• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of geneticsGeneticsCurrent IssueInformation for AuthorsEditorial BoardSubscribeSubmit a Manuscript
Genetics. Jul 2009; 182(3): 851–861.
PMCID: PMC2710164

Unraveling the Complex Trait of Crop Yield With Quantitative Trait Loci Mapping in Brassica napus

Abstract

Yield is the most important and complex trait for the genetic improvement of crops. Although much research into the genetic basis of yield and yield-associated traits has been reported, in each such experiment the genetic architecture and determinants of yield have remained ambiguous. One of the most intractable problems is the interaction between genes and the environment. We identified 85 quantitative trait loci (QTL) for seed yield along with 785 QTL for eight yield-associated traits, from 10 natural environments and two related populations of rapeseed. A trait-by-trait meta-analysis revealed 401 consensus QTL, of which 82.5% were clustered and integrated into 111 pleiotropic unique QTL by meta-analysis, 47 of which were relevant for seed yield. The complexity of the genetic architecture of yield was demonstrated, illustrating the pleiotropy, synthesis, variability, and plasticity of yield QTL. The idea of estimating indicator QTL for yield QTL and identifying potential candidate genes for yield provides an advance in methodology for complex traits.

YIELD is the most important and complex trait in crops. It reflects the interaction of the environment with all growth and development processes that occur throughout the life cycle (Quarrie et al. 2006). Crop yield is directly and multiply determined by yield-component traits (such as seed weight and seed number). Yield-related traits (such as biomass, harvest index, plant architecture, adaptation, resistance to biotic and abiotic constraints) may also indirectly affect yield by affecting the yield-component traits or by other, unknown mechanisms. Increasing evidence suggests that “fine-mapped” quantitative trait loci (QTL) or genes identified as affecting crop yield involve diverse pathways, such as seed number (Ashikari et al. 2005; Tian et al. 2006b; Burstin et al. 2007; Xie et al. 2008; Xing et al. 2008; Xue et al. 2008), seed weight (Ishimaru 2003; Song et al. 2005; Shomura et al. 2008; Wang et al. 2008; Xie et al. 2006, 2008; Xing et al. 2008; Xue et al. 2008), flowering time (Cockram et al. 2007; Song et al. 2007; Xie et al. 2008; Xue et al. 2008), plant height (Salamini 2003; Ashikari et al. 2005; Xie et al. 2008; Xue et al. 2008), branching (Clark et al. 2006; Burstin et al. 2007; Xing et al. 2008), biomass yield (Quarrie et al. 2006; Burstin et al. 2007), resistance and tolerance to biotic and abiotic stresses (Khush 2001; Brown 2002; Yuan et al. 2002; Waller et al. 2005; Zhang 2007; Warrington et al. 2008), and root architecture (Hochholdinger et al. 2008).

Many experiments have explored the genetic basis of yield and yield-associated traits (yield components and yield-related traits) in crops. Summaries of identified QTL have been published for wheat (MacCaferri et al. 2008), barley (Von Korff et al. 2008), rice, and maize (http://www.gramene.org/). The results show several common patterns. First, QTL for yield and yield-associated traits tend to be clustered in the genome, which suggests that the QTL of the yield-associated traits have pleiotropic effects on yield. Second, this kind of pleiotropy has not been well analyzed genetically. The QTL for yield (complicated factor), therefore, have not been associated with any yield-associated traits (relatively simple factors, such as plant height). Therefore, they are unlikely to predict accurately potential candidate genes for yield. Third, only a few loci (rarely >10) have been found for each of these traits. Thus, the genetic architecture of yield has remained ambiguous. Fourth, trials were carried out in a few environments and how the mode of expression of QTL for these complex traits might respond in different environments is unclear.

In this study, the genetic architecture of crop yield was analyzed through the QTL mapping of seed yield and eight yield-associated traits in two related populations of rapeseed (Brassica napus) that were grown in 10 natural environments. The complexity of the genetic architecture of seed yield was demonstrated by QTL meta-analysis. The idea of estimating indicator QTL (QTL of yield-associated traits, which are defined as the potential genetic determinants of the colocalized QTL for yield) for yield QTL in conjunction with the identification of candidate genes is described.

MATERIALS AND METHODS

Plant material, field experiments, and trait measurements:

A population of doubled haploids (DHs) of 202 lines was developed by microspore culture from the F1 cross between Tapidor (a European winter cultivar) and Ningyou7 (a Chinese semi-winter cultivar) and named TNDH (Qiu et al. 2006). A reconstructed F2 (RC-F2) population was made from 101 crosses per round, between pairs of DH lines randomly chosen from the 202 lines of the TNDH population. In 2004 and 2005, three and four rounds of crossing by hand emasculation and pollination resulted in 303 and 404 crosses, respectively.

The two populations and both parents were grown in five locations in China over 1, 3, or 4 years, for 10- and 3-year–location combinations for the DH and RC-F2 populations, respectively. Year–location combinations were treated as microenvironments. Microenvironment–population combinations were treated as experiments. These microenvironments were split into two contrasting macroenvironments (winter and semi-winter) and three agroecological areas (the Loess Plateau, the middle valley of the Yangtze River, and the lower valley of the Yangtze River) (supporting information, Table S1). Details of the climatic conditions during the growing season for each of the 10 microenvironments are described in Figure S1. The planting followed a randomized complete-block design with three replicates. Each plot was 3.0 m2 with 30 plants in all microenvironments (except S5, which was 4.0 m2 with 40 plants), with 40 cm between rows and 25 cm between individual plants. The seeds were sown by hand and the field management followed standard agricultural practice. In each plot, 9 representative individuals from the middle of each row in the S3, S4, N3, and N4 microenvironments, or 12 in the S5, S6, and N6 microenvironments, or all individuals in the S7, N7, and E7 microenvironments were harvested by hand from ground level at maturity.

