当前位置: 首页 > >

Association mapping for pre-harvest sprouting resistance in white winter wheat

发布时间:

Theor Appl Genet (2012) 125:793–805 DOI 10.1007/s00122-012-1872-0

ORIGINAL PAPER

Association mapping for pre-harvest sprouting resistance in white winter wheat
Pawan Kulwal ? Goro Ishikawa ? David Benscher Zongyun Feng ? Long-Xi Yu ? Ashok Jadhav ? Subhash Mehetre ? Mark E. Sorrells
?

Received: 27 November 2011 / Accepted: 4 April 2012 / Published online: 1 May 2012 ? Springer-Verlag 2012

Abstract Association mapping identi?ed quantitative trait loci (QTLs) and the markers linked to pre-harvest sprouting (PHS) resistance in an elite association mapping panel of white winter wheat comprising 198 genotypes. A total of 1,166 marker loci including DArT and SSR markers representing all 21 chromosomes of wheat were used in the analysis. General and mixed linear models were used to analyze PHS data collected over 4 years. Association analysis identi?ed eight QTLs linked with 13 markers mapped on seven chromosomes. A QTL was detected on each arm of chromosome 2B and one each on chromosome arms 1BS, 2DS, 4AL, 6DL, 7BS and 7DS. All except the QTL on 7BS are located in a location similar to previous

reports and, if veri?ed, the QTL on 7BS is likely to be novel. Principal components and the kinship matrix were used to account for the presence of population structure but had only a minor effect on the results. Although, none of the QTLs was highly signi?cant across all environments, a QTL on the long arm of chromosome 4A was detected in three different environments and also using the best linear unbiased predictions over years. Although previous reports have identi?ed this as a major QTL, its effects were minor in our biparental mapping populations. The results of this study highlight the bene?ts of association mapping and the value of using elite material in association mapping for plant breeding programs.

Introduction
Communicated by P. Hayes.

Electronic supplementary material The online version of this article (doi:10.1007/s00122-012-1872-0) contains supplementary material, which is available to authorized users.
P. Kulwal ? A. Jadhav ? S. Mehetre State Level Biotechnology Centre, Mahatma Phule Agricultural University, Rahuri 413 722, MS, India G. Ishikawa Tohoku Agricultural Research Center, National Agriculture and Food Research Organization, 4 Akahira, Shimokuriyagawa, Morioka, Iwate 020-0198, Japan D. Benscher ? Z. Feng ? L.-X. Yu ? M. E. Sorrells (&) Department of Plant Breeding and Genetics, Cornell University, 240 Emerson Hall, Ithaca, NY 14853, USA e-mail: mes12@cornell.edu Z. Feng Department of Plant Genetics and Breeding, College of Agronomy, Sichuan Agricultural University, 211 Huimin Rd., Wenjiang District, Chengdu 611130, Sichuan, China

Pre-harvest sprouting (PHS) is characterized by the premature germination of the seed while still attached to the mother plant and is an important quality constraint in hexaploid wheat (Triticum aestivum L.). PHS in wheat results from early break of seed dormancy under humid and wet weather prior to harvest. Pre-harvest sprouting has been a major problem in different parts of the world resulting in the loss of grain weight and poor end use quality. Wild relatives of wheat have extended seed dormancy that has been reduced through domestication so that harvested seed will germinate after a short after-ripening period. There are several factors and mechanisms that affect PHS in wheat including kernel color, phytohormones such as ABA and GA and after ripening (AR). These and other factors have been adequately studied and reviewed in several articles (Gerjets et al. 2010; Kulwal et al. 2010). There have been several studies using biparental mapping

123

794

Theor Appl Genet (2012) 125:793–805

populations in the past to understand the genetic basis of PHS through QTL mapping and almost all the 21 chromosomes of common wheat have been reported to contain QTLs for PHS (reviewed in Kulwal et al. 2010). Of these 21, chromosomes 2B and 2D as well as chromosomes of groups 3 (3A, 3B and 3D), and 4A are of particular importance as they have been reported to contain major QTLs for PHS in both white and red wheat. Chromosomes 2B and 2D are important because major and consistent QTLs have been identi?ed for PHS in several studies (Kulwal et al. 2004; Mohan et al. 2009; Munkvold et al. 2009); whereas group 3 chromosomes are important because, besides major QTLs (Kulwal et al. 2005; Mori et al. 2005; Imtiaz et al. 2008; Nakamura et al. 2011), they also contain the red kernel color gene ‘R’ (Sears 1944) and the viviparous gene TaVp1 (Bailey et al. 1999; Nakamura and Toyoma 2001). Chromosome 4A has been reported to contain a ‘Phs’ gene and several studies have identi?ed major QTLs for PHS and seed dormancy on the long arm of chromosome 4A (Anderson et al. 1993; Kato et al. 2001; Flintham et al. 2002; Lohwasser et al. 2005; Mares et al. 2005; Torada et al. 2005; Chen et al. 2008; Ogbonnaya et al. 2008; Liu et al. 2011). Because previous QTL mapping studies for PHS in wheat were conducted using biparental mapping populations, they sampled only allelic variability present between the two parental genotypes. The limitations include low resolution for the QTL location and the effects of individual QTLs are overestimated. Use of these QTL in the marker-assisted selection program often does not give expected results, because the biparental population may not represent the breeding material or the important QTLs do not segregate in the biparental population and are not detected. Linkage disequilibrium (LD) based association mapping (AM) has a number of advantages over other mapping techniques including the potential for increased QTL resolution, and an increased sampling of genetic variation. This is because the panel of accessions used can be more informative than the offspring of biparental crosses, can be representative of breeding germplasm, and can take into account historical recombination events (Flint-Garcia et al. 2003; Gupta et al. 2005; Breseghello and Sorrells 2006a; Yu and Buckler 2006; Nordborg and Weigel 2008; Zhu et al. 2008). Therefore, AM is receiving increasing attention as a method complementary to bi-parental mapping populations to associate the genotype to phenotype based on LD by exploiting the genetic variation present in germplasm collections (Breseghello and Sorrells 2006a). Often, the AM studies are carried out using exotic germplasm, which facilitates the discovery of novel alleles, but can limit the application in practical plant breeding programs and multiple alleles reduce the power of detection. Although

there have been a number of AM studies on wheat (Breseghello and Sorrells 2006b; Crossa et al. 2007; Reif et al. 2011), to date there is no genome wide association study for PHS. In nature, developmental traits are under the control of genes, which are expressed at speci?c stages of development in response to the existing environmental conditions. The mapping of such genes has been referred to as functional genome-wide association mapping (fGWAS) (Wu and Lin 2006; He et al. 2010). For a trait like PHS, the period of AR is very important for the expression of the trait and maximizing the differences among genotypes. Evaluation of genotypes for PHS in response to AR periods of different duration may identify signi?cant QTLs that could escape detection when the genotypes are tested for PHS after a single AR period. The present study was undertaken with the objective of mapping the QTLs for PHS in an elite association mapping panel of white winter wheat genotypes using DArT and SSR markers. In addition, fGWAS was implemented by testing the spikes for PHS response at different durations of the AR period. Here we report the identi?cation of new QTLs, validation of some of the earlier known QTLs for PHS in white winter wheat and the importance of using elite material in association mapping studies.

