis divided common wheat accessions into six clusters (Pop-A, -B, -C, -D, -E and -F) (Table 1) and most of them were admixtures (Fig. 1). However, there are some non-admixture accessions that help to infer the characteristics of each cluster (Fig. 1). Accessions classified as Pop-A belong to Eastern Asian lines present at the eastern limit of wheat distribution (mostly Japanese cultivars; Table 1). Landraces derived from the Himalayan region to Southern Asia are categorized into Pop-B (Table 1). Accessions included in Pop-C are mostly Western Asian lines (Afghanistan and Iran; Table 1). Landraces derived from the Middle East are categorized into Pop-D and Pop-E (Iran, Turkey, Lebanon and Syria) (Table 1). Most European landraces are classified as Pop-F (Table 1). Overall, the common wheat landraces of the NBRP-Wheat hexaploid wheat core collection are genetically divergent and form at least six groups each consisting of accessions from close geographical regions.

The accessions analyzed in this study form a triangular shape on the PC1-PC2 plane (Fig. 2). Accessions classified as Pop-A and Pop-C form two vertices of the triangle, and accessions of Pop-B are plotted on the side of the triangle between Pop-A and Pop-C (Fig. 2). Accessions classified as Pop-F form the third vertex of the triangle, and accessions of Pop-D and Pop-E were plotted on the side of the triangle between Pop-C and Pop-F (Fig. 2). The configurations indicate that the Himalayan region group (Pop-B) is situated between Eastern Asian (Pop-A) and Western Asian (Pop-C) groups, which may represent the dispersal pathway of hexaploid wheat through the Silk Road. Middle Eastern groups (Pop-D and Pop-E) are situated between Western Asian and European groups (Pop-F, Fig. 2), likely indicating that the latter two groups originated from the Middle Eastern groups. The finding of accessions that connect the vertices of the triangle (Fig. 2A) suggests that the differentiation of the six groups among common wheat landraces was genetically and geographically not completely punctuated.

Most modern cultivars are categorized into Pop-A, Pop-E and Pop-F (Table 1). Non-admixture accessions in Pop-A are Japanese cultivars and most are soft wheat cultivars used for Japanese noodles. Non-admixture accessions in Pop-F are early British cultivars developed before the 1920s. These results suggest that they developed from regional landraces. Many modern cultivars categorized into Pop-E are admixtures (Fig. 1) and most are cultivars developed after the 1930s.

Genetic relationships and origin of exotic hexaploid wheat species

The NBRP-Wheat hexaploid wheat core collection contains five exotic non-aestivum hexaploid wheat species (T. spelta, T. compactum, T. sphaerococcum, T. macha and T. vavilovii). Population structure analysis showed that these accessions were divided into seven clusters (Fig. 1). Six of them were composed mainly of common wheat accessions and only one cluster (Pop-G) contained T. spelta, T. macha and synthetic hexaploid wheat (Fig. 1). Accessions categorized as Pop-G were sharply distinguished from other accessions by PCA (Fig. 2). These results indicate that accessions included in Pop-G are genetically different to common wheat accessions. Spelt wheat (T. spelta) accessions used in this study were derived from Germany, Spain and Iran (Table 1). All European spelt wheat accessions were categorized as Pop-G, whereas Iranian spelt wheat was categorized as Pop-C (Fig. 1), in agreement with a previous study suggesting that European and Asian spelt wheats originated independently (Dvorak et al., 2012). Macha wheat (T. macha) is endemic to the Caucasus (Cao et al., 1998; Zohary and Hopf, 2000; Matsuoka, 2011). Previous work has suggested that this hexaploid wheat originated from a hybrid cross between wild emmer wheat (T. turgidum ssp. dicoccoides) and T. aestivum (Matsuoka, 2011). All macha wheat accessions are admixtures with genetic exchange from the Pop-D accessions including Caucasian common wheat; in contrast, all European spelt wheats are non-admixture accessions (Table 1 and Fig. 1). This result suggests that T. aestivum lines genetically close to Pop-D accessions contributed to the origin of macha wheat. Comparative genetic studies indicated that European spelt wheat originated from the introgression of cultivated emmer wheat (tetraploid, 2n = 4× = 28) into free-threshing hexaploid wheat (Matsuoka, 2011). Furthermore, the synthetic hexaploid wheat accession, W7984, was classified into Pop-G and its genomic composition was very similar to macha wheat accessions (Fig. 1). Moreover, PCA showed that W7984 clustered with macha wheat accessions (Fig. 2). These results indicate that the accessions in Pop-G have emmer wheat genetic influences which distinguish this group from others.