Seed yield and eight yield-associated traits were investigated. Flowering time was measured as the interval between the date of sowing and the date when the first flowers emerged on 50% of the plants in a plot. Maturity time was measured as the interval between the date of sowing and the date when pods on most of the plants in a plot were yellow. Plant height was the height of each harvested individual in a plot, measured from the base of the stem to the tip of the main shoot. Branch number was the number of branches arising from the main shoot of each harvested individual in a plot. Pod number was the number of well-filled, normally developed pods on each harvested individual in a plot. Seed number per pod was the average number of well-filled seeds from 100 well-developed pods, sampled from the primary branch in the middle of the harvested individuals in a plot. Seed weight was the average dry weight of 1000 well-filled seeds from three replicate samples, from the mixed seeds of the harvested individuals in a plot. Biomass yield per plant was measured as the average total above-ground (except the seeds) dry weight of the harvested individuals in a plot. Seed yield per plant was measured as the average dry weight of seeds of the harvested individuals in a plot.

Construction of the linkage map and alignment with the genome of Arabidopsis:

A total of 786 markers were mapped to the new linkage map generated with the TNDH population using JoinMap 3.0 (http://www.kyazma.nl/index.php/mc.JoinMap). This covered 19 chromosomes identified as A1–A10 and C1–C9, with an average distance of 2.7 cM between markers (Table S2 A). The threshold for goodness of fit was set to ≤5.0 with logarithm of the odds ratio (LOD) scores >1.0 and a recombination frequency <0.4. The order of the markers on the linkage map agreed well with our published maps (Qiu et al. 2006; Long et al. 2007). The genotype of each RC-F2 line was deduced from the corresponding genotype of their parents.

Two hundred and seventy-seven markers with known sequence information were employed as anchored markers for the map alignment between B. napus and Arabidopsis thaliana, on the basis of comparative mapping reported previously (Parkin et al. 2005). Details of the alignment are descried in Long et al. (2007). Several syntenic blocks or insertion fragments (islands) were identified between the Arabidopsis genome and the TNDH linkage map (Table S2 B).

A total of 425 genes of Arabidopsis with known functions relating to flowering time, plant height, branch number, and other traits investigated in this study were collected from the TAIR website (http://www.arabidopsis.org/) and published articles (Table S2 C). These genes were located on each syntenic block according to their physical positions in the genome of Arabidopsis. The positions of putative genes were aligned to the TNDH linkage map according to their closest anchored markers in the same syntenic block (Table S2 D).

Statistical analysis, QTL mapping, and meta-analysis:

Environment was treated as a fixed effect, while genotype (DH or RC-F2 lines) was treated as a random effect. The broad-sense heritability was calculated as: An external file that holds a picture, illustration, etc.
Object name is GEN1823851inf1.jpg, where An external file that holds a picture, illustration, etc.
Object name is GEN1823851inf2.jpg is the genotypic variance, An external file that holds a picture, illustration, etc.
Object name is GEN1823851inf3.jpg is the interaction variance of genotype with environment, An external file that holds a picture, illustration, etc.
Object name is GEN1823851inf4.jpg was the error variance, n was the number of environments, and r was the number of replicates. The genetic correlation was calculated as: An external file that holds a picture, illustration, etc.
Object name is GEN1823851inf5.jpg, where covGxy, An external file that holds a picture, illustration, etc.
Object name is GEN1823851inf6.jpg, and An external file that holds a picture, illustration, etc.
Object name is GEN1823851inf7.jpg were the genotypic covariance and variance of the pairwise traits, respectively. The significance of each genetic correlation was determined using a t-test of the correlation coefficients (Kong 2005). Estimates of components of variance and covariance were obtained using the SAS general linear model (GLM) procedure (Sas Institute 2000).

QTL were detected by composite interval mapping (Zeng 1994) using WinQTL cartographer 2.5 software (http://statgen.ncsu.edu/qtlcart/WQTLCart.htm). The number of control markers, window size, and walking speed were set to 5, 10 cM, and 1 cM, respectively. The default genetic distance (5 cM) was used to define a QTL in an experiment. The experimentwise LOD threshold was determined by permutation analysis (Churchill and Doerge 1994) with 1000 repetitions. LOD scores corresponding to P = 0.05 (2.5–3.2 for DH and 3.8–4.4 for the RC-F2 population) were used for identifying significant QTL. To avoid missing QTL with very small effects, a lower LOD score corresponding to P = 0.50 (2.1–2.3 for the DH population and 2.9–3.1 for the RC-F2 population) was adopted for the presence of suggestive QTL. The overlapping suggestive QTL and all significant QTL were admitted (Long et al. 2007) and named as identified QTL.