Materials and methods Plant material The association mapping panel used in the present study comprised a set of 208 elite soft white winter wheat lines including some released varieties, parental genotypes used in the breeding program at Cornell University as well as advanced breeding materials. Except for two genotypes, all genotypes were selections from breeding populations generated by the Cornell wheat breeding program. Reexamination of the kernel color of these accessions revealed ten genotypes with either red kernel color or segregation for kernel color and they were removed from the analysis leaving 198 genotypes for association mapping. PHS experimental design The experimental material was planted in 1 m rows in a randomized complete block design with two replications over the period of 4 years from 2008 through 2011 in the ?elds near Ithaca, NY, USA. Field, location means, standard deviations, and average temperatures and precipitation for a fortnight before the last harvest date of spikes are given in Table 1.

123

Theor Appl Genet (2012) 125:793–805

795

Table 1 Descriptive statistics of the 198 genotypes in the association mapping panel for pre-harvest sprouting (PHS), heading date (HD) and plant height (Pl Ht) Trait PHS PHS PHS PHS PHS PHS HD (days) Pl Ht (cm) Field Location Snyder Snyder McGowan McGowan McGowan McGowan McGowan McGowan Year 2008 2009 2010 2011 (3 days) 2011 (5 days) 2011 (7 days) 2011 2011 Mean ± SD 1.7 ± 0.80 1.38 ± 0.94 2.25 ± 1.31 2.50 ± 1.16 2.60 ± 1.13 2.96 ± 1.12 0.00 ± 2.11 94.01 ± 10.63 Range (min/max) 0.00 to 5.20 0.00 to 4.70 0.00 to 5.20 0.30 to 5.70 0.20 to 5.95 0.55 to 5.85 -3.48 to 4.52 68.00 to 114.00 Temp. range (oC) 26.3/13.8 23.2/11.2 25.9/13.7 26.8/12.9 Precipitation (cm) 5.7 4.5 2.2 3.9

The HD trait is plus or minus days based on the location mean set equal to zero SD standard deviation, min minimum, max maximum

The PHS evaluation was as described by Anderson et al. (1993). Brie?y, for the 2008 through 2010 trials, ?ve spikes were harvested from each plot when they reached physiological maturity and were dried for 5 days at ambient temperature and humidity and later placed in the mist chamber for PHS evaluation for 4 days. During 2011, a total of ?fteen spikes were harvested from each plot at physiological maturity and ?ve spikes each were transferred to the mist chamber for PHS evaluation after 3, 5 and 7 days of drying (after-ripening period). PHS data were recorded for each spike on a scale of 0 (no visible sprouting) to 9 (extensive sprouting) and averaged for the ?ve spikes. All the genotypes were not harvested on the same day because of differences in the date of physiological maturity. Consequently, the ?nal PHS scores were adjusted for harvest date by subtracting the mean of the PHS scores of the genotypes harvested on the same day from the PHS score of the individual genotypes and then adding the mean PHS score of all 198 genotypes. Heading date and plant height During the 2011 trial, heading date (HD) was measured by recording the date on which spike emergence was complete for 50 % of the plot, whereas plant height (Pl Ht) was measured to the top of the spike for the mean of the plot. The data on HD and Pl Ht were recorded to compare the results of PHS with these traits, particularly HD, for which a major QTL has been reported on chromosome 2BS in our earlier study (Munkvold et al. 2009) in the vicinity of the photoperiod gene Ppd-B1 (Hanocq et al. 2004). DNA isolation and genotyping Two types of molecular markers, Diversity Array Technology (DArT) and simple sequence repeat (SSR), were

used in the present study. DNA was extracted from young seedlings following Heun et al. (1991) and sent to Triticarte Pty Ltd, Australia (http://www.triticarte.com.au/) for whole-genome pro?ling using DArT markers. The SSR marker data were generated using direct-labeled primers as well as M-13 tailed primers and genotyped using an ABI 3730 at Cornell University Life Sciences Core Laboratory Center as described in Yu et al. (2010). Some of the SSR marker data were generated using polyacrylamide gel electrophoresis followed by silver staining. The DArT genotypic data consisted of 1,103 markers, whereas the SSR marker data comprised 62 loci. In addition, one polymerase chain reaction (PCR) based landmark unique gene (PLUG) marker (Ishikawa et al. 2007) was also used, thus making a total of 1,166 markers. The Wheat Interpolated Maps v4 (Triticarte Pty Ltd, Australia, personal communication) were used as references to locate the positions of the majority of the DArT markers. Apart from this, some DArT markers were assigned to respective chromosomes based on earlier published literature as well as on the association (r2 [ 0.8) with other mapped markers. The chromosome location of some of the DArT markers used in the present study was not yet known, however, the other DArT markers covered the entire genome of wheat. The SSR markers corresponded to previously reported locations of PHS QTL and were on chromosomes 2B, 2D, 4A, 6D and 7B. The total number of markers after removing markers with\10 % MAF was 841 and after removing markers with \10 % MAF and r2 = 1 was 657. Population structure Population structure was assessed using a set of 42 unlinked DArT and SSR markers representing all the chromosomes. Software package Structure 2.3.3 (Pritchard

123

796

Theor Appl Genet (2012) 125:793–805

et al. 2000) was used with the admixture model and the number of sub-populations (K) as 1 through 20 with the variable length of burn-in period and number of iterations set at 20,000–250,000. Statistical analysis Simple correlations between PHS scores across the years as well as Spearman’s rank correlations for PHS scores, broad sense heritability estimates, and descriptive statistics for different traits were calculated in a spreadsheet. The best linear unbiased predictors (BLUP) for the PHS data of 4 years (2008 through 2011) were computed using the statistical software package JMP (http://www.jmp.com) treating genotypes as random effects. Additive main effects and multiplicative interaction (AMMI) bi-plot was constructed using MATMODEL 3.0 (Gauch 2007). Association analysis Association analysis was performed using general linear model (GLM) and mixed linear model (MLM) approaches as implemented in the software package TASSEL 3.0 (Bradbury et al. 2007). Out of a total of 1,166, the markers with minor allele frequencies (MAF) less than 10 % were excluded from analysis. The markers with r2 = 1 (r2 represents the correlation between alleles at two loci and a value of 1 indicates that markers were in complete LD) were also excluded from the analysis and the one marker with the least amount of missing data was retained. To take into account the population structure, Q values were determined using the software STRUCTURE, and principal components (PCs) were obtained using TASSEL 3.0 and used as covariates in GLM and MLM analysis. For MLM analysis, marker-based kinship matrix (K) obtained using TASSEL was used along with the Q matrix to correct for both family and population structure and the phenotypic variation explained (r2) by the marker is reported. In the case where a marker was detected only in GLM, the corresponding r2 values are reported. The variation explained by a marker locus is calculated differently for each method (GLM or MLM) so the r2 values are not directly comparable. The optimum levels of compression in combination with P3D for variance component estimation were used as criteria in MLM. The PHS data of individual years as well as BLUP for the PHS scores of 4 years and for the two locations [Snyder (2008 and 2009 environments) and McGowan (2010 and 2011 environments)] were analyzed separately. For calculating the BLUPs for 4 years (2008–2011), the PHS data from the 5-day after-ripening period from 2011 trial were used along with data from 2008 to 2010 trials. For

functional GWAS, PHS data for the 2011 experiment from the three different durations of after ripening periods (3, 5 and 7 days) were analyzed separately as well as by taking the average of the three periods. In addition to this, analysis was also performed using all the 1,166 markers as well as after excluding markers with \10 % MAF following GLM and MLM approaches with and without the Q matrix. However, we have reported here the results obtained using the robust set of data after removing markers with \10 % MAF and r2 = 1. Signi?cant marker-trait associations were assigned based on the false discovery rate (FDR) criteria of Benjamini and Hochberg (1995) (called stringent FDR criteria hereafter) and the results are presented in Tables. At the same time other FDR criteria of q values corresponding to the p values obtained in GLM and MLM analyses were used to assign the signi?cant associations using the software QVALUE (Storey 2002). More markers were signi?cant using these criteria and results are presented in supplementary Tables.