Triticum sphaerococcum is endemic to southern Pakistan and northwestern India (Percival, 1921; Ellerton, 1939; Mori et al., 2013). Together with Pakistani common wheat, all T. sphaerococcum accessions were classified as Pop-B according to population structure analysis (Fig. 1). This result suggests a single origin for T. sphaerococcum in Southern Asia.

Triticum vavilovii is endemic to the southern Caucasus. All T. vavilovii accessions were classified into Pop-C, together with T. aestivum accessions. They are admixture accessions, being genetically related to Pop-D; in addition, their genomic structure was similar to common wheat accessions from Afghanistan (Fig. 1). This result suggests that T. vavilovii is genetically close to common wheat derived from Western Asia: whereas common wheat is free-threshing, T. vavilovii is non-free-threshing. The Iranian T. spelta accession in Pop-C is non-free-threshing (Fig. 1), which may indicate that Asian spelt wheat contributed to the origin of T. vavilovii.

Club wheat (T. compactum) differs from other hexaploid varieties principally due to its compact spike, which results from the activity of the dominant compactum (C) allele on chromosome 2D (Rao, 1972; Johnson et al., 2008). Six club wheat accessions derived from various regions (Japan, China, Afghanistan, Iran, Turkey and Ethiopia) were selected from the NBRP-Wheat hexaploid wheat core collection. Population structure analyses classified five of the accessions, with the exception of the Ethiopian club wheat accession, into four clusters (Pop-A, -C, -D and -F, Fig. 1). This result suggests that club wheat accessions lacked genetic homogeneity as a taxonomic group, in contrast to other exotic hexaploid wheat. Kosuge et al. (2012) showed that genes other than C affect compact spike size in hexaploid wheat; therefore, we propose that the club wheat accessions originated independently in various regions through mutations in different loci.

Association analyses

We performed association analyses for nine quantitative traits and seven qualitative traits and found a total of 322 significant marker-trait associations (MTAs, Table 6). Here, we take MTAs for heading time (HT), culm length (CL), awnedness (AWN) and anthocyanin pigmentation on coleoptile (AC) as examples to compare our results to previous studies and test the potential usefulness of the NBRP-Wheat hexaploid wheat core collection for association mapping studies.

Eight QTLs for ear emergence are located on chromosomes 2B, 2D, 4A, 5A, 5D and 7D (for a summary of the wheat gene catalogue, see Gene Catalogue 2013, https://shigen.nig.ac.jp/wheat/komugi/genes/download.jsp). Meta-QTL analysis in elite European winter wheat germplasms found QTLs for ear emergence on chromosomes 1B, 1D, 2A, 3A, 3B, 4B, 4D, 5A, 6A, 6B, 7A, 7B and 7D (Griffiths et al., 2009). Neumann et al. (2011) detected MTAs for HT on chromosomes 1B, 1D, 2D, 4A, 4B, 5D, 6A, 6B and 7A by genome-wide association analysis of 96 winter wheats with 874 DArT markers. Edae et al. (2014) reported MTAs for HT on chromosomes 1D, 2A, 2B, 3A, 3B, 4B and 7D by genome-wide association analyses in 285-294 spring wheats grown in five different environments with 1,863 DArT markers. Genome-wide association analysis of a core collection of 227 bread wheats with 760 DArT markers revealed 62 markers corresponding to 33 chromosomal regions individually associated with earliness components (Le Gouis et al., 2012). These studies suggest that the HT trait is strongly related to the photoperiod response gene (Ppd-1) on the short arms of homoeologous group-2 chromosomes and the vernalization response gene (Vrn-1) on the long arms of homoeologous group-5 chromosomes. We found 68 MTAs (30 SNPs and 28 PAVs) for HT on 20 chromosomes, except for 4D (Fig. 3A and Table 6). The significant markers were located on the short arms of chromosomes 2B and 2D and on the long arms of chromosomes 5A and 5D (Supplementary Table S1), which may indicate the detection of the Ppd-1 and Vrn-1 homoeoloci in the association study. The significant markers on chromosome 2A were all located on the long arms (Supplementary Table S1), indicating that HT was not associated with the Ppd-1A locus on chromosome arm 2AL in our conditions. Association of HT with markers on 2AL was also reported by Edae et al. (2014).