Since QTL of the same or related traits, which were detected in different experiments and mapped to the same region of a chromosome, might be several estimates of the position of a single QTL, algorithms for QTL meta-analysis were used to estimate the number and positions of the meta-QTL underlying the analyzed QTL (Goffinet and Gerber 2000). This approach provides a modified Akaike criterion that can be used to determine the number of meta-QTL that best fitted the results on a given linkage group. It also groups the QTL that were detected in the different experiments into classes that correspond to the same QTL and provides a consensus estimate of QTL positions. Computations were conducted using the BioMercator 2.1 software (Arcade et al. 2004). At present, the method used in this software cannot distinguish between models with more than four meta-QTL on the same linkage group. If the estimated number of meta-QTL is more than four, Biomercator 2.1 declares that the most probable model is one with a number of meta-QTL that is equal to the number of analyzed QTL. The delete function of the software was used to select specific segments of a linkage group separated by regions with no QTL, and to apply QTL meta-analysis to these segments separately. The software also provides a method for calculating 95% confidence intervals for the meta-QTL,

equation M1

where equation M2 is the variance of position of the QTLi, and k is the total number of QTL integrated into the meta-QTL.

A two-round strategy of QTL meta-analysis was adopted. The QTL identified in different experiments were first integrated into consensus QTL, trait by trait (Figure 1A). Two types of consensus QTL were defined: major QTL (those occurring at least once with R2 ≥20% or at least twice with R2 ≥ 10%), and minor QTL (the remainder with relatively small effect) (Price 2006; MacCaferri et al. 2008). In the second round of QTL meta-analysis, the consensus QTL for the different traits were integrated into unique QTL (Figure 1B).

Figure 1.
Demonstration of the process and result of two rounds of QTL meta-analysis for QTL mapped on 30- to 60-cM segments of A2 linkage group. (A) Thirteen identified QTL for seed yield detected from five experiments were integrated into four consensus QTL (which ...

For each unique QTL in which a consensus QTL for seed yield was integrated, one or more consensus QTL of other yield-associated traits were chosen as indicator QTL, which were defined as potential genetic determinants of the colocalized QTL for seed yield. The criteria for selecting indicator QTL were the magnitude of the LOD scores, the number of experiments in which they were detected, the overlapping of the confidence interval, and the number of common environments within which the two QTL (i.e., the QTL for yield-associated traits and those for seed yield) were identified.

RESULTS

Phenotypic variation and correlation among traits and across environments:

The two parents (Tapidor and Ningyou7) differed significantly for most trait–environment combinations (Table S3). Ningyou7, a semi-winter cultivar, had larger seeds and flowered and matured earlier than the winter cultivar Tapidor, in all microenvironments. However, for the other six traits, the performance of the two parents in the two macroenvironments (winter and semi-winter) was reversed, such that the means in the two macroenvironments were similar (Table 1). This reflects the adaptability of the two cultivars to the different macroenvironments (e.g., the climatic factors, Figure S1). Normal or near-normal distribution (data not shown) with wide transgressive segregation was observed in both populations (DH and RC-F2), for all traits in all microenvironments (Table S3), which indicated a quantitative genetic control.

TABLE 1
Phenotypic variation of the two parents for seed yield and eight yield-associated traits in two macroenvironments

Most pairwise genetic correlations between the two populations within the same microenvironment differed little in degree (Table S4 A–J). In contrast, pairwise genetic correlations in different microenvironments differed considerably (mostly in degree, a few in direction), which suggested that genetic correlation depended strongly on the environment. Genetic correlations among traits were also calculated across environments for both populations (Table 2). In general, seed yield was correlated with all investigated traits in both populations: negatively for flowering and maturity times, and positively for the other six traits. As expected, all pairwise combinations among the three yield-component traits (pod number, seed number, and seed weight) were significantly negatively correlated, which suggested competition among the sinks for assimilates. Significant correlations were also observed for most pairwise combinations among the five yield-related traits, and for half of the pairwise combinations between the five yield-related and three yield-component traits. Seed yield, among all the investigated traits, notably showed the highest correlation with the other eight traits, and on average, showed the highest coefficients of determination (mean r2 = 21 and 22% for DH and RC-F2 populations, respectively).

TABLE 2
Genetic correlations and broad-sense heritability in DH (below diagonal) and RC-F2 (above diagonal) populations

The broad-sense heritability of the DH population was slightly higher than that of the RC-F2 population for all investigated traits, except biomass yield, and varied from 60% and 55% (seed yield) to 93% and 89% (flowering time), respectively (Table 2).

Genomewide detection, meta-analysis, and expression mode of QTL from multiple environments:

An analysis of variance indicated that, despite strong genotypic and environmental effects for all investigated traits, there were significant genotype × environment interactions (Table S5). Consequently, each environment and population was analyzed separately for QTL. A total of 1022 QTL (614 significant QTL and 408 suggestive QTL) were detected for the nine traits in both populations in the 10 microenvironments. After deleting 152 nonoverlapping suggestive QTL (Table S6 A), a total of 870 identified QTL were identified for the nine traits, of which nearly 10% (85) were for seed yield (Table 3). The identified QTL can explain 2.2–33.4% of the phenotypic variance, although the majority (84%) showed only moderate effects, with R2 < 10% (Table S6 A). Most of the Tapidor alleles in the identified QTL increased flowering and maturity times but decreased seed weight, which accorded well with the performance of the two parents for the traits (Table 3).

TABLE 3
Overview of identified QTL and consensus QTL for seed yield and eight yield-associated traits

Where the confidence intervals of the identified QTL for each trait in different experiments overlapped (Table S6 B), meta-analysis integrated these identified QTL into single consensus QTL trait by trait (Figure 1A). As a result, the 694 overlapping identified QTL were integrated into 225 reproducible consensus QTL with reduced confidence intervals, from an average of 6.5 to 3.7 cM (Table 3). The proportion of the reproducible consensus QTL was highest (80.7% = 46/57) for flowering time and lowest (36.4% = 20/55) for seed yield, a result that reflected the heritability of these traits.