Results Phenotypic data The PHS data across years were normally distributed except for 2009. Therefore, 2009 data were log transformed and standardized before analysis using an option available in TASSEL. The PHS scores were higher for the McGowan location as compared to the Snyder location (Table 1). Analysis of variance (ANOVA) indicated highly signi?cant differences (p \ 0.01) for PHS scores across years (data not shown). The broad sense heritability (h2) across the 4 years was 0.37, whereas by locations, it was much higher for the McGowan location (0.71) as compared to the Snyder location (0.50). For the different durations of after-ripening period, heritabilities were 0.68, 0.70 and 0.63 for the 3, 5, and 7 day AR durations, respectively. Signi?cant correlations as well as rank correlations were observed for the PHS scores across the years except for the PHS scores of 2008 with 2009 and 2011 (Table 2). Similarly, signi?cant correlations and rank correlations were observed for PHS scores for the different durations of AR period in the 2011 experiment (Table 2). There were nonsigni?cant negative correlations between the PHS score and plant height and days to heading, and a signi?cant positive correlation between plant height and days to heading (Table 2). As the values of correlations and the rank correlations were similar, only the values of the correlations are given in Table 2.

123

Theor Appl Genet (2012) 125:793–805 Table 2 Correlation coef?cients for the pre-harvest sprouting (PHS) scores (a) across the years and (b) for different days of after-ripening period and plant height (Pl Ht) and heading date (HD) for the 2011 trial PHS08 (a) PHS09 PHS10 PHS11 Meann BLUPs 0.06 0.14* 0.06 0.40*** 0.40*** 3 days (b) 5 days 7 days Pl Ht HD 0.65*** 0.58*** -0.06 0.01 0.65*** -0.10 -0.08 -0.08 -0.08 0.23*** 0.37*** 0.25*** 0.63*** 0.63*** 0.46*** 0.82*** 0.82*** 0.74*** 0.74*** 7 days 0.99*** Pl Ht PHS09 PHS10 PHS11 Meann

797

5 days

Signi?cant at * p \ 0.05 and *** p \ 0.001, respectively
n

Mean of 4 years PHS data from 2008 through 2011

group consisted of genotypes related to NY91017-8080. As some of the genotypes from sub-groups 1 and 2 overlapped, it was considered to be one larger sub-group. To examine the effect of population structure, the association analysis was conducted without the population structure covariarates as well as with different numbers of subpopulations and their respective Q matrices. In addition to this, different numbers of PCs were used as covariates in the analysis. Because of the relatively small contribution of individual components using PCA (Fig. 1), two subgroups/two components were considered to be optimal. Moreover, similar results obtained using Q matrices from STRUCTURE and principal components considering Q = 2 supported the same sub-groups. Only results obtained using PCs as covariates are presented in Tables. As such, there was little in?uence of population structure on the signi?cant marker-trait associations. The only effect that was observed was for the marker gwm397 mapped on chromosome 4AL. This marker was not signi?cant when population structure was not taken into account, although another linked marker, gwm610, was signi?cant (see ‘‘Discussion’’). Signi?cant marker-trait associations for PHS

Population structure A set of 42 markers (one marker from the short and long arm of each chromosome) was used to estimate the underlying population structure. The logarithm of the probability of likelihood [LnP(D)] value declined after K = 2 and then again increased. The principal component analysis (PCA) based on all the markers identi?ed three sub-groups (Fig. 1). The ?rst sub-group consisted of genotypes mostly related to ‘Whatford’ and NY87048W7388, while the second sub-group consisted of genotypes mostly related to ‘Cayuga’ and ‘Caledonia’. The third subFig. 1 Scatterplot of principal component 1 (PC1) drawn against principal component 2 (PC2). Each circle or dot or plus sign represents an individual genotype in different subpopulations of the association mapping panel. Subpopulations are enclosed by ovals in continuous and dashed lines and labeled according to the genetic background by arrows. Proportion of the variance explained by the principal components is given in parenthesis
20 15 10
NY91017-8080 background

A total of eight QTLs linked with 13 markers representing seven chromosomes were detected for PHS in this study (Table 3; Supplementary Table S1). One QTL was detected on each arm of chromosome 2B and one each on chromosome arms 1BS, 2DS, 4AL, 6DL, 7BS and 7DS. The signi?cant markers obtained in GLM and MLM were different for the Snyder (2008 and 2009 environments) and McGowan locations (2010 and 2011 environments) and none of these QTLs was signi?cant in all the environments.

PC 2 (7.48%)

5 0
-5 -10 -15
Background of Cayuga, Caledonia

Background of Whatford, NY87048W-7388, other

-20 -25 -20 -15 -10 -5 0 5 10 15 20

PC 1 (18.45%)

123

798

Theor Appl Genet (2012) 125:793–805

However, a QTL on 4AL linked with marker gwm397 was signi?cant in three out of four environments as well as for the BLUPs over environments (Table 3). This QTL was also linked with two other markers (gwm610 and wPt730913) on the same chromosome arm and explained up to 12 % of the phenotypic variation. Although the genetic position of the marker wPt-730913 is far from the markers gwm397 and gwm610 based on two different maps (Triticarte personal communication; Somers et al. 2004), because they were on the same arm of the 4A chromosome, we counted it as a single QTL. However, it is possible that these are two separate QTLs. The QTLs on chromosomes 4AL and 7DS were prominent in the Snyder location (2008 and 2009 environment), while QTLs on chromosomes 1BS, 2BS, 2BL, 6DL and 7BS were prominent at McGowan locations (2010 and 2011 environments). All these QTLs were also detected with varying level of signi?cance using BLUPs calculated over four environments. The QTL on 2DS was signi?cant using BLUPs, but not in the individual environments (Table 3). As most signi?cant marker-trait associations were not consistent across individual environments and locations, the additive main effects and multiplicative interaction (AMMI) biplot were constructed following Gauch et al. (2011). The AMMI bi-plot as seen in Fig. 2 showed that the two locations, Synder (2008 and 2009 environment) and McGowan (2010 and 2011 environments), not only differ from each other for their main effects (horizontal distribution) but also for their interaction effects (vertical distribution). Therefore, BLUP values were also calculated for individual locations across 2 years and used for analyzing the data to facilitate interpretation of the results. Using this approach, two QTLs representing chromosomes 4AL and 5B/7B were detected across both locations (Table 4). The genetic location of marker wPt3873 is uncertain and has been mapped on chromosomes 5B and 7B in different studies (Triticarte Pty Ltd, Australia, personal communication; Crossa et al. 2007). This QTL was not signi?cant using the data from individual environments. Contrary to this, four QTLs that were detected in individual environments (2DS, 6DL, 7BS and 7DS; from Table 3) did not pass the stringent FDR criteria using this approach, although three of them (excluding 7DS) were signi?cant for q values (Table 4; Supplementary Table S2). All the markers that were signi?cant in multiple environments had allelic effects in the same direction in each environment (Supplementary Table S1). The results of marker-trait association presented in Tables 3, 4 and 5 as well as Supplementary Tables show all markers that were signi?cant for at least one location in a single year according to the stringent false discovery rate criteria and the q value criteria, respectively. When the analysis was conducted using all 1,166 markers, as well as after excluding markers with \10 %