Wheat plant height is controlled by genes located on at least 17 of the 21 chromosomes (Snape et al., 1977). Previously reported QTLs and genes for plant height (culm length; CL in this study) are summarized in Gene Catalogue 2013. Neumann et al. (2011) detected MTAs for plant height on chromosomes 1A, 1B, 2B, 4A, 6A, 6B, 7A and 7B; meanwhile, Edae et al. (2014) reported MTAs for plant height on chromosomes 3B, 5B, 6A, 7A and 7B. In this study, we found 47 MTAs (26 SNPs and 21 PAVs) for CL on chromosomes 1A, 1B, 1D, 2A, 2B, 2D, 3A, 3B, 4A, 4B, 5A, 5B, 6A, 6B and 7B (Fig. 3B and Table 6). The semi-dwarfing genes Rht8, Rht-B1 and Rht-D1 on chromosomes 2D, 4B and 4D, respectively, which have been derived from Japanese common wheat, are widely used in modern cultivars (Gale and Marshall, 1976; Börner et al., 1996; Korzun et al., 1998; Worland et al., 1998; Peng et al., 1999; Singh et al., 2001; Pearce et al., 2011). Accordingly, both Japanese wheat accessions and modern cultivars selected from the NBRP-Wheat hexaploid wheat core collection are expected to carry these semi-dwarfing genes. However, we did not detect MTAs on chromosome 4D. We suspect that allelic variation at the Rht-D1 locus was too low to detect MTAs due to the predominance of the semi-dwarf accessions, as was also indicated by Neumann et al. (2011).

Awning of wheat is controlled by three genes, Hd, B1 and B2 (Gene Catalogue 2013), which are located on the short arm of chromosome 4A (Sears, 1954; Rao, 1981; Sourdille et al., 2002), the long arm of chromosome 5A (Sears, 1954; Sourdille et al., 2002) and the long arm of chromosome 6B (Sears, 1954; Sourdille et al., 2002), respectively. Wheat varieties homozygous recessive for hd, b1 and b2 alleles are fully awned, while those homozygous dominant for either Hd B2 or B1 B2 are awnless; those with Hd B2 are expected to be awnless (Sourdille et al., 2002). We detected significant MTAs on chromosomes 2A, 4D and 5A (Fig. 3C and Table 6), and MTAs on chromosome 5A likely correspond to B1. However, we could not find MTAs corresponding to Hd or B2. QTL analysis for three traits related to awning (awn length at the base, middle and top of spike) in a doubled-haploid population derived from the cross between T. aestivum ‘Courtot’ (awned) and T. aestivum ‘Chinese Spring’ (awnless) revealed two QTLs on chromosomes 4A and 6B that cosegregated with Hd and B2, respectively. The QTL on chromosome 4A (Hd) had a stronger effect on the awns at both ends of the spike and that on chromosome 6B (B2) on the awns in the middle of the spike (Sourdille et al., 2002). Recently, Yoshioka et al. (2017) reported finding a variety of awned and awnless spikes in a segregating population between T. aestivum ‘Mironovskaya 808’ (awnless) and T. aestivum ‘Chinese Spring’ (awnless). Additionally, they found that all phenotypic variation could be explained by the three loci Hd, B1 and B2 by a combination of close observation and genotype prediction. In this study, however, we rated awnedness as a qualitative trait with a three-point scale, which would have resulted in poor phenotypic resolution in detecting significant MTAs in other predicted genomic regions.