Nearly half of the 401 consensus QTL for the nine traits, particularly those for seed yield, were specifically detected in 1 of the 10 microenvironments, and no consensus QTL, except 2 for flowering time, appeared consistently in >8 microenvironments (Figure 2A). It is interesting that the proportions of the consensus QTL for the three traits (flowering time, maturity time, and seed weight) specific to the macroenvironments were all lower than for the other six traits. This was in accordance with the variation in parental performance for these traits in the two macroenvironments. Two-thirds of the consensus QTL were detected in either the winter or semi-winter macroenvironment, and only one-third of the consensus QTL appeared in both. This strongly suggested that more than half of the QTL were expressed principally in response to the specific environment in which the population was grown. For example, the three major QTL clustered on the C7 linkage group (qSY.C7-2, qSY.C7-3, and qSY.C7-4) were all specific to the semi-winter macroenvironment. They were strongly expressed in Hangzhou (in the E7 microenvironment; LOD scores >10; R2 > 20%) and two of them were weakly expressed in Wuhan (in the S4 and S7 microenvironments, respectively; LOD scores <3; R2 <5%) (Figure 2B and Table S6 B). Wuhan and Hangzhou are in the agroecological area of the middle and lower valley, respectively, of the Yangtze River, both of which are semi-winter macroenvironments in China (Table S1). The favorable alleles come from the semi-winter parent Ningyou7 (which was bred near Hangzhou), which indicated that these major QTL played a key role in regulating the adaptability of Ningyou7 to the corresponding semi-winter agroecological area.

Figure 2.
Expression response of 401 consensus QTL in natural environments. (A, top) Number of consensus QTL appeared in 1 to 10 (all) microenvironments. (A, bottom) Number of consensus QTL appeared in winter, semi-winter, or both macroenvironments. Most consensus ...

Dissection of pleiotropic unique QTL and identification of candidate genes responsible for seed yield:

Most of the consensus QTL determined for each trait overlapped with those determined for other traits, as shown on linkage groups A1 and C6 (Figure 3 and Figure S2). For example, a consensus QTL for seed yield (qSY.A2-2), detected in several microenvironments, colocalized precisely with three consensus QTL for flowering time, maturity time, and plant height (qFT.A2-4, qPH.A2-2, and qMT.A2-2, respectively) in a small region of the A2 linkage group (Figure 4). These consensus QTL were therefore subjected to the second round of meta-analysis (Figure 1B), which resulted in the integration of 329 overlapping consensus QTL into 111 pleiotropic unique QTL (Table 3) with reduced average putative confidence intervals of 2.5 cM (Table S6 C).

Figure 3.
Distribution of 401 consensus QTL for seed yield and eight yield-associated traits on 19 linkage groups. The abscissa represents the 19 linkage groups that were divided into 10-cM bins and the ordinate represents the number of QTL. The long dotted line ...
Figure 4.
Dissection of the overlapping QTL cluster for seed yield and yield-associated traits in a 7.2-cM region on linkage group A2. Identified QTL are indicated as LOD curves. Confidence intervals (CI) for each identified QTL are shown as horizontal lines under ...

The 55 consensus QTL for seed yield were classified into two types: (i) eight nonoverlapping QTL and (ii) 47 overlapping QTL, which were integrated into pleiotropic unique QTL with shorter average confidence intervals of 2.5 cM (Table S6 C). Confidence intervals for the new QTL were reduced by 2.5 and 1.8 times, compared with the average confidence intervals of the 85 identified QTL and 55 consensus QTL for seed yield (6.2 cM and 4.5 cM, respectively; Table S6 B). The two rounds of meta-analysis resulted in a more precise estimate of the QTL for seed yield. One to four consensus QTL of the other eight yield-associated traits were chosen as indicator QTL, which potentially controlled seed yield for each of the 47 pleiotropic unique QTL (Figure S2 and Table S6 C). The total of 63 indicator QTL could be associated with any of the eight yield-associated traits, but the highest proportion was associated with plant height and the lowest proportion with maturity time (Table 4).

TABLE 4
The character and indirect evidence that indicator QTL influence seed yield

Indirect evidence that indicator QTL influenced seed yield came from the extent to which the confidence intervals of the indicator QTL overlapped those of consensus QTL for seed yield, from an average of 71.9% to 92.5% (total mean = 80.9%) for different traits (Table 4). In addition, the average distance between peak positions of the indicator QTL and consensus QTL for seed yield was small: 1.1 cM on average. Furthermore, in some cases, the putative change in the value of seed yield estimated from the effect of indicator QTL was close to the additive effect of the corresponding QTL for seed yield. For example, the additive effect (a = 172 kg/ha) of the QTL for seed yield, qSY.A1-4, was about half that (a = 328 kg/ha) of its indicator QTL, qBY.A1-2 (a QTL for biomass yield), which was involved in the pleiotropic unique QTL, uq.A1-9 (Table S6 C). This ratio (172/328 = 52.1%) was close to the general ratio (57.1%; J. Shi, unpublished data) of seed yield to biomass yield in the corresponding experiment N6R. More importantly, the additive effect direction of the QTL for seed yield is generally opposite (−) to those of the corresponding indicator QTL for the two traits (flowering and maturity times), but mostly the same (+) for the other six traits. This is in accordance with the signs of the genetic-correlation coefficients of seed yield with the above eight yield-associated traits (Table 2).