MAF using GLM and MLM approaches with and without Q matrix, the results obtained for the signi?cant associations were similar to those reported here (data not shown). Ignoring the stringent false discovery rate criteria, large numbers of marker-trait associations were observed for PHS across the years at p \ 0.05 using GLM and MLM. The majority of signi?cant associations obtained through MLM was also observed using GLM (Supplementary Table S3). Functional GWAS As expected, mean PHS scores were lower for the 3-day AR period and higher for the 7-day AR period. However, the greatest range for PHS scores was at the 5-day AR period re?ecting a higher genetic variance (Table 1). Four DArT markers linked with three QTLs were signi?cant for PHS for different durations of AR period. However, only two of these markers [wPt-0697 (chromosome 2BL) and wPt-666223 (chromosome unknown)] were common for all three durations of AR period (Table 5), whereas the remaining two markers [(wPt-666931 (chromosome 2BL) and wPt-8283 (chromosome 7BS)] were signi?cant for only the 5 and 3 days of AR, respectively. All the markers signi?cant for both the 5- and 7-day AR periods were also detected for the 3-day AR period and all were detected using the mean data calculated over the 3-, 5- and 7-day AR periods with varying degrees of signi?cance (Table 5). There were additional markers with lower p values that were signi?cant using q value criteria, but because they did not pass the stringent FDR criteria, they are not listed in Table 5. Details of these markers are given in Supplementary Table S4. Signi?cant QTLs for heading date and plant height The data on heading date and plant height were recorded during 2011 only. For heading date, a highly signi?cant QTL linked with the marker gwm429 mapped on chromosome 2BS was identi?ed. For plant height, a signi?cant QTL was linked with DArT marker wPt-730772 on chromosome 6AS (Supplementary Table S5). Neither of these two marker loci was signi?cantly associated with PHS QTLs in the present study.

Discussion Because of the practical limitations of QTL mapping in biparental populations, association mapping has gained wide acceptance as an ef?cient method for mapping QTLs in plant populations. This has been facilitated by the

123

Table 3 Markers showing signi?cant association with pre-harvest sprouting in general linear model (Q) and mixed linear model (Q ? K) by location and year
Snyder location 2008 p Q – 5.71E205 9.74E205 1.92E–04 2.99E-03 5.49E-03 4.15E-04 – – 1.59E-03 – 1.16E204 1.12E-02 – – – 1.97E-02 3.91E-02 – – – 3.66E-03 – 0.04 0.05 – – 0.05 – 0.03 0.02 – 4.51E-02 0.02 2.00E-02 0.03 2.33E-03 0.05 – – – – 3.23E-02 4.05E-03 3.63E208 1.01E207 5.03E205 – 2.22E-02 – 1.85E204
2

Theor Appl Genet (2012) 125:793–805

Marker 2009 R Q?K – – – – – – 2.87E-03 3.59E206 5.78E206 3.94E-04 – – – 6.01E-03 0.04 – 0.03 – 0.07 0.11 3.02E-03 – 9.27E-04 – – – – – – 0.06 0.02 0.05 0.05 0.12 0.02 – – – – 1.00E-02 – 0.03 – – 4.83E-02 0.02 – – – – – – – – – – – – 1.40E-03 0.06 3.54E-02 2.19E-02 8.31E-03 3.70E-03 3.83E205 – – – – – – 2.83E204 – Q Q?K Q Q?K Q
2

Chromosome armA 2010 R
2

McGowan location 2011 R
2

Putative position (cM)B p p p Q?K – – – 2.12E-02 4.18E204 – – – – – – 2.61E-03 –

BLUP (2008–2011) R
2

p Q 0.02 0.03 0.03 0.03 0.07 – – – – – – 0.05 – 2.47E-03 1.80E-03 5.18E-04 2.33E204 2.22E204 2.43E204 1.35E-03 1.42E-03 3.91E205 – 3.62E-03 7.82E-04 3.65E-02 Q?K 1.84E-02 3.14E-02 4.19E-02 1.01E-02 7.96E-03 7.59E-03 7.29E-03 4.39E-03 1.99E-03 – – 4.12E-03 –

R2

wPt-666564

1BS

33.5

a

0.03 0.02 0.02 0.04 0.04 0.04 0.04 0.04 0.05 – 0.04 0.04 0.02

barc328.217

2BS

20.2b

stm773.183

2BS

48.0c

wPt-0697

2BL

86.6a

wPt-666931

2BL

97.8a

wPt-6003 4.80E-03 3.39E-02 2.32E204 – – –

2DS

14.7b

gwm610.162

4AL

12.0d

gwm610.172

4AL

12.0d

gwm397.191

4AL

18.0d

wPt-730913

4AL

93.1a

cfd37.208

6DL

60.0d

wPt-8283

7BS

46.5a

wPt-663918

7DS

00.5a

R : variation explained by the marker in MLM (when marker is signi?cant only in GLM, R value corresponding to it is shown)

2

– Denotes p [ 0.05; p values which passed the false discovery rate criteria are highlighted in bold and their corresponding p values in different years are given only if they were \0.05

A

Markers on the same chromosome arm are considered to be linked with the same QTL

B

Putative genetic position based on aTriticarte map (Triticarte Pty Ltd., Australia personal communication), bMunkvold et al. (2009), cLehmensiek et al. (2005), dSomers et al. (2004)

799

123

800 Fig. 2 The AMMI1 biplot for pre-harvest sprouting (PHS) data used in association mapping. The abscissa shows main effects, namely genotype means and environment means, and the ordinate shows interaction principal component 1 (IPC1) scores. The vertical line indicates the grand mean (a PHS score of 1.99) and the horizontal line indicates an IPC1 score of 0; Sny08 Snyder 2008, Sny09 Snyder 2009, McG10 McGowan 2010 and McG11 McGowan 2011
3 Sny08 2

Theor Appl Genet (2012) 125:793–805

Sny09

1

IPC1

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

-1

McG11

-2 McG10 -3

Main effects

Table 4 Signi?cant marker-trait associations for pre-harvest sprouting obtained through general linear model (Q) and mixed linear model (Q ? K) using best linear unbiased predictors (BLUPs) of individual locations Marker chromosome armA Putative position (cM)B BLUP (Snyder) p Q wPt-666564 barc328.217 stm773.183 wPt-666931 gwm610.162 gwm610.172 gwm397.191 wPt-3873 1BS 2BS 2BS 2BL 4AL 4AL 4AL 5B/7B 33.5a 20.2b 48.0c 97.8
a