Anthocyanin pigments in wheat are found in the anthers (gene symbol Pan), auricle (Ra), leaf blade (Plb), leaf sheath (Pls), coleoptile (Rc), culm (Pc) or pericarp (Pp) (Gene Catalogue 2013). The Pan, Plb, Pls, Rc and Pc genes were closely linked on the short arms of homoeologous group-7 (Khlestkina et al., 2002, 2009). Ahmed et al. (2006) noted that the Rc gene likely encodes a transcriptional activator of late biosynthesis genes involved in light-regulated anthocyanin synthesis. We detected significant MTAs with AC on the short arm of chromosome 7A, as expected for the Rc gene (Fig. 3D and Table 6).

The measurements for each trait in this study were obtained from a single cultivation season at one location. Therefore, we are not taking account of phenotypic variance due to environmental effects. However, as discussed above, some MTAs detected in this study are consistent with genes, QTLs and MTAs that have been reported previously. The genotype information acquired from the NBRP-Wheat hexaploid wheat core collection in this study provides suitable research tools to identify novel QTLs by marker-trait association analyses. The collection has previously been used in forward genetics studies (Yokota et al., 2016; Yoshioka et al., in press). The public release of the core collection with intensive genotype information should promote gene discovery in hexaploid wheat.

ACKNOWLEDGMENTS

We thank the staff of NBRP-Wheat for their excellent technical support. This work was supported by the National BioResource Project-Wheat from the Ministry of Education, Culture, Sports, Science and Technology, Japan. The accessions in the core collection are available from NBRP-Wheat. The genotype data including SNPs and PAVs are also to be available from the NBRP-Wheat Database KOMUGI (https://shigen.nig.ac.jp/wheat/komugi/). This manuscript is given the contribution number 620 by the Laboratory of Plant Genetics, Graduate School of Agriculture, Kyoto University.