On the basis of comparative mapping with the genome of Arabidopsis, candidate genes were located in the confidence interval of the reproducible (for the higher stringency) consensus QTL. Some candidate genes underlay the indicator QTL that potentially control seed yield (Table S6 C) and thus are potential candidate genes for seed yield. For example, two genes for flowering time, ZTL and VIN3, were located in the confidence interval of the QTL for flowering time, qFT.A2-4, which had been chosen as the indicator QTL of the QTL for seed yield, qSY.A2-2, and were thus potential candidate genes for seed yield (Figure 4).

DISCUSSION

Although much research into the genetic basis of yield and yield-associated traits has been reported, the genetic architecture of yield has remained ambiguous because few QTL have been identified in each such experiment. Here, we report 870 identified QTL: 85 for seed yield and 785 for eight yield-associated traits. These QTL were detected on a single genetic map of B. napus in two related populations grown in 10 natural environments. Three key factors for successful screening for QTL with high resolution are a high-density genetic map to separate multiple QTL (Jiang and Zeng 1995), a large population, and replicated experiments in multiple environments to assess genotype × environment interactions in the expression of QTL. The analysis of large numbers of QTL, however, is problematic.

Meta-analysis originally integrated QTL of the same and/or related traits from different populations on the basis of an integrated map (Arcade et al. 2004). Due to variability in order and genetic distance of the common markers among the different linkage maps, the accuracy of the integrated map and estimate of the number and position of meta-QTL were equivocal (Veyrieras et al. 2007). We adopted meta-analysis to combine QTL detected in multiple environments and related populations for multiple yield-associated traits, however, based on a single linkage map. This increased the accuracy of the estimated positions of the meta-QTL and facilitated the dissection of the genetic architecture of yield.

Nearly 80% of the 401 consensus QTL were from only one or two specific microenvironments, and no consensus QTL, except one for flowering time (qFT.C6-4), appeared in all microenvironments (Figure 2A). This QTL has not been detected in two spring-cropped microenvironments with the same population (Long et al. 2007). This suggests that few QTL are likely to be expressed universally. The high proportion of environment-specific QTL indicates the large impact the natural environment has on the genes underlying seed yield and yield-associated traits.

One hundred and twenty-nine of the 225 reproducible consensus QTL were found in the two macroenvironments (Table S6 B). These QTL are likely to be important targets for marker-assisted selection and the development of varieties with wide adaptability. Another 96 reproducible consensus QTL were only detected in either the winter or semi-winter macroenvironment. These QTL are also likely to be important targets for the development of varieties that have special adaptability.

Thirteen of the 15 major QTL were repeatedly found 2–10 times (mean = 7) in the different experiments (Table S6 B). Fine mapping these QTL and validating the potential candidate genes is a reliable and feasible strategy for QTL cloning. The clusters of major QTL for flowering time on linkage group C6 have been dissected successfully into several sub-QTL through substitution mapping (J. Wang, unpublished data). More than half of the 386 minor QTL were also repeatedly identified 2–10 times. Cloning such QTL via fine mapping may be unlikely, but the application of bioinformatics (such as meta-analysis, comparative genomics, and haplotype association mapping) and/or transcriptomics to narrow QTL/QTG (quantitative trait genes) without fine mapping seems promising (Price 2006; Burgess-Herbert et al. 2008; Norton et al. 2008). Recently, four minor QTL for disease resistance in rice, each explaining <5% of the phenotypic variation, have been isolated successfully by a strategy involving candidate genes that integrated expression profiling, bioinformatics, and functional complementation analysis (Hu et al. 2008). In other words, a minor QTL might be a major QTL if the variability of the parental alleles is sufficiently large and/or the environment allows the induction of differential expression of the two alleles. Therefore, the minor QTL that were repeatedly detected also provide an important genetic resource for marker-assisted selection, cloning of QTL, and the functional analysis of genes.

Another important feature of the consensus QTL was the high proportion (82% = 329/401) that colocalized at the genomic level. Forty-seven of the 55 QTL for seed yield colocalized with other QTL. Additionally, six of eight independent QTL for seed yield overlapped other QTL for traits not published in this article, i.e., glucosinolate content in seeds (J. Feng, unpublished data) and resistance to Sclerotinia sclerotiorum (C. Jiang, unpublished data). These colocalizations indicate that the 55 QTL for seed yield are all dependent on the QTL of other yield-associated traits. This is not surprising, given that yield is determined by the cumulative effect of many processes/traits of growth and development, from sowing to harvest, when the conventional yield-associated traits become dominant. Most functional genes in the genome, therefore, might contribute directly or indirectly to the trait of yield (Slafer 2003). When a locus controlling a trait for a component of yield (e.g., seed weight) or a yield-related trait (e.g., plant height) shows a phenotypic difference resulting from allelic variation between the parents at that locus, a detected QTL for yield might be the result of gene pleiotropy (Li et al. 1997). In other words, QTL for yield may result from pleiotropic QTL: genomic regions that affect multiple traits by containing multiple, tightly linked, trait-specific genes or genes that affect multiple traits (Hall et al. 2006). Most published fine-mapped QTL and the genes identified as affecting yield exhibit pleiotropic effects on at least one trait (Peng et al. 1999; Brown 2002; Yuan et al. 2002; Li et al. 2004; Ashikari et al. 2005; Waller et al. 2005; Clark et al. 2006; Tian et al. 2006a; Cockram et al. 2007) or multiple yield-associated traits (Quarrie et al. 2006; Xie et al. 2006, 2008; Burstin et al. 2007; Song et al. 2007; Shomura et al. 2008; Xue et al. 2008). Even a short region with a confidence interval of 0.4 cM and a physical distance of 37.4 kb has revealed pleiotropic effects between seven QTL of different yield-associated traits and the QTL of grain yield in rice (Xie et al. 2008).