BLUP (McGowan) R
2

p Q Q?K 2.80E-03 9.03E-03 1.92E-02 3.18E-03 – – – 4.11E-02

R2

Q?K – – – – 6.75E-06 6.63E-06 4.10E-05 1.36E-03 – – – – 0.11 0.11 0.09 0.06

– – – – 1.69E-07 2.29E-07 1.17E-05 1.71E-04

2.77E-04 1.16E-04 1.19E-04 1.18E-04 – – 4.57E-03 2.72E-02

0.05 0.04 0.03 0.05 – – 0.04 0.02

12.0d 12.0d 18.0d 27.3a/61.5e

R2: variation explained by the marker in MLM (when marker is signi?cant only in GLM, R2 value corresponding to it is shown) – denotes p [ 0.05; p values which passed the false discovery rate criteria are highlighted in bold and their corresponding p values in different years are given only if they were \0.05
A B

Markers on the same chromosome arm are considered to be linked with the same QTL

Putative genetic position based on aTriticarte map (Triticarte Pty Ltd., Australia personal communication), bMunkvold et al. (2009), cLehmensiek et al. (2005), dSomers et al. (2004), eCrossa et al. (2007)

availability of increasing numbers of markers and advances in computational tools. As a result, several association mapping studies have been reported in wheat that mostly focused on quality traits (Breseghello and Sorrells 2006b; Emebiri et al. 2010; Reif et al. 2011), disease resistance (Crossa et al. 2007; Miedaner et al. 2011; Yu et al. 2011) and yield and related traits (Crossa et al. 2007; Yao et al. 2009; Neumann et al. 2011). In the majority of these

studies, either a small set of genotypes (91 genotypes and 395 markers; Emebiri et al. 2010) or markers (96 lines and 93 markers; Breseghello and Sorrells 2006b) or both have been used to ?nd the marker-trait associations for different traits. In this study, we used a large number of genotypes and markers for identifying the QTLs for PHS in elite wheat germplasm using both general-linear and mixedlinear models.

123

Theor Appl Genet (2012) 125:793–805

801

Table 5 Signi?cant marker-trait associations for pre-harvest sprouting obtained through general linear model (Q) and mixed linear model (Q ? K) for the data of 2011 for different durations of after-ripening period
Marker chromosome arma Putative position (cM)b 3 days p Q wPt-666223 wPt-0697 wPt-666931 wPt-8283
2

5 days R Q?K 4.01E-03 1.82E-02 – 4.07E-03 0.04 0.03 0.03 0.04
2

7 days R Q?K 1.17E-02 2.12E-02 4.18E-04 2.61E-03
2 2

Mean R Q?K 2.85E-02 4.49E-04 – – 0.03 0.07 – –
2

p Q 7.52E-04 3.70E-03 3.83E-05 2.83E-04

p Q

p Q 2.27E-04 2.75E-04 1.46E-03 5.72E-04 Q?K 5.22E-03 2.33E-03 6.41E-03 7.80E-03

R2

? 2BL 2BL 7BS

– 86.6 97.8 68.9

2.22E-04 5.30E-03 2.58E-02 1.74E-04

0.04 0.03 0.07 0.05

1.31E-02 1.83E-04 – –

0.04 0.05 0.04 0.04

R : Variation explained by the marker in MLM (when marker is signi?cant only in GLM, R value corresponding to it is shown) – denotes p [ 0.05; p values which passed the false discovery rate are highlighted in bold and their corresponding p values in different years are given only if they were \0.05
a b

Markers on the same chromosome arm are considered to be linked with the same QTL Genetic position based on Triticarte map (Triticarte Pty Ltd., Australia personal communication)

Effect of population structure The results of association mapping can be in?uenced by the presence of underlying population structure. However, the association mapping panel used in the present study had minimal subpopulation structure that was evident from the relatively small contribution of the individual components in PCA (Fig. 1). The only effect of population structure was on the highly signi?cant marker gwm397.191 on chromosome 4AL. Without correction for population structure (GLM without Q matrix), this marker was not signi?cant, whereas another marker from chromosome 4AL (gwm610) linked with the same QTL was highly signi?cant. After taking into account the population structure using either the Q matrix from STRUCTURE or principal components (PCs) and/or using the kinship matrix (K), marker gwm397.191 was signi?cant. Genetically, both of these markers (gwm397 and gwm610) have been mapped very close (6 cM) to each other in the wheat consensus map (Somers et al. 2004) and were linked with the same QTL in the material used in the present study. In the majority of the AM studies, the Q matrix obtained using STRUCTURE was used to account for population structure, while in some studies PCs were also used. Price et al. (2006) have advocated the use of PCs to effectively account for population strati?cation. In the present study, we used both (Q matrix from STRUCTURE, as well as PCs) and they produced similar results. Although little gain was achieved by accounting for population structure in this study, the use of Q and Q ? K matrices reduced the probability of false positives (Fig. S1). The uni?ed mixedmodel method used in the present study has been reported to reduce the Type I error in association mapping (Yu et al. 2006). Removing up to 20 % of the related genotypes based on cluster analysis as suggested by Breseghello and

Sorrells (2006b) did not change the results except that slightly higher p values were obtained, suggesting minimal population structure in this population. It has also been suggested that for some traits and populations, the K-only model may be as good as or better than the Q ? K model (TASSEL user manual). In our study, the results obtained using both K-only and Q ? K were not different for the most signi?cant markers (data not shown). Although the Q-model (GLM) was not as effective as the K-only or Q ? K model (MLM) for controlling population structure, inclusion of both GLM and MLM results in the present study suggests that some signi?cant associations may not be detected using MLM because they do not pass the FDR criteria. It has also been suggested by Moncada et al. (2001) that in QTL mapping, the most conservative approach is to focus on those QTLs that are detected by more than one analysis or in multiple environments. The use of both models also facilitates comparisons with previous studies. Importance of elite material for association mapping In genome-wide AM studies, the results are strongly in?uenced by the choice of germplasm, size of the population, number and distribution of markers, and the trait under study. Often, in AM studies, exotic germplasm or diverse materials representing different geographic regions of the world are used to minimize LD and ?nd QTLs for a variety of traits. Although the results obtained from such studies may provide some insight about the diversity, it is not usually optimal for application in a breeding program. Entries from a core collection from the germplasm bank or exotic material may be suited for AM of traits less in?uenced by adaptation, however, elite material is more useful for studying quantitative traits such as PHS that can be