REFERENCES

•  Ahmed,  N.,  Maekawa,  M.,  Utsugi,  S.,  Rikiishi,  K.,  Ahmad,  A., and  Noda,  K. (2006) The wheat Rc gene for red coleoptile colour codes for a transcriptional activator of late anthocyanin biosynthesis genes. J. Cereal Sci. 44, 54–58.
•  Akbari,  M.,  Wenzl,  P.,  Caig,  V.,  Carling,  J.,  Xia,  L.,  Yang,  S.,  Uszynski,  G.,  Mohler,  V.,  Lehmensiek,  A.,  Kuchel,  H., et al. (2006) Diversity arrays technology (DArT) for high-throughput profiling of the hexaploid wheat genome. Theor. Appl. Genet. 113, 1409–1420.
•  Alexander,  D. H.,  Novembre,  J., and  Lange,  K. (2009) Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 19, 1655–1664.
•  Börner,  A.,  Plaschke,  J.,  Korzun,  V., and  Worland,  A. J. (1996) The relationships between the dwarfing genes of wheat and rye. Euphytica 89, 69–75.
•  Bradbury,  P. J.,  Zhang,  Z.,  Kroon,  D. E.,  Casstevens,  T. M.,  Ramdoss,  Y., and  Buckler,  E. S. (2007) TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics 23, 2633–2635.
•  Breseghello,  F., and  Sorrells,  M. E. (2006) Association mapping of kernel size and milling quality in wheat (Triticum aestivum L.) cultivars. Genetics 172, 1165–1177.
•  Buckler IV,  E. S., and  Thornsberry,  J. M. (2002) Plant molecular diversity and applications to genomics. Curr. Opin. Plant Biol. 5, 107–111.
•  Caldwell,  K. S.,  Dvorak,  J.,  Lagudah,  E. S.,  Akhunov,  E.,  Luo,  M.,  Wolters,  P., and  Powell,  W. (2004) Sequence polymorphism in polyploid wheat and their D-genome diploid ancestor. Genetics 167, 941–947.
•  Cao,  W.,  Hucl,  P.,  Scoles,  G., and  Chibbar,  R. N. (1998) Genetic diversity within spelta and macha wheats based on RAPD analysis. Euphytica 104, 181–189.
•  Chao,  S.,  Zhang,  W.,  Akhunov,  E.,  Sherman,  J.,  Ma,  Y.,  Luo,  M. C., and  Dubcovsky,  J. (2009) Analysis of gene-derived SNP marker polymorphism in US wheat (Triticum aestivum L.) cultivars. Mol. Breed. 23, 23–33.
•  Crossa,  J.,  Burgueño,  J.,  Dreisigacker,  S.,  Vargas,  V.,  Herrera-Foessel,  S. A.,  Lillemo,  M.,  Singh,  R. P.,  Trethowan,  R.,  Warburton,  M.,  Franco,  J., et al. (2007) Association analysis of historical bread wheat germplasm using additive genetic covariance of relatives and population structure. Genetics 177, 1889–1913.
•  Dvorak,  J.,  Deal,  K. R.,  Luo,  M. C.,  You,  F. M.,  von Borstel,  K., and  Dehghani,  H. (2012) The origin of spelt and free-threshing hexaploid wheat. J. Hered. 103, 426–441.
•  Edae,  E. A.,  Byrne,  P. F.,  Haley,  S. D.,  Lopes,  M. S., and  Reynolds,  M. P. (2014) Genome-wide association mapping of yield and yield components of spring wheat under contrasting moisture regimes. Theor. Appl. Genet. 127, 791–807.
•  Ellerton,  S. (1939) The origin and geographical distribution of Triticum sphaerococcum Perc. and its cytogenetical behaviour in crosses with T. vulgare VILL. J. Genet. 38, 307–324.
•  Gale,  M. D., and  Marshall,  G. A. (1976) The chromosomal location of Gai1 and Rht1, genes for gibberellin insensitivity and semi-dwarfism, in a derivative of Norin 10 wheat. Heredity 37, 283–289.
•  Giles,  R. J., and  Brown,  T. A. (2006) GluDy allele variations in Aegilops tauschii and Triticum aestivum: implications for the origins of hexaploid wheats. Theor. Appl. Genet. 112, 1563–1572.
•  Griffiths,  S.,  Simmonds,  J.,  Leverington,  M.,  Wang,  Y.,  Fish,  L.,  Sayers,  L.,  Alibert,  L.,  Orford,  S.,  Wingen,  L.,  Herry,  L., et al. (2009) Meta-QTL analysis of the genetic control of ear emergence in elite European winter wheat germplasm. Theor. Appl. Genet. 119, 383–395.
•  Harlan,  J. R. (1975) Our vanishing genetic resources. Science 188, 617–621.
• International Wheat Genome Sequencing Consortium (IWGSC) (2018) Shifting the limits in wheat research and breeding using a fully annotated reference genome. Science 361, eaar7191.
•  Jaccoud,  D.,  Peng,  K.,  Feinstein,  D., and  Kilian,  A. (2001) Diversity arrays: a solid state technology for sequence information independent genotyping. Nucleic Acids Res. 29, e25.
•  Johnson,  E. B.,  Nalam,  V. J.,  Zemetra,  R. S., and  Riera-Lizarazu,  O. (2008) Mapping the compactum locus in wheat (Triticum aestivum L.) and its relationship to other spike morphology genes of the Triticeae. Euphytica 163, 193–201.
•  Jolliffe,  I. T. (2002) Principal Component Analysis, Second Edition. Springer, New York, USA.
•  Khlestkina,  E. K.,  Pestsova,  E. G.,  Röder,  M. S., and  Börner,  A. (2002) Molecular mapping, phenotypic expression and geographical distribution of genes determining anthocyanin pigmentation of coleoptiles in wheat (Triticum aestivum L.). Theor. Appl. Genet. 104, 632–637.
•  Khlestkina,  E.,  Pshenichnikova,  T.,  Röder,  M., and  Börner,  A. (2009) Clustering anthocyanin pigmentation genes in wheat group 7 chromosomes. Cereal Res. Commun. 37, 391–398.
•  Kihara,  H. (1944) Discovery of the DD-analyser, one of the ancestors of Triticum vulgare. Agric. Hortic. 19, 889–890.
•  Korzun,  V.,  Röder,  M. S.,  Ganal,  M. W.,  Worland,  A. J., and  Law,  C. N. (1998) Genetic analysis of the dwarfing gene (Rht8) in wheat. Part I. Molecular mapping of Rht8 on the short arm of chromosome 2D of bread wheat (Triticum aestivum L.). Theor. Appl. Genet. 96, 1104–1109.
•  Kosuge,  K.,  Watanabe,  N.,  Melnik,  V. M.,  Laikova,  L. I., and  Goncharov,  N. P. (2012) New sources of compact spike morphology determined by the genes on chromosome 5A in hexaploid wheat. Genet. Resour. Crop Evol. 59, 1115–1124.
•  Le Gouis,  J.,  Bordes,  J.,  Ravel,  C.,  Heumez,  E.,  Faure,  S.,  Praud,  S.,  Galic,  N.,  Remoué,  C.,  Balfourier,  F.,  Allard,  V., et al. (2012) Genome-wide association analysis to identify chromosomal regions determining components of earliness in wheat. Theor. Appl. Genet. 124, 597–611.
•  Li,  G.,  Xu,  X.,  Bai,  G.,  Carver,  B. F.,  Hunger,  B.,  Bonman,  J. M.,  Kolmer,  J., and  Dong,  H. (2016) Genome-wide association mapping reveals novel QTL for seedling leaf rust resistance in a worldwide collection of winter wheat. Plant Genome 9, plantgenome2016.06.0051.
•  Matsuoka,  Y. (2011) Evolution of polyploid Triticum wheats under cultivation: the role of domestication, natural hybridization and allopolyploid speciation in their diversification. Plant Cell Physiol. 52, 750–764.
•  McCouch,  S.,  Baute,  G. J.,  Bradeen,  J.,  Bramel,  P.,  Bretting,  P. K.,  Buckler,  E.,  Burke,  J. M.,  Charest,  D.,  Cloutier,  S.,  Cole,  G., et al. (2013) Agriculture: Feeding the future. Nature 499, 23–24.
•  McFadden,  E. S., and  Sears,  E. R. (1944) The artificial synthesis of Triticum spelta. Rec. Genet. Soc. Am. 13, 26–27.
•  Mizuno,  N.,  Yamasaki,  M.,  Matsuoka,  Y.,  Kawahara,  T., and  Takumi,  S. (2010) Population structure of wild wheat D-genome progenitor Aegilops tauschii Coss.: implications for intraspecific lineage diversification and evolution of common wheat. Mol. Ecol. 19, 999–1013.
•  Mori,  N.,  Ohta,  S.,  Chiba,  H.,  Takagi,  T.,  Niimi,  Y.,  Shinde,  V.,  Kajale,  M. D., and  Osada,  T. (2013) Rediscovery of Indian dwarf wheat (Triticum aestivum L. ssp. sphaerococcum (Perc.) MK.) an ancient crop of the Indian subcontinent. Genet. Resour. Crop Evol. 60, 1771–1775.
•  Nesbitt,  M., and  Samuel,  D. (1996) From staple crop to extinction? The archaeology and history of the hulled wheats. In Hulled wheats, Proceedings of the First International Workshop on Hulled Wheats, 21. (eds.: Padulosi, S., Hammer, K., and Heller, J.), pp. 41–100, International Plant Genetic Resources Institute, Rome, Italy.
•  Neumann,  K.,  Kobiljski,  B.,  Denčić,  S.,  Varshney,  R. K., and  Börner,  A. (2011) Genome-wide association mapping: a case study in bread wheat (Triticum aestivum L.). Mol. Breed. 27, 37–58.
•  Nielsen,  N. H.,  Backes,  G.,  Stougaard,  J.,  Andersen,  S. U., and  Jahoor,  A. (2014) Genetic diversity and population structure analysis of European hexaploid bread wheat (Triticum aestivum L.) varieties. PLoS One 9, e94000.
•  Pearce,  S.,  Saville,  R.,  Vaughan,  S. P.,  Chandler,  P. M.,  Wilhelm,  E. P.,  Sparks,  C. A.,  Al-Kaff,  N.,  Korolev,  A.,  Boulton,  M. I.,  Phillips,  A. L., et al. (2011) Molecular characterization of Rht-1 dwarfing genes in hexaploid wheat. Plant Physiol. 157, 1820–1831.
•  Peng,  J.,  Richards,  D. E.,  Hartley,  N. M.,  Murphy,  G. P.,  Devos,  K. M.,  Flintham,  J. E.,  Beales,  J.,  Fish,  L. J.,  Worland,  A. J.,  Pelica,  F., et al. (1999) ‘Green revolution’ genes encode mutant gibberellin response modulators. Nature 400, 256–261.
•  Percival,  J. (1921) The Wheat Plant: A Monograph. Duckworth and Co, London, UK.
•  Rao,  M. V. P. (1972) Mapping of the compactum gene C on chromosome 2D of wheat. Wheat Inf. Serv. 35, 9.
•  Rao,  M. V. P. (1981) Telocentric mapping of the awn inhibitor gene Hd on chromosome 4B of common wheat. Cereal Res. Commun. 9, 335–337.
•  Reich,  D.,  Price,  A. L., and  Patterson,  N. (2008) Principal component analysis of genetic data. Nat. Genet. 40, 491–492.
•  Remington,  D. L.,  Thornsberry,  J. M.,  Matsuoka,  Y.,  Wilson,  L. M.,  Whitt,  S. R.,  Doebley,  J.,  Kresovich,  S.,  Goodman,  M. M.,  Edward,  S., and  Buckler,  E. S. IV. (2001) Structure of linkage disequilibrium and phenotypic associations in the maize genome. Proc. Natl. Acad. Sci. USA 98, 11479–11484.
•  Roldán-Ruiz,  I.,  Dendauw,  J.,  Van Bockstaele,  E.,  Depicker,  A., and  De Loose,  M. (2000) AFLP markers reveal high polymorphic rates in ryegrasses (Lolium spp.). Mol. Breed. 6, 125–134.
•  Salamini,  F.,  Özkan,  H.,  Brandolini,  A.,  Schäfer-Pregl,  R., and  Martin,  W. (2002) Genetics and geography of wild cereal domestication in the Near East. Nat. Rev. Genet. 3, 429–441.
•  Sears,  E. R. (1954) The aneuploids of common wheat. Mo. Agr. Expt. Stat. Res. Bull. 572, 1–58.
•  Semagn,  K.,  Bjørnstad,  Å.,  Skinnes,  H.,  Marøy,  A. G.,  Tarkegne,  Y., and  William,  M. (2006) Distribution of DArT, AFLP, and SSR markers in a genetic linkage map of a doubled-haploid hexaploid wheat population. Genome 49, 545–555.
•  Singh,  R. P.,  Huerta-Espino,  J.,  Rajaram,  S., and  Crossa,  J. (2001) Grain yield and other traits of tall and dwarf isolines of modern bread and durum wheats. Euphytica 119, 241–244.
•  Snape,  J. W.,  Law,  C. N., and  Worland,  A. J. (1977) Whole chromosome analysis of height in wheat. Heredity 38, 25–36.
•  Sourdille,  P.,  Cadalen,  T.,  Gay,  G.,  Gill,  B., and  Bernard,  M. (2002) Molecular and physical mapping of genes affecting awning in wheat. Plant Breed. 121, 320–324.
•  Wang,  G.,  Leonard,  J. M.,  von Zitzewitz,  J.,  Peterson,  C. J.,  Ross,  A. S., and  Riera-Lizarazu,  O. (2014) Marker–trait association analysis of kernel hardness and related agronomic traits in a core collection of wheat lines. Mol. Breed. 34, 177–184.
•  Worland,  A. J.,  Korzun,  V.,  Röder,  M. S.,  Ganal,  M. W., and  Law,  C. N. (1998) Genetic analysis of the dwarfing gene Rht8 in wheat. Part II. The distribution and adaptive significance of allelic variants at the Rht8 locus of wheat as revealed by microsatellite screening. Theor. Appl. Genet. 96, 1110–1120.
•  Yokota,  H.,  Iehisa,  J. C. M.,  Nishijima,  R.,  Nitta,  M.,  Takenaka,  S.,  Nasuda,  S., and  Takumi,  S. (2016) Variation in abscisic acid responsiveness at the early seedling stage is related to line differences in seed dormancy and in expression of genes involved in abscisic acid responses in common wheat. J. Cereal Sci. 71, 167–176.
•  Yoshioka,  M.,  Iehisa,  J. C. M.,  Ohno,  R.,  Kimura,  T.,  Enoki,  H.,  Nishimura,  S.,  Nasuda,  S., and  Takumi,  S. (2017) Three dominant awnless genes in common wheat: fine mapping, interaction and contribution to diversity in awn shape and length. PLoS One 12, e0176148.
•  Yoshioka,  M.,  Takenaka,  S.,  Nitta,  M.,  Li,  J.,  Mizuno,  N., and  Nasuda,  S. Genetic dissection of grain morphology in hexaploid wheat by analysis of the NBRP-Wheat core collection. Genes Genet. Syst. (in press)
•  Zanke,  C. D.,  Ling,  J.,  Plieske,  J.,  Kollers,  S.,  Ebmeyer,  E.,  Korzun,  V.,  Argillier,  O.,  Stiewe,  G.,  Hinze,  M.,  Neumann,  K., et al. (2014) Whole genome association mapping of plant height in winter wheat (Triticum aestivum L.). PLoS One 9, e113287.
•  Zanke,  C. D.,  Ling,  J.,  Plieske,  J.,  Kollers,  S.,  Ebmeyer,  E.,  Korzun,  V.,  Argillier,  O.,  Stiewe,  G.,  Hinze,  M.,  Neumann,  F., et al. (2015) Analysis of main effect QTL for thousand grain weight in European winter wheat (Triticum aestivum L.) by genome-wide association mapping. Front. Plant Sci. 6, 644.
•  Zanke,  C. D.,  Rodemann,  B.,  Ling,  J.,  Muqaddasi,  Q. H.,  Plieske,  J.,  Polley,  A.,  Kollers,  S.,  Ebmeyer,  E.,  Korzun,  V.,  Argillier,  O., et al. (2017) Genome-wide association mapping of resistance to eyespot disease (Pseudocercosporella herpotrichoides) in European winter wheat (Triticum aestivum L.) and fine-mapping of Pch1. Theor. Appl. Genet. 130, 505–514.
•  Zhang,  L.,  Liu,  D.,  Guo,  X.,  Yang,  W.,  Sun,  J.,  Wang,  D.,  Sourdille,  P., and  Zhang,  A. (2011) Investigation of genetic diversity and population structure of common wheat cultivars in northern China using DArT markers. BMC Genet. 12, 42.
•  Zhang,  Z.,  Schwartz,  S.,  Wagner,  L., and  Miller,  W. (2000) A greedy algorithm for aligning DNA sequences. J. Comput. Bio. 7, 203–214.
•  Zohary,  D., and  Hopf,  M. (2000) Domestication of Plants in the Old World: The Origin and Spread of Cultivated Plants in West Asia, Europe, and the Nile Valley. Oxford University Press, Oxford, UK.
•  Zondervan,  K. T., and  Cardon,  L. R. (2004) The complex interplay among factors that influence allelic association. Nat. Rev. Genet. 5, 89–100.