Seed yield showed the highest correlation with the other eight traits (mean r2 = 21 and 22% for DH and RC-F2 populations, respectively), and a high percentage (85%) of QTL for seed yield colocalized with QTL for other yield-associated traits. These findings indicate that the QTL of yield-associated traits are potential contributors to the colocalized QTL for seed yield. On average, a QTL for seed yield involved 2.5 QTL of yield-associated traits, and at least one QTL for all eight yield-associated traits involved the QTL for seed yield, qSY.A1-4 (Table S6 C). To estimate which QTL of yield-associated trait(s) were more likely to have pleiotropic effects on seed yield at a particular locus, the idea of indicator QTL was proposed to identify the probable genetic determinant of the colocalized QTL for seed yield. These indicator QTL are determined by their larger LOD scores, reproducibility, overlapping confidence intervals, and presence in common environment(s) with the QTL for seed yield. There is indirect evidence that supports the idea that pleiotropy is likely to be the genetic cause of the colocalization of indicator QTL and seed yield QTL: (1) the peak positions of the two kinds of QTL are very close to each other; (2) the proportion of the overlapping confidence intervals of the two kinds of QTL is very high; (3) the additive-effect directions of the seed yield QTL are generally opposite (−) to that of indicator QTL for two traits (flowering time and maturity time) while mostly the same (+) for the other six traits, which accorded well with the signs of the genetic-correlation coefficients of seed yield with the above eight yield-associated traits; (4) in many cases, the additive effect of the seed yield QTL was very near to the putative change in the value of seed yield that was estimated from the additive effect of the indicator QTL. Indicator QTL, from easily measured traits (such as flowering time and seed weight), usually had more stable expression, higher LOD scores, larger R2 values, and identifiable candidate genes than the colocalized QTL for seed yield (Table S6 C). Indicator QTL will thus facilitate the cloning of these QTL for seed yield, assuming that pleiotropy and not tight linkage is the cause of the colocalization of the indicator QTL with the corresponding QTL for seed yield (MacCaferri et al. 2008). For example, the cloning of qSY.A2-2, a major QTL for seed yield, should be facilitated using its indicator QTL qFT.A2-4. This QTL is a major QTL for flowering time with a large genetic effect, stable expression in all six semi-winter microenvironments, and available candidate genes (Figure 4). Flowering time is also measured more easily and precisely than seed yield. The successful cloning of QTL for yield in rice (Xue et al. 2008) and wheat (Quarrie et al. 2006) was facilitated by indicator QTL: a QTL for flowering time and a QTL for biomass yield, respectively.

Twenty-nine of 47 pleiotropic QTL for seed yield colocalized with more than two consensus QTL. This indicated that, in addition to pleiotropy, the effect of the QTL for seed yield could be a synthetic effect of several underlying tightly linked QTL of different yield-associated traits. A previously mapped QTL may be dissected into several sub-QTL after subsequent fine mapping (Ashikari et al. 2005; Thomson et al. 2006; Christians and Senger 2007). Multiple indicator QTL were chosen for 14 pleiotropic QTL for seed yield, half of which involved QTL for seed yield in different environments. For example, qSY.A2-4 colocalized two indicator QTL (qPH.A2-4 and qSN.A2-1) in two (S4 and S6) and one (N6) microenvironments, respectively (Table S6 C). Multiple indicator QTL from different environments also revealed that environmental conditions contribute strongly to the variability and plasticity of QTL for yield.

In conclusion, tens of QTL for seed yield and hundreds of QTL for yield-associated traits were identified from multiple environments and populations using a single genetic map in B. napus. The complexity of the genetic architecture of yield was demonstrated by meta-analysis, illustrating the pleiotropy, synthesis, variability, and plasticity of yield QTL. The assignment of indicator QTL enabled the QTL for seed yield to be associated with specific yield-associated trait(s) in specific environment(s). In addition, in silico comparative mapping aligned many orthologs from the genome of Arabidopsis into particular regions of the reproducible QTL with tighter confidence intervals. Yield, the most complex trait in crops, was unraveled into tens of unique QTL indicated by one or multiple indicator QTL, with potential underlying candidate genes. The idea of estimating indicator QTL for yield QTL and identifying potential candidate genes for yield provides an advance in methodology for complex traits.

Acknowledgments

The authors thank Eddie Arthur for his critical reading of the manuscript. The authors also thank Dianrong Li and Hao Wang (Hybrid Rapeseed Research Center of Shaanxi, Dali 715105, China) for the field work and collecting the phenotypic data. Financial support for this work was provided by the National Basic Research and Development Program (2006CB101600).

Notes

Supporting information is available online at http://www.genetics.org/cgi/content/full/genetics.109.101642/DC1.

References

  • Arcade, A., A. Labourdette, M. Falque, B. Mangin, F. Chardon et al., 2004. BioMercator: integrating genetic maps and QTL towards discovery of candidate genes. Bioinformatics 20 2324–2326. [PubMed]
  • Ashikari, M., H. Sakakibara, S. Lin, T. Yamamoto, T. Takashi et al., 2005. Cytokinin oxidase regulates rice grain production. Science 309 741–745. [PubMed]
  • Brown, J. K., 2002. Yield penalties of disease resistance in crops. Curr. Opin. Plant Biol. 5 339–344. [PubMed]
  • Burgess-Herbert, S. L., A. Cox, S. W. Tsaih and B. Paigen, 2008. Practical applications of the bioinformatics toolbox for narrowing quantitative trait loci. Genetics 180 2227–2235. [PMC free article] [PubMed]
  • Burstin, J., P. Marget, M. Huart, A. Moessner, B. Mangin et al., 2007. Developmental genes have pleiotropic effects on plant morphology and source capacity, eventually impacting on seed protein content and productivity in pea. Plant Physiol. 144 768–781. [PMC free article] [PubMed]
  • Christians, J. K., and L. K. Senger, 2007. Fine mapping dissects pleiotropic growth quantitative trait locus into linked loci. Mamm. Genome 18 240–245. [PubMed]
  • Churchill, G. A., and R. W. Doerge, 1994. Empirical threshold values for quantitative trait mapping. Genetics 138 963–971. [PMC free article] [PubMed]
  • Clark, R. M., T. N. Wagler, P. Quijada and J. Doebley, 2006. A distant upstream enhancer at the maize domestication gene tb1 has pleiotropic effects on plant and inflorescent architecture. Nat. Genet. 38 594–597. [PubMed]
  • Cockram, J., H. Jones, F. J. Leigh, D. O'Sullivan, W. Powell et al., 2007. Control of flowering time in temperate cereals: genes, domestication, and sustainable productivity. J. Exp. Bot. 58 1231–1244. [PubMed]
  • Goffinet, B., and S. Gerber, 2000. Quantitative trait loci: a meta-analysis. Genetics 155 463–473. [PMC free article] [PubMed]
  • Hall, M. C., C. J. Basten and J. H. Willis, 2006. Pleiotropic quantitative trait loci contribute to population divergence in traits associated with life-history variation in Mimulus guttatus. Genetics 172 1829–1844. [PMC free article] [PubMed]
  • Hochholdinger, F., T. J. Wen, R. Zimmermann, P. Chimot-Marolle, O. da Costa e Silva et al., 2008. The maize (Zea mays L.) roothairless 3 gene encodes a putative GPI-anchored, monocot-specific, COBRA-like protein that significantly affects grain yield. Plant J. 54 888–898. [PMC free article] [PubMed]
  • Hu, K. M., D. Y. Qiu, X. L. Shen, X. H. Li and S. P. Wang, 2008. Isolation and manipulation of quantitative trait loci for disease resistance in rice using a candidate gene approach. Mol. Plant 1 786–793. [PubMed]
  • Ishimaru, K., 2003. Identification of a locus increasing rice yield and physiological analysis of its function. Plant Physiol. 133 1083–1090. [PMC free article] [PubMed]
  • Jiang, C., and Z. B. Zeng, 1995. Multiple trait analysis of genetic mapping for quantitative trait loci. Genetics 140 1111–1127. [PMC free article] [PubMed]
  • Khush, G. S., 2001. Green revolution: the way forward. Nat. Rev. Genet. 2 815–822. [PubMed]
  • Kong, F., 2005. Quantitaive Genetics in Plants. China Agricultural University Press, Beijing.
  • Li, J., M. Thomson and S. R. McCouch, 2004. Fine mapping of a grain-weight quantitative trait locus in the pericentromeric region of rice chromosome 3. Genetics 168 2187–2195. [PMC free article] [PubMed]
  • Li, Z., S. R. Pinson, W. D. Park, A. H. Paterson and J. W. Stansel, 1997. Epistasis for three grain yield components in rice (Oryza sativa L.). Genetics 145 453–465. [PMC free article] [PubMed]
  • Long, Y., J. Shi, D. Qiu, R. Li, C. Zhang et al., 2007. Flowering time quantitative trait loci analysis of oilseed brassica in multiple environments and genomewide alignment with Arabidopsis. Genetics 177 2433–2444. [PMC free article] [PubMed]
  • MacCaferri, M., M. C. Sanguineti, S. Corneti, J. L. Ortega, M. B. Salem et al., 2008. Quantitative trait loci for grain yield and adaptation of durum wheat (Triticum durum Desf.) across a wide range of water availability. Genetics 178 489–511. [PMC free article] [PubMed]
  • Norton, G. J., M. J. Aitkenhead, F. S. Khowaja, W. R. Whalley and A. H. Price, 2008. A bioinformatic and transcriptomic approach to identifying positional candidate genes without fine mapping: an example using rice root-growth QTLs. Genomics 92 344–352. [PubMed]
  • Parkin, I. A., S. M. Gulden, A. G. Sharpe, L. Lukens, M. Trick et al., 2005. Segmental structure of the Brassica napus genome based on comparative analysis with Arabidopsis thaliana. Genetics 171 765–781. [PMC free article] [PubMed]
  • Peng, J., D. E. Richards, N. M. Hartley, G. P. Murphy, K. M. Devos et al., 1999. ‘Green revolution’ genes encode mutant gibberellin response modulators. Nature 400 256–261. [PubMed]
  • Price, A. H., 2006. Believe it or not, QTLs are accurate! Trends Plant Sci. 11 213–216. [PubMed]
  • Qiu, D., C. Morgan, J. Shi, Y. Long, J. Liu et al., 2006. A comparative linkage map of oilseed rape and its use for QTL analysis of seed oil and erucic acid content. Theor. Appl. Genet. 114 67–80. [PubMed]
  • Quarrie, S., S. Pekic Quarrie, R. Radosevic, D. Rancic, A. Kaminska et al., 2006. Dissecting a wheat QTL for yield present in a range of environments: from the QTL to candidate genes. J. Exp. Bot. 57 2627–2637. [PubMed]
  • Salamini, F., 2003. Hormones and the green revolution. Science 302 71–72. [PubMed]
  • Sas Institute, 2000. SAS/STAT User's Guide, Version 8. SAS Institute, Cary, NC.
  • Shomura, A., T. Izawa, K. Ebana, T. Ebitani, H. Kanegae et al., 2008. Deletion in a gene associated with grain size increased yields during rice domestication. Nat. Genet. 40 1023–1028. [PubMed]
  • Slafer, G. A., 2003. Genetic basis of yield as viewed from a crop physiologist's perspective. Ann. Appl. Biol. 142 117–128.
  • Song, B. K., K. Nadarajah, M. N. Romanov and W. Ratnam, 2005. Cross-species bacterial artificial chromosome (BAC) library screening via overgo-based hybridization and BAC-contig mapping of a yield enhancement quantitative trait locus (QTL) yld1.1 in the Malaysian wild rice Oryza rufipogon. Cell. Mol. Biol. Lett. 10 425–437. [PubMed]
  • Song, X. J., W. Huang, M. Shi, M. Z. Zhu and H. X. Lin, 2007. A QTL for rice grain width and weight encodes a previously unknown RING-type E3 ubiquitin ligase. Nat. Genet. 39 623–630. [PubMed]
  • Thomson, M. J., J. D. Edwards, E. M. Septiningsih, S. E. Harrington and S. R. McCouch, 2006. Substitution mapping of dth1.1, a flowering-time quantitative trait locus (QTL) associated with transgressive variation in rice, reveals multiple sub-QTL. Genetics 172 2501–2514. [PMC free article] [PubMed]
  • Tian, F., D. J. Li, Q. Fu, Z. F. Zhu, Y. C. Fu et al., 2006. a Construction of introgression lines carrying wild rice (Oryza rufipogon Griff.) segments in cultivated rice (Oryza sativa L.) background and characterization of introgressed segments associated with yield-related traits. Theor. Appl. Genet. 112 570–580. [PubMed]
  • Tian, F., Z. Zhu, B. Zhang, L. Tan, Y. Fu et al., 2006. b Fine mapping of a quantitative trait locus for grain number per panicle from wild rice (Oryza rufipogon Griff.). Theor. Appl. Genet. 113 619–629. [PubMed]
  • Veyrieras, J. B., B. Goffinet and A. Charcosset, 2007. MetaQTL: a package of new computational methods for the meta-analysis of QTL mapping experiments. BMC Bioinformatics 8 49. [PMC free article] [PubMed]
  • von Korff, M., S. Grando, A. Del Greco, D. This, M. Baum et al., 2008. Quantitative trait loci associated with adaptation to Mediterranean dryland conditions in barley. Theor. Appl. Genet. 117 653–669. [PubMed]
  • Waller, F., B. Achatz, H. Baltruschat, J. Fodor, K. Becker et al., 2005. The endophytic fungus Piriformospora indica reprograms barley to salt-stress tolerance, disease resistance, and higher yield. Proc. Natl. Acad. Sci. USA 102 13386–13391. [PMC free article] [PubMed]
  • Wang, E., J. Wang, X. Zhu, W. Hao, L. Wang et al., 2008. Control of rice grain-filling and yield by a gene with a potential signature of domestication. Nat. Genet. 40 1370–1374. [PubMed]
  • Warrington, C. V., S. Zhu, W. A. Parrott, J. N. All and H. R. Boerma, 2008. Seed yield of near-isogenic soybean lines with introgressed quantitative trait loci conditioning resistance to corn earworm (Lepidoptera: Noctuidae) and soybean looper (Lepidoptera: Noctuidae) from PI 229358. J. Econ. Entomol. 101 1471–1477. [PubMed]
  • Xie, X., M. H. Song, F. Jin, S. N. Ahn, J. P. Suh et al., 2006. Fine mapping of a grain weight quantitative trait locus on rice chromosome 8 using near-isogenic lines derived from a cross between Oryza sativa and Oryza rufipogon. Theor. Appl. Genet. 113 885–894. [PubMed]
  • Xie, X., F. Jin, M. H. Song, J. P. Suh, H. G. Hwang et al., 2008. Fine mapping of a yield-enhancing QTL cluster associated with transgressive variation in an Oryza sativa × O. rufipogon cross. Theor. Appl. Genet. 116 613–622. [PubMed]
  • Xing, Y. Z., W. J. Tang, W. Y. Xue, C. G. Xu and Q. Zhang, 2008. Fine mapping of a major quantitative trait loci, qSSP7, controlling the number of spikelets per panicle as a single Mendelian factor in rice. Theor. Appl. Genet. 116 789–796. [PubMed]
  • Xue, W., Y. Xing, X. Weng, Y. Zhao, W. Tang et al., 2008. Natural variation in Ghd7 is an important regulator of heading date and yield potential in rice. Nat. Genet. 40 761–767. [PubMed]
  • Yuan, J., V. N. Njiti, K. Meksem, M. J. Iqbal, K. Triwitayakorn et al., 2002. Quantitative trait loci in two soybean recombinant inbred line populations segregating for yield and disease resistance. Crop Sci. 42 271–277. [PubMed]
  • Zeng, Z. B., 1994. Precision mapping of quantitative trait loci. Genetics 136 1457–1468. [PMC free article] [PubMed]
  • Zhang, Q., 2007. Strategies for developing Green Super Rice. Proc. Natl. Acad. Sci. USA 104 16402–16409. [PMC free article] [PubMed]

Articles from Genetics are provided here courtesy of Genetics Society of America
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...