123

802

Theor Appl Genet (2012) 125:793–805

strongly in?uenced by biotic and abiotic stresses and differences in maturity. For an adaptive trait like PHS, it is not feasible to accurately phenotype an association panel of diverse wheat genotypes from different regions because diseases and maturation at different times under different climatic conditions can affect seed development and confound the phenotype. Moreover, the QTLs identi?ed using elite material can have direct application in the breeding program through marker-assisted selection because many of the same alleles will be represented in an association panel (Breseghello and Sorrells 2006a). Another important advantage of association mapping is that more QTLs can be detected than in a biparental mapping population. The present study is an excellent example of such a situation, wherein, we have identi?ed a major QTL for PHS on chromosome 4AL (previously reported by other investigators) that was present in our breeding lines, but was a minor QTL (LOD score of 3.56 in one environment) based on the results of a biparental mapping population (Munkvold et al. 2009). In the same study, the QTL on chromosome 2BS was detected in 16 environments with high signi?cance (Munkvold et al. 2009). In addition, this study has identi?ed additional markers around the previously identi?ed major QTL on 2BS that are useful for ?ne mapping. The availability of large numbers of markers and improved statistical tools greatly increases the ef?ciency of identifying QTLs and the use of MAS in breeding programs. QTLs 9 environment interaction In QTL mapping studies, large numbers of QTLs are often identi?ed, but only a few of them are signi?cant in across all environments. One of the important reasons for this is the QTL 9 environment interactions that result in an allele in?uencing the trait in only certain environments (Gauch et al. 2011). In this study, none of the QTLs was detected in all four environments, although a QTL on chromosome 4AL linked with marker gwm397 was detected in three environments. While some QTLs were signi?cant at the Snyder location (2008 and 2009 environment), others were signi?cant at the McGowan location (2010 and 2011 environment). QTL 9 environment interactions have been reported for PHS in other studies (Kulwal et al. 2004; Imtiaz et al. 2008; Liu et al. 2011). However, those studies did not offer any solution to deal with inconsistencies of QTL detection across environments. To address this concern and to increase the ef?ciency of QTL detection under such circumstances, recently Gauch et al. (2011) proposed two new strategies. The additive main effects and multiplicative interaction (AMMI) model proposed by them improved the accuracy of the phenotypic data which

increased the likelihood of QTL detection. They demonstrated improved accuracy using the PHS data from the study of Munkvold et al. (2009). The AMMI bi-plot (Fig. 2) of our data clearly separated the Snyder and McGowan locations (these locations are actually in close proximity) in terms of main effects and interaction effects. Therefore, when we analyzed BLUPs of the two locations separately, besides the QTL on 4AL, another QTL linked with marker wPt-3873 (5B/7B) was signi?cant across environments/locations, whereas the QTLs on chromosomes 1BS, 2BS and 2BL were signi?cant at the McGowan location (Table 4). Functional GWAS The concept of functional mapping of QTL may provide insight into the genetic control of developmental traits. This has been referred to as time-related mapping (TRM) versus time-?xed mapping (TFM) by Wu et al. (1999). It is logical that during the process of development, different QTLs are expressed at different times in response to the existing environmental conditions and developmental stage (Wu and Lin 2006; He et al. 2010), although they may have a similar effect on the ?nal phenotype. In this study, we used three different periods of after-ripening to perform functional mapping for PHS. Although, the broad sense heritability estimate among the three durations of AR period was highest for the 5-day period (0.70), and can be considered as an optimum time period of AR for the spikes to be tested for PHS, it was not signi?cantly different from that observed for the 3-day period (0.68). Moreover, all the QTLs detected for the 5-day AR period were also detected for the 3-day AR period. Therefore, given similar environmental conditions, the 3-day AR might also be used for PHS evaluation in future studies. As these results were based on the data from one location (2011), additional studies are needed to con?rm this. Comparisons with previous results All 21 chromosomes of bread wheat have been reported to contain minor or major QTLs for PHS, seed dormancy and other related parameters including grain color, falling number and a-amylase (reviewed by Kulwal et al. 2010). In this study, a total of eight QTLs were signi?cantly associated with PHS. The QTL linked with the markers gwm397 and gwm610 on the long arm of chromosome 4A was highly signi?cant and explained up to 12 % of the variation, and is likely the same QTL on the long arm of chromosome 4A that has been detected in the majority of PHS studies (Flintham et al. 2002). Among these studies, SSR marker gwm397 was linked to a major QTL for PHS

123

Theor Appl Genet (2012) 125:793–805

803

(Ogbonnaya et al. 2008), seed dormancy (Mares et al. 2005; Torada et al. 2005; Tan et al. 2006) and both, PHS and seed dormancy (Chen et al. 2008; Liu et al. 2011; Fig. S2). Besides gwm397, another SSR marker gwm610 was linked with the same QTL in this study. Although, gwm610 was used in many of the earlier studies, it was never reported to be linked with PHS and would be useful for a MAS program aimed at PHS improvement. In our earlier study, a minor QTL in the marker interval gwm397wmc617 was detected in the Cayuga 9 Caledonia DH population explaining only 3 % of the variation (Munkvold et al. 2009). Munkvold et al. (2009) reported a major QTL on the short arm of chromosome 2B in the marker interval barc55wmc474 that was detected in all 16 environments with a LOD score of up to 27.10 that explained up to 24 % of the variation. Although, this QTL was not consistent in this marker interval (barc55-wmc474) in all environments in this study, based on the earlier results (Munkvold et al. 2009) and our ?ne mapping studies for this QTL, additional markers were mapped in this region. Based on this study and a related ?ne mapping study, we were able to resolve this as two different QTLs linked in coupling (unpublished results). Although, neither barc55 nor wmc474 was signi?cant, another nearby SSR marker, barc328, was significant, along with SSR marker stm773 that explained up to 5 % of the variation. Interestingly, another QTL on chromosome arm 2BL linked to DArT marker wPt-2397 was detected in only one of the 16 environments in the earlier study (Munkvold et al. 2009) but was signi?cant in this study. This QTL was linked with the DArT markers wPt0697 and wPt-666931. This again illustrates the ef?ciency of AM to detect and validate QTLs in elite breeding germplasm. In both of these studies, this QTL explained up to 4 % of the variation. In the biparental population, out of 15 QTLs detected for PHS, only two of them (one each on 2BS and 6DL) explained more than 10 % of the variation (Munkvold et al. 2009) indicating the complexity of this trait. Similarly, in the present study, out of eight QTLs, only one on 4AL was a major one. In addition to the important QTLs on chromosome 2BS, 2BL and 4AL, we identi?ed ?ve more QTLs located on chromosome arms 1BS, 2DS, 6DL, 7BS and 7DS. Significant QTLs for PHS and dormancy have been identi?ed on these chromosomes in earlier studies (Kulwal et al. 2004; Munkvold et al. 2009; Knox et al. 2012; Liu et al. 2011). For example, two QTLs for PHS were identi?ed by Munkvold et al. (2009) on chromosome 1B. However, it is dif?cult to cross reference QTLs in different studies. Comparative analysis of different wheat maps (Roder et al. 1998; Somers et al. 2004; Sourdille et al. 2004; Crossa et al. 2007; data available on GrainGenes) assigned the centromere to different positions leading to differences in

chromosome arms assignments, particularly for the markers near the centromeric region. This may be the case for marker wPt-666564 mapped on chromosome 1B, although we putatively assigned it to the short arm of 1B. Similarly, for the marker wPt-8283 mapped on chromosome 7B, the arm location is unclear but based on several published maps, it appears to be on 7BS. Kulwal et al. (2004) identi?ed a QTL for PHS on the short arm of chromosome 2D in the marker interval gwm261-cdo1379 in the ITMI population. Comparison of the maps of Roder et al. (1998), Somers et al. (2004) and that of Munkvold et al. (2009) suggests that the signi?cant QTL associated with marker wPt-6003 identi?ed in the present study on 2DS corresponds to the QTL identi?ed by Kulwal et al. (2004) on 2DS. Similarly, signi?cant marker cfd37 mapped on 6DL identi?ed in this study is the same that was earlier determined to be linked to the PHS QTL by Munkvold et al. (2009). However, in the biparental population this QTL explained 14 % of the variation as compared to 4 % in the association mapping panel used in this study. In a very recent study, Knox et al. (2012) identi?ed two QTLs for germination index in a durum wheat population on chromosome 7B. Although there are other reports of PHS QTLs on chromosome 7B (Kulwal et al. 2004; Mohan et al. 2009; Liu et al. 2011), they have all been identi?ed to be interacting QTLs without any main effect. Therefore, we believe that the main effect QTL identi?ed in this sutdy on 7B is novel in hexaploid wheat. Using a DH population, Munkvold et al. (2009) identi?ed four QTLs on chromosome 7D and one was linked with the marker wPt-0934. According to the unpublished maps from Triticarte Pty Ltd., Australia, the marker, wPt-663918, that was associated with PHS in our study is located *3 cM from the marker wPt-0934 suggesting that a similar QTL was identi?ed in both of these studies. Although some of the QTLs identi?ed in this study were detected in the same genomic regions identi?ed earlier, it should be noted that the association panel used in this study was unrelated to germplasm in earlier reports except for Anderson et al. (1993) and Munkvold et al. (2009). However, given the statistical issues associated with individual studies and the lack of common markers, direct comparisons between the different studies need to be further veri?ed.

Conclusion This study has identi?ed a previously reported major QTL for PHS on chromosome 4AL in unrelated germplasm and that was not previously identi?ed in our breeding program. Also, a novel QTL on chromosome 7BS in our elite white winter wheat materials was identi?ed. This highlights the importance of association mapping for identifying QTL on

123

804

Theor Appl Genet (2012) 125:793–805 University, Ithaca. Available at http://www.css.cornell.edu/ staff/gauch (veri?ed 20 September 2011) Gauch HG Jr, Rodrigues PC, Munkvold JD, Hefner EL, Sorrells M (2011) Two new strategies for detecting and understanding QTL 9 environment interactions. Crop Sci 51:96–113 Gerjets T, Schole?eld D, Foulkes MJ, Lenton JR, Holdsworth MJ (2010) An analysis of dormancy, ABA responsiveness, afterripening and pre-harvest sprouting in hexaploid wheat (Triticum aestivum L.) caryopses. J Exp Bot 61:597–607 Gupta PK, Rustgi S, Kulwal PL (2005) Linkage disequilibrium and association studies in higher plants: present status and future prospects. Plant Mol Biol 57:461–485 Hanocq E, Niarquin M, Heumez E, Rousset M, Le Gouis J (2004) Detection and mapping of QTL for earliness components in a bread wheat recombinant inbred lines population. Theor Appl Genet 110:106–115 He Q, Berg A, Li Y, Vallejos CE, Wu R (2010) Mapping genes for plant structure, development and evolution: functional mapping meets ontology. Trends Genet 26:39–46 Heun M, Kennedy AE, Anderson JA, Lapitan NLV, Sorrells ME, Tanksley SD (1991) Construction of a restriction fragment length polymorphism map for barley (Hordeum vulgare). Genome 34:437–447 Imtiaz M, Ogbonnaya FC, Oman J, Van Ginkel M (2008) Characterization of quantitative trait loci controlling genetic variation for preharvest sprouting in synthetic backcross derived wheat lines. Genetics 178:1725–1736 Ishikawa G, Yonemaru J, Saito M, Nakamura T (2007) PCR-based landmark unique gene (PLUG) markers effectively assign homoeologous wheat genes to A, B and D genomes. BMC Genomics 8:135 Kato K, Nakamura W, Tabiki T, Miura H, Sawada S (2001) Detection of loci controlling seed dormancy on group 4 chromosomes of wheat and comparative mapping with rice and barley genomes. Theor Appl Genet 102:980–985 Knox RE, Clarke FR, Clarke JM, Fox SL, DePauw RM, Singh AK (2012) Enhancing the identi?cation of genetic loci and transgressive segregants for preharvest sprouting resistance in a durum wheat population. Euphytica. doi:10.1007/s10681-011-0557-0 Kulwal PL, Singh R, Balyan HS, Gupta PK (2004) Genetic basis of pre-harvest sprouting tolerance using single-locus and two-locus QTL analyses in bread wheat. Funct Integr Genomics 4:94–101 Kulwal PL, Kumar N, Gaur A, Khurana P, Khurana JP, Tyagi AK, Balyan HS, Gupta PK (2005) Mapping of a major QTL for preharvest sprouting tolerance on chromosome 3A in bread wheat. Theor Appl Genet 111:1052–1059 Kulwal PL, Mir RR, Kumar S, Gupta PK (2010) QTL analysis and molecular breeding for seed dormancy and pre-harvest sprouting tolerance in bread wheat. J Plant Biol 37:59–74 Lehmensiek A, Eckermann PJ, Verbyla AP, Appels R, Sutherland MW, Daggard GE (2005) Curation of wheat maps to improve map accuracy and QTL detection. Aust J Agric Res 56:1347–1354 Liu S, Bai G, Cai S, Chen C (2011) Dissection of genetic components of preharvest sprouting resistance in white wheat. Mol Breed 27:511–523 Lohwasser U, Roder MS, Borner A (2005) QTL mapping of the domestication traits pre-harvest sprouting and dormancy in wheat. Euphytica 143:247–249 Mares DJ, Mrva K, Cheong J, Williams K, Watson B, Storlie E, Sutherland M, Zou Y (2005) A QTL located on chromosome 4A associated with dormancy in white- and red grained wheats of diverse origin. Theor Appl Genet 111:1357–1364 Miedaner T, Wurschum T, Maurer HP, Korzun V, Ebmeyer E, Reif JC (2011) Association mapping for Fusarium head blight resistance in European soft winter wheat. Mol Breed 28:647–655

other chromosomes in germplasm that is more diverse than a biparental population, but still representative of a breeding program. In addition, we have identi?ed another marker (gwm610) linked with this important QTL. This information will be useful for pyramiding important QTLs for PHS resistance in breeding programs along with the QTL on 2BS. This study also illustrates the importance of using elite material in association mapping for adaptive traits of complex inheritance such as PHS for effective use of QTL mapping results in the plant breeding programs.
Acknowledgments PLK acknowledges Department of Science and Technology, Govt. of India for the award of BOYSCAST Fellowship and also the Vice Chancellor MPAU, Rahuri. We also gratefully acknowledge Hatch project 149502, Wheat CAP; USDA-NIFA NRI CAP Grant No. 2005-05130), and the Triticeae Coordinated Agriculture Project (USDA-NIFA, NRI CAP No. 2011-68002-30029). Thanks are also due to three anonymous reviewers for very thorough and constructive suggestions.

References
Anderson JA, Sorrells ME, Tanksley SD (1993) RFLP analysis of genomic regions associated with resistance to preharvest sprouting in wheat. Crop Sci 33:453–459 Bailey PC, McKibbin RS, Lenton JR, Holdsworth MJ, Flintham JE, Gale MD (1999) Genetic map locations for orthologous Vp1 genes in wheat and rice. Theor Appl Genet 98:281–284 Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate:a practical and powerful approach to multiple testing. J R Stat Soc 57:289–300 Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES (2007) TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics 23:2633–2635 Breseghello F, Sorrells ME (2006a) Association analysis as a strategy for improvement of quantitative traits in plants. Crop Sci 46:1323–1330 Breseghello F, Sorrells ME (2006b) Association mapping of kernel size and milling quality in wheat (Triticum aestivum L.) cultivars. Genetics 172:1165–1177 Chen C-X, Cai S-B, Bai G-H (2008) A major QTL controlling seed dormancy and pre-harvest sprouting resistance on chromosome 4A in a Chinese wheat landrace. Mol Breed 21:351–358 Crossa J, Burgueno J, Dreisigacker S, Vargas M, Herrera-Foessel SA, Lillemo M, Singh RP, Trethowan R, Warburton M, Franco J, Reynolds M, Crouch JH, Ortiz R (2007) Association analysis of historical bread wheat germplasm using additive genetic covariance of relatives and population structure. Genetics 177:1889– 1913 Emebiri LC, Oliver JR, Mrva K, Mares D (2010) Association mapping of late maturity a-amylase (LMA) activity in a collection of synthetic hexaploid wheat. Mol Breed 26:39–49 Flint-Garcia SA, Thornsberry JM, Buckler ES IV (2003) Structure of linkage disequilibrium in plants. Annu Rev Plant Biol 54:357–374 Flintham JE, Adlam R, Bassoi M, Holdsworth M, Gale M (2002) Mapping genes for resistance to sprouting damage in wheat. Euphytica 126:39–45 Gauch HG (2007) MATMODEL version 3.0: open source software for AMMI and related analyses. Crop and Soil Science, Cornell

123

Theor Appl Genet (2012) 125:793–805 Mohan A, Kulwal PL, Singh S, Kumar V, Mir RR, Kumar J, Prasad M, Balyan HS, Gupta PK (2009) Genome-wide QTL analysis for pre-harvest sprouting tolerance in bread wheat. Euphytica 168:319–329 Moncada P, Martinez CP, Borrero J, Chatel M, Gauch H Jr, Guimaraes E, Tohme J, McCouch SR (2001) Quantitative trait loci for yield and yield components in an Oryza sativa 9 Oryza ru?pogon BC2F2 population evaluated in an upland environment. Theor Appl Genet 102:41–52 Mori M, Uchino N, Chono M, Kato K, Miura H (2005) Mapping QTLs for grain dormancy on wheat chromosome 3A and group 4 chromosomes, and their combined effect. Theor Appl Genet 110:1315–1323 Munkvold JD, Tanaka J, Benscher D, Sorrells ME (2009) Mapping quantitative trait loci for preharvest sprouting resistance in white wheat. Theor Appl Genet 119:1223–1235 Nakamura S, Toyoma T (2001) Isolation of a Vp1 homologue from wheat and analysis of its expression in embryos of dormant and non-dormant cultivars. J Exp Bot 52:875–876 Nakamura S, Abe F, Kawahigashi H, Nakazono K, Tagiri A, Matsumoto T, Utsugi S, Ogawa T, Handa H, Ishida H, o Mori M, Kawaura K, Ogihara Y, Miura H (2011) A wheat homolog of MOTHER OF FT AND TFL1 acts in the regulation of germination. Plant Cell 23:3215–3229 Neumann K, Kobiljski B, Dencic S, Varshney RK, Borner A (2011) Genome-wide association mapping: a case study in bread wheat (Triticum aestivum L.). Mol Breed 27:37–58 Nordborg M, Weigel D (2008) Next-generation genetics in plants. Nature 456:720–723 Ogbonnaya FC, Imtiaz M, Ye G, Hearnden PR, Hernandez E, Eastwood RF, Van Ginkel M, Shorter SC, Winchester JM (2008) Genetic and QTL analyses of seed dormancy and preharvest sprouting resistance in the wheat germplasm CN10955. Theor Appl Genet 116:891–902 Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D (2006) Principal components analysis corrects for strati?cation in genome-wide association studies. Nat Genet 38:904–909 Pritchard JK, Stephens M, Donnelly PJ (2000) Inference of population structure using multilocus genotype data. Genetics 155:945–959 Reif JC, Gowda M, Maurer HP, Longin CFH, Korzun V, Ebmeyer E, Bothe R, Pietsch C, Wurschum T (2011) Association mapping for quality traits in soft winter wheat. Theor Appl Genet 122:961–970 Roder MS, Korzun V, Wendehake K, Plaschke J, Tixier M-H, Leroy P, Ganal MW (1998) A microsatellite map of wheat. Genetics 149:2007–2023

805 Sears ER (1944) Cytogenetic studies with polyploidy species of wheat. II. Additional chromosome aberrations in Triticum vulgare. Genetics 29:232–246 Somers D, Isaac P, Edwards K (2004) A high-density microsatellite consensus map for bread wheat (Triticum aestivum L.). Theor Appl Genet 109:1105–1114 Sourdille P, Singh S, Cadalen T, Brown-Guedira GL, Gay G, Qi L, Gill BS, Dufour P, Murigneux A, Bernard M (2004) Microsatellite-based deletion bin system for the establishment of geneticphysical map relationships in wheat (Triticum aestivum L.). Funct Integr Genomics 4:12–25 Storey JD (2002) A direct approach to false discovery rates. J R Stat Soc Series B 64:479–498 Tan M-K, Sharp PJ, Lu M-Q, Howes N (2006) Genetics of grain dormancy in a white wheat. Aust J Agric Res 57:1157–1165 Torada A, Ikeguchi S, Koike M (2005) Mapping and validation of PCR-based markers associated with a major QTL for seed dormancy in wheat. Euphytica 143:251–255 Wu R, Lin M (2006) Functional mapping—how to map and study the genetic architecture of dynamic complex traits. Nat Rev Genet 7:229–237 Wu W-R, Li W-M, Tang D-Z, Lu H-R, Worland AJ (1999) Timerelated mapping of quantitative trait loci underlying tiller number in rice. Genetics 151:297–303 Yao J, Wang L, Liu L, Zhao C, Zheng Y (2009) Association mapping of agronomic traits on chromosome 2A of wheat. Genetica 137:67–75 Yu J, Buckler ES (2006) Genetic association mapping and genome organization of maize. Curr Opin Biotech 17:155–160 Yu J, Pressoir G, Briggs WH, Vroh Bi I, Yamasaki M, Doebley JF, McMullen MD, Gaut BS, Nielsen DM, Holland JB et al (2006) A uni?ed mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat Genet 38:203–208 Yu L-X, Liu S, Anderson JA, Singh RP, Jin Y, Dubcovsky J, BrownGuidera G, Bhavani S, Morgounov A, He Z, Huerta-Espino J, Sorrells ME (2010) Haplotype diversity of stem rust resistance loci in uncharacterized wheat lines. Mol Breed 26:667–680 Yu L-X, Lorenz A, Rutkoski J, Singh RP, Bhavani S, Huerta-Espino J, Sorrells ME (2011) Association mapping and gene–gene interaction for stem rust resistance in CIMMYT spring wheat germplasm. Theor Appl Genet 123:1257–1268 Zhu C, Gore M, Buckler ES, Yu J (2008) Status and prospects of association mapping in plants. Plant Genome 1:5–20

123



相关推荐


友情链接: