JXB Advance Access originally published online on December 6, 2006
Journal of Experimental Botany 2007 58(2):351-360; doi:10.1093/jxb/erl214
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RESEARCH PAPER |
Marker-assisted selection to improve drought adaptation in maize: the backcross approach, perspectives, limitations, and alternatives
1CIMMYT, Apartado Postal 6-641, 06600 Mexico DF, Mexico
2Syngenta Seeds SAS, 12 chemin de l'Hobit, F31790 St-Sauveur, France
* To whom correspondence should be addressed. E-mail: J.Ribaut{at}cgiar.org
Received 4 April 2006; Accepted 22 September 2006
 |
Abstract |
|---|
 Top Abstract IntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
A number of different marker-assisted selection (MAS) approaches do exist for the improvement of polygenic traits. Results of a marker-assisted backcross (MABC) selection experiment aimed at improving grain yield under drought conditions in tropical maize are presented and compared with alternative MAS strategies. The introgression of favourable alleles at five target regions involved in the expression of yield components and flowering traits increased grain yield and reduced the asynchrony between male and female flowering under water-limited conditions. Eighty-five per cent of the recurrent parent's genotype at non-target loci was recovered in only four generations of MABC by screening large segregating populations (2200 individuals) for three of the four generations. Selected MABC-derived BC2F3 families were crossed with two testers and evaluated under different water regimes. Mean grain yield of MABC-derived hybrids was consistently higher than that of control hybrids (crosses from the recurrent parent to the same two testers as the MABC-derived families) under severe water stress conditions. Under those conditions, the best five MABC-derived hybrids yielded, on average, at least 50% more than control hybrids. Under mild water stress, defined as resulting in <50% yield reduction, no difference was observed between MABC-derived hybrids and the control plants, thus confirming that the genetic regulation for drought tolerance is dependent on stress intensity. MABC conversions involving several target regions are likely to result in partial rather than complete line conversion. Simulations were conducted to assess the utility of such partial conversions, i.e. containing favourable donor alleles at non-target regions, for subsequent phenotypic selection. The results clearly showed that selecting several genotypes (10–20) at each MABC cycle was most efficient. In the light of these results, alternative approaches to MABC are discussed, including recurrent selection, illustrated by an example of improving the adaptation of maize to low temperatures. Given the current approaches for MAS and the choices of marker technologies available now and potential for future developments, the use of MAS techniques in further improving grain yield under abiotic stresses in maize appears very promising.
Key words: Backcross, drought, marker-assisted selection, recurrent selection, Zea mays
|
Introduction |
|---|
|
TopAbstractIntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
Recent developments in plant molecular genetics have provided plant breeders with powerful tools to identify and select Mendelian components underlying both simple and complex agronomic traits (Ribaut and Hoisington, 1998; Dekkers and Hospital, 2002). Associations between markers and traits were first reported in maize by Stuber and Moll (1972) using isozymes. The advent of abundant DNA-based molecular markers allowed the construction of genetic maps (Botstein et al., 1980; Helentjaris et al., 1986). Continuous progress in the areas of DNA technology, such as the dicovery of the polymerase chain reaction (PCR) (Saiki et al., 1988), and genomics, have allowed the development of powerful genotyping systems, the construction of highly saturated maps (www.maizegdb.org; Lawrence et al., 2004; Schaeffer et al., 2006), and the establishment of countless marker–trait associations (Eathington, 2005). In addition to marker technologies, phenotyping technologies and statistical tools are critical to the ability to establish quality marker–trait associations (also called QTLs for quantitative trait loci). Over the past two decades, statistical methods to identify QTLs have evolved from rudimentary to very sophisticated. Significant milestones included the development of mixed model-based approaches (Jansen, 1993; Zeng, 1994; Jourjon et al., 2005), models for the analysis of non-experimental populations (Thornsberry et al., 2001), or approaches accounting for epistatic or genotype by environment interactions (GEIs; Jiang and Zeng, 1995; Vargas et al., 2006). Phenotyping technologies and methodologies have also evolved. Spatial analysis (Gleeson, 1997), for instance, represents an improvement over alpha lattice or incomplete block analyses in terms of data quality. In addition, the benefits of spatial analysis are greater for stressed environments where GEI is usually large. Developments in large-scale low-cost technologies for the quantification of metabolites also represent significant improvements in phenotypic characterization. There is no doubt that the collection and use of very high quality phenotypic data are critical for the application of marker-assisted selection (MAS).
Despite numerous reports of QTLs in maize (www.maizegdb.org/qtl.php), including QTLs for adaptation to water-limited conditions, reviewed by Tuberosa et al. (2002), Ribaut et al. (2004), and Sawkins et al. (2006), little has been published on the implementation of MAS based on these QTLs in breeding programmes. Successful MAS applications have been reported for introgression breeding in maize, including introgressions of transgenes (Ragot et al., 1995) and conversions involving simple (Ho et al., 2002; Morris et al., 2003) or complex traits (Bouchez et al., 2002; Willcox et al., 2002). Several marker-assisted recurrent selection (MARS) strategies have also been proposed and evaluated. When aimed at population improvement, MARS involves selecting individuals based on their marker genotypes and intermating them at random to produce the next generation (Hospital et al., 1997; Knapp, 1998; Lande and Thompson, 1990; Moreau et al., 1998). MARS can also be based on directed recombination between selected individuals of a segregating population, driven by an ideal genotype defined as a mosaic of favourable chromosomal segments from the parents of the segregating population (Stam, 1995; Van Berloo and Stam, 1998, 2001; Charmet et al., 1999). Although the use of MAS is increasing, it is not yet a common breeding practice, especially in the public sector. Besides cost, major limitations to the development of MAS might include the limited understanding of genetic phenomena such as gene networks, epistasis, and GEI.
Maize is the third most important food grain crop in the developing world. It is estimated that the demand for maize in developing countries will grow by 50%, from 558 million tons in 1995 to 837 million tons in 2020, outstripping the demand for wheat and rice. Much of this increased demand will be accommodated by increasing developing countries’ domestic supply, which will require intensifying production on existing agricultural land (IFPRI, 2000). Since the majority of tropical maize is grown under rainfed conditions, the possibilities for alleviation of water stress are limited. Consequently, there is a strong need to assist farmers in the developing world by providing them with well-characterized cultivars that can not only withstand greater levels of drought but also perform well under optimum conditions. In an attempt to achieve this objective, CIMMYT has allocated significant resources over the past 15 years to improve grain yield under water-limited conditions in tropical maize. The research and discussions presented here are part of this effort.
The objective of this article is to review how MAS can contribute to maize breeding. Results of a marker-assisted backcross (MABC) experiment conducted at CIMMYT to improve grain yield under water-limited conditions in tropical maize are presented and discussed. Potential extensions of the MABC protocol aimed at opportunistically capturing favourable attributes from the donor parent are assessed through computer simulations. Finally, other approaches, including MARS, are discussed as possible alternatives to MABC and illustrated with an independent MARS experiment.
|
The drought adaptation MABC experiment in tropical maize |
|---|
|
TopAbstractIntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
Plant material, phenotyping, and QTL detection
An MABC project was initiated in 1994 by crossing the drought-tolerant line Ac7643, the donor parent, and the drought-susceptible line CML247, the recurrent parent. CML247 is an elite tropical inbred line developed by CIMMYT and often used as a tester because of its outstanding combining ability and good yield per se under well-watered conditions. However, under water stress, CML247 displays a very large male–female flowering asynchrony (also known as ASI or anthesis–silking interval) and, as a consequence, is very susceptible to drought. A large ASI typically affects ovule pollination and therefore jeopardizes grain production (Bolaños and Edmeades, 1993). Ac7643 is tolerant to drought and has been previously used as the drought-tolerant parent of a segregating population (Ac7643xc7729/TZSRW) used to identify QTLs for secondary traits and yield components under water-limited conditions (Ribaut et al., 1996, 1997a). An F2:3 segregating population with 234 individuals was developed from the F1 between Ac7643 and CML247. F1 plants were also crossed to CML247 to generate a BC1F1 population.
All phenotypic evaluations were conducted during the dry winter season (November–April) in Tlaltizapan, Mexico (18°N, 940 m above sea level). Experimental designs consisted of alpha (0,1) lattices with two replications, similar to that described in Ribaut et al. (1996). Different water regimes were applied by restricting irrigation before and during flowering to obtain the desired drought-stress intensities. Irrigation was resumed at the end of the flowering stage, corresponding to the end of silk emergence, and maintained until harvest to allow grain filling. A well-watered regime was also conducted with irrigation applied every 12 d throughout the experiment.
Male (MFLW), and female flowering (FFLW) dates and yield component traits were collected. ASI was calculated as the difference between FFLW and MFLW for each plant (Ribaut et al., 1996). Yield component traits including grain yield, number of ears per plant, 100 kernel fresh weight, and number of kernels per ear were recorded on a family basis. Means were adjusted using spatial analysis (Gleeson, 1997).
A genetic map was constructed from polymorphic loci using MAPMAKER (Lander et al., 1987). Ten linkage groups were identified for a total map length of 1920 cM, and a mean marker density of 15.2 cM. The largest gap between two loci, on chromosome 9, was 46.2 cM. The order of the markers was found to be consistent with previously published maps.
QTL detection was performed using both simple interval mapping and two models of composite interval mapping (Zeng, 1994) as implemented in QTLCartographer (Basten et al., 1994). Co-factors were selected by forward and backward regressions with in and out thresholds set at a P-value of 0.01. Outputs of the three models were compared to make final decisions on the presence or absence of QTLs. QTLs were considered to be significant when their LOD value was >3.0. Allelic effects, additivity, and dominance were obtained from QTLCartographer at each significant QTL.
Marker-assisted backcrossing
Target regions for MABC were identified primarily from QTLs for ASI. Significant QTLs detected in 1994 and 1996 from the segregating population Ac7643xAc7729 (Ribaut et al., 1996, 1997a) and in 1996 from Ac7643xCML247, where, in both cases, Ac7643 was the drought-tolerant parent, were used. Five QTLs for ASI, located on chromosomes 1, 2, 3, 8, and 10, were selected across these two segregating populations. These five QTLs either explained >5% of the phenotypic variance for ASI in at least one of the segregating populations, were detected in both populations, or co-localized with QTLs for yield components. Based on these observations, these five QTLs were used as target loci to transfer favourable alleles for yield production from Ac7643 into CML247 through MABC. A cluster of QTLs for flowering traits and yield components was identified on chromosome 3, where CML247 displayed the favourable alleles. Polymorphic molecular markers were used to ensure the fixation of the CML247 allele at an early stage of the line conversion in this region. Together, selected QTLs explained 
38% of the total phenotypic variance for ASI in the Ac7643xCML247 segregating population.
Details of the MABC experiment are shown in Fig. 1. Selected plants from the first two generations, BC1F1 and BC2F1, were advanced by backcrossing to the recurrent parent. Selected plants from the following two generations, BC2F2 and BC2F3, were advanced by self-pollinating. About 400 BC1F1 plants were genotyped for the first generation of marker-based selection. Subsequent generations involved the genotyping of 2200 plants at each generation in an effort to increase the speed of the conversion process.
|
Marker-based selection was conducted in a stepwise manner on all four generations, from BC1F1 to BC2F3 (Fig. 1). Plants were first genotyped at three PCR-based markers which mapped within three of the target regions (on chromosomes 1, 2, and 10) and had been selected based on the quality of their polymorphisms and their combinability for multiplexed PCR amplifications (Ribaut et al., 1997b). Plants displaying heterozygous genotypes at these markers were selected before pollination to reduce the population size by 8-fold. Other genotypes were discarded. Then plants were genotyped at four pairs of restriction fragment length polymorphism (RFLP) markers flanking ASI QTLs located on chromosomes 1, 2, 8, and 10, plus one RFLP marker within the ASI QTL on chromosome 3. Plants displaying heterozygous genotypes at these markers were selected. Finally, for the second BC and the two self-pollinations, 60 polymorphic RFLP markers distributed throughout the genome were used to recover recurrent parent genotype at non-target regions. Plants with the highest proportion of CML247 alleles at these markers were selected. These last two selection steps involving RFLPs were conducted in the laboratory during grain filling and throughout harvest. At the end of each generation, ears from 10–15 individuals having Ac7643 alleles at target loci (within or flanking target regions) and the largest proportion of CML247 alleles at non-target loci (outside target regions) were selected to develop the next generation.
After the four MABC cycles, the 70 BC2F3 individuals presenting the closest allelic composition at target and non-target loci compared with the target genotype were crossed to two CIMMYT testers, CML254 and CML274. Ten individual CML247 plants were also crossed to the same two testers to produce 10 control hybrids. Eighty hybrids (70 MABC-derived and 10 controls) were thus evaluated for grain production in 1998, 1999, and 2001 under two water-stress regimes each year as well as a well-watered treatment in 1998 and 1999. From those 70 genotypes, 30 were selected based on their agronomic traits and yield performance. Doing so allowed capturing favourable alleles from Ac7643 outside the target loci and removing the genotypes presenting negative epistatic effects. Such selection was necessary as absolutely no phenotypic selection was applied during the MAS process.
Genetic composition of MABC-derived lines
Due to the large-scale selection that was conducted, the frequency of Ac7643 alleles at non-target loci for the best individuals was 13% after two backcross generations. These primarily flanked target regions due to linkage drag. This frequency decreased to 9% after the first self-pollination and 7% after the second self-pollination. The average frequency of donor genome at non-target loci across the 70 MABC-derived BC2F3 families was 14%.
Simulation analysis of partial conversions
Selecting several genotypes at each cycle, as conducted in this MABC experiment, increases the probability of identifying at least one genotype that carries donor alleles at some non-target loci. However, increasing the number of selected genotypes per cycle delays the line conversion process due to an increase in the average of donor genome contributions at non-target loci across selected genotypes. This trade-off was explored through simulations including screening at each cycle using a large population of 2000 genotypes and following the same model as described in Ribaut et al. (2002). The simulation results present the different average level of line conversion (left axis) expected across BC cycles considering that different numbers of genotypes had been selected at each cycle (5, 10, and 20) and the transfer of three or five QTL/target regions (15 cM each) (Fig. 2). The same figure indicates the expected percentage of the genome outside the selected regions at which the donor allele can be found in at least one selected genotype on both the carrying (with selected region) and the non-carrying (without selected region) chromosomes. Because the presence of donor alleles at non-target loci is random following Mendelian segregation on the non-carrying chromosome, but is affected by linkage drag at non-target loci linked to selected regions on the carrying chromosome, the probability of having a donor allele at non-target loci is always higher on non-carrying than on carrying chromosomes.
|
Independent of the level of conversion, the selection at each cycle of 20 genotypes compared with the selection of five genotypes increases the line conversion process by
0.3 and 0.6 cycles for three and five QTL/target regions, respectively. Under the simulated conditions, 80–90% line conversion can be obtained after two (86% and 82% for five and 20 selected plants, respectively) and three (86% and 81% for five and 20 selected plants, respectively) BC cycles for the transfer of three and five target regions, respectively. After two BC cycles and the transfer of three regions, the donor allele would be expected to be found at 52, 77, and 98% of non-target loci on non-carrying chromosomes when selecting five, 10, and 20 genotypes per cycle, while those values decrease to 38, 64, and 90% on carrying chromosomes. For the transfer of five regions, after three BCs, donor alleles would be expected to be found at 45, 78, and 99% of non-target loci on non-carrying chromosomes for selecting five, 10, and 20 genotypes per cycle, while those values decrease to 25, 50, and 82% on carrying chromosomes. Those simulations clearly indicate that if the recovery of donor alleles at non-target loci is part of the selection strategy, then it is beneficial to select several genotypes at each MABC cycle. Increasing the number of selected genotypes from five to 20 almost doubles and triples the percentage of the genome where remaining Ac7643 alleles can be expected at non-target loci on non-carrying and carrying chromosomes, respectively. The gain from selecting several genotypes is larger for five compared with three regions selected for the MABC. As the self-pollination cycles will reduce the percentage presented in Fig. 3 following Mendelian segregation, it is important that elite donor alleles at non-target loci can be present in more than one genotype, another advantage for the selection of 20 genotypes at each cycle.
|
Field performance of MABC-derived lines
Among the 30 MABC-derived hybrids, a subset of five best hybrids performing consistently among the best genotypes for grain yield production across the drought trials was identified. Field performance (grain yield) of the three different groups of genotypes (controls, 30 MABC hybrids, and five best MABC hybrids) is presented in Fig. 3. From this figure, it is clear that performance of the three types of genotypes is significantly dependent on the level of stress. Under most severe stress conditions [severe stress (SS) and intermediate stress (IS) 1998), grain yield reduction (grain yield under stress versus well-watered for each group) for both IS and SS represented
85% for the controls, 68% for the 30 MABC hybrids, and 43% for the five best MABC hybrids when crossed with CML254. With CML274, the yield reduction represented 86% for the control, 74% for the 30 MABC hybrids, and 68% for the five best MABC hybrids. Under those severe stress conditions, the five best MABC-derived hybrids yielded on average 2–3 times more than control hybrids, and the difference was larger for the hybrids obtained by using CML254 as tester compared with CML274. During 2001, where the stress level was less severe compared with 1998, as demonstrated by the grain yield performance for the controls, the same tendency was observed but the difference among groups was reduced. The 1999 cycle was favourable for grain yield production due to the absence of high temperature stress during flowering and the presence of light rain, which reduced the level of stress. Those climatic conditions contributed to a very high yield under well-watered conditions, almost 73% and 39% more, compared with the yield performance under 1998 well-watered conditions for CML254 and CML 274, respectively. Although the grain yield reduction between stress and non-stress conditions was still 45%, the trend indicated that the control-derived hybrids performed better compared with the 30 MABC- and five best MABC-derived hybrids with both testers. These results demonstrate that genetic gains to improve grain production under severe stress conditions have been achieved through the use of molecular markers. Under those severe stress conditions (1998 and 2001), a subset of the five best derived-hybrids has been identified that performed significantly better in seven out of eight stress trials over those two years. The five BC2F3 lines used to derived the five best MABC hybrids were passed to CIMMYT's breeders to be evaluated and used in their respective target environments. When water-limited conditions did not reduce grain yield by more than 45% (1999), MABC-derived hybrids performed equally or less well compared with the controls, demonstrating that no genetic gain was obtained under those mild stress conditions. It is important to underline that no yield penalty was observed under well-watered conditions (1998 and 1999), neither for the 30 MABC hybrids, nor for the five best MABC hybrids compared with the control. Those results clearly indicate that different genetic mechanisms control the plant phenotype depending on the stress intensity, and that under moderate stress conditions grain yield is highly dependent on the yield potential of the cultivar and drought adaptation is less of an issue (Blum, 2006).
|
The MABC approach |
|---|
|
TopAbstractIntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
Although several parameters influence what selection strategy should be followed, today the design of the most appropriate MABC strategy is relatively straightforward, and genetic gain can be predicted through simulation (Frisch et al., 2000). The transfer of five target regions using an optimal selectable population size of 50 genotypes (Ribaut et al., 2002) can only be obtained by screening a large population, as was the case in this study for three cycles of selection (2200 genotypes). Due to the optimal selectable population size and because a target partial line conversion of 85% was chosen, the MABC presented here was completed in four cycles. Although the genetic gain has been evaluated in a single location, which had also been used for the QTL identification, thus reducing the potential impact of GEI, this experiment demonstrates the impact of MABC on grain yield improvement under water-limited conditions.
Selection against the donor genome in gene introgression is more effectively achieved by using molecular markers than by using conventional backcrossing (Tanksley et al., 1989). Therefore, there is no doubt that MABC remains an efficient approach to transfer a single favourable allele (e.g. transgene) into different genetic backgrounds, or to improve a given genotype for a target trait. This latter approach is particularly important when breeding for human consumption, where the adoption of novel germplasm is challenging. The introgression of favourable alleles through MABC into farmer-favoured material, to improve it for a specific trait, is certainly one of the most suitable approaches for MAS in the South. Although the use of molecular markers to improve popular germplasm might have the largest short-term impact in developing countries, in the long term the transfer of favourable alleles from a donor to a recipient line is limited, as the output of such an approach can ‘only’ generate an improved version of an existing elite genotype. To overcome this limitation and take advantage of elite alleles coming from two or more parental lines, other MAS approaches can be considered, as described in the following cases.
|
MARS: a viable alternative? |
|---|
|
TopAbstractIntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
MARS and MABC
Although phenotypic recurrent selection is well suited for improvement of complex traits in maize (Hallauer and Miranda, 1981), it has rarely been used due to the prohibitively long cycle length. With the advent of molecular markers and the dissection of traits into QTLs, an ideotype or ideal genotype can be defined as a mosaic of chromosomal segments (QTLs) carrying favourable alleles from various parents. In maize, performance is often measured as the simultaneous outcome of several traits. Consequently, ideal genotypes will often be complex genotypes. Stam (1995) provides an explanation why, in the case of bi-parental populations, these genotypes will never occur in any Fn population of realistic size. Similarly, because of the number of loci involved and the relative contribution of the parents, the ideal genotype will generally not be attainable through MABC. MARS schemes, which involve several successive generations of crossing individuals based on their molecular marker genotypes, and the use of multitrait selection indices, would permit one to get close to or even achieve an ideal genotype (Peleman and Van Der Voort, 2003; Stam, 1995). The genetic gain achievable through MARS is probably higher than that achievable through MABC. For instance, when considering grain yield under both well-watered and water-limited conditions, MARS is likely to allow one to improve both traits, whereas MABC might restrict improvements to one of the two, as in the example shown above. More resources are likely to be needed for MARS than for MABC, although these may not be significantly larger if background markers are used to select for the recovery of the recurrent parent in the MABC approach.
MARS and phenotypic selection
Results from computer simulations have suggested that MARS was generally superior to phenotypic selection in accumulating favourable alleles in one individual (Van Berloo and Stam, 1998, 2001; Charmet et al., 1999). The advantage of MARS over phenotypic selection was largest when the populations under selection were very diverse, as is often the case when breeding with tropical or subtropical material. Further simulation studies showed that the relative advantage of MARS over phenotypic selection decreases rapidly when the fraction of the total genotypic variance explained by the QTLs included in the selection index decreased (Van Berloo and Stam, 2001).
In terms of practical MARS applications, a number of published studies have shown contrasting results. Both Moreau et al. (2004) and Openshaw and Frascaroli (1997) showed little, if any, advantage of MARS over phenotypic selection, even though genetic gains obtained through MARS were reported. In contrast, several studies (Ragot et al., 2000; Johnson, 2004; Crosbie et al., 2006) reported successful large-scale applications of MARS in private maize breeding programmes. Ragot et al. (2000) identified QTLs in a bi-parental maize population and conducted MARS using a genetic index constructed from several traits including agronomic performance (grain yield, grain moisture at harvest) and adaptation to abiotic stress (early vigour under cold conditions). Genetic values were computed for the index as well as for individual traits for all generations of the MARS process, including the two parents of the population, demonstrating an increase of genetic values of individual traits in the desired direction when selecting individuals based on an index (Fig. 4). A report of an independent breeding programme by (Eathington, 2005) demonstrated that the rate of genetic gain achieved through MARS was about twice that from phenotypic selection. Several recent accounts of commercial maize hybrids for which at least one of the parental lines had been derived through MARS also seem to suggest that MARS is being successfully implemented in private maize breeding programmes.
|
Multiallelic MARS
The most favourable alleles for traits of interest are likely to be spread across more than two lines, therefore requiring the assembly of alleles from many different sources in a single inbred line in order to achieve the greatest possible genetic gain. MARS can be used in such situations (Stam, 1995; Peleman and Van Der Voort, 2003). One obvious strategy would be the decomposition of the multiallelic situation into series of bi-allelic situations where the products of bi-allelic MARS schemes are brought together to assemble the final inbred lines. The use of MABC or partial backcrossing sequentially to pyramid favourable alleles from different sources might be a valuable alternative to the use of MARS if the number of target loci is low.
Yet, breeding with multiple sources of favourable alleles at once rather than conducting successive bi-allelic line development programmes could result in considerable time savings and therefore significantly improve the rate of genetic gain. The potential benefits justify investing in a thorough evaluation of multiparental MAS through simulation and practical research, and in a comparison of various realistic scenarios.
|
Other alternatives to MABC |
|---|
|
TopAbstractIntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
The most critical challenge for MAS remains the ability to predict the phenotype of any given genotype in a segregating population based on the allelic constitution at target loci. The complexity of gene networks and interactions among genetic elements or between genetic elements and the environment has often led to the discovery of fewer than expected stable significant QTLs across genetic pools and environments. This is a limitation which would strongly indicate that the probability of success of a MAS experiment can be increased where one detects QTLs in the target segregating population. However, the large number of QTLs characterized today cannot be ignored, nor can the large amounts of genomic sequence data available and the increasing amount of information produced through the ‘-omics’ technologies allowing the identification of candidate genes and helping validating target regions of interest when co-localizing with QTLs. In this context, new approaches based on existing genetic information, without the need to map QTLs in any new cross—a very time-consuming and expensive step—becomes more and more appealing. One can envisage determining target regions identified through the construction of consensus genetic linkage maps from the compilation of genetic data generated through different approaches and collected over crosses and environments. The underlying rationale for this approach is that genes involved in a target trait are most likely to be located in the same position in the crop genome, and that phenotypic differences across germplasm are created by the nature/quality and interaction of the alleles at those genes. Therefore, one can postulate that if a genomic region has been identified from various segregating populations for a target trait, there is a high probability that the same region will be involved in the regulation of a target phenotype in new genetic backgrounds. This type of MAS based on the selection of target regions identified from previous studies is currently being conducted at CIMMYT for the improvement of drought tolerance in tropical maize. Preliminary results from this study are very promising as genetic gains have been observed for several target traits in a couple of new segregating populations.
All strategies discussed above have one thing in common. They pyramid favourable alleles using linked markers to target genes, or cluster of genes, at an early stage of recombination. However, there is no doubt that to predict allelic performance in a new segregating population or in multiparental design, informative markers at candidate genes will represent a major advantage in increasing the efficiency of the selection. Association studies have great potential for resolving which individual genes may be responsible for the expression of a target plant phenotype and have the power to evaluate and characterize a wide range of alleles (Buckler and Thornsberry, 2002; Flint-Garcia et al., 2003a). Although common in human genetics in elucidating the genetic basis of a number of important human diseases (Lander and Schork, 1994; Risch and Merikangas, 1996), association approaches have only recently been applied to plant populations (Flint-Garcia et al., 2003b). In maize, due to extensive recombination, only polymorphisms at no more than a few hundreds of bases to a locus responsible for the phenotypic effect are likely to be significantly associated with the trait in a randomly mating population (Thornsberry et al., 2001). Thus informative polymorphisms characterized as gene haplotypes and associated with a contrasting phenotype can be converted into DNA markers for use in MAS experiments. Since the association approaches work with a wide range of germplasm, informative markers for most alleles affecting plant phenotype at a specific gene can be identified through association tests and used as a prediction tool in new and diverse genetic backgrounds. Combining genetic mapping with association studies has great potential for maize breeding.
|
Conclusions |
|---|
|
TopAbstractIntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
There is no doubt that the use of molecular markers in public and private maize plant breeding programmes has increased significantly over the last few years, as demonstrated by the increasing number of published papers on this topic and the successful applications of MARS in the private sector. However, the potential of new approaches described herein and the present state-of-the-art technologies for the discovery and the use of markers also indicate that achieving the full potential of molecular genetics and genomics in contributing to maize improvement will necessitate further research on methodologies. Considering the seemingly exponential growth in the amount of genetic information being generated, there is no doubt that the major challenge for MAS will be in the integration of this diverse and disparate information and interpretation in a specific biological context to convert it into knowledge. For example, with the evolution of the current technology, one can state with some certainty that the haplotyping of several thousand genes of maize lines as a pre-breeding tool will soon become a reality. What may take longer and will be more challenging is to determine the biological value of the different alleles at those genes under target environmental conditions. This challenge is even greater for target environments exhibiting a large GEI such as water-limited environments. Modern maize breeding for complex traits, a combination of molecular and phenotypic selection, should become routine in breeding programmes, but success will largely be dependent on the accuracy of plant phenotyping and the capacity to determine the GEI; two major components that affect the prediction of the allelic value on the plant phenotype in new populations.
|
Acknowledgements |
|---|
The authors would like to thank Changjian Jiang for running the simulations, Mateo Vargas for conducting the QTL analysis, and Kaitlin Lesnick and Mark Sawkins for helpful review of the manuscript.
|
Abbreviations |
|---|
ASI, anthesis–silking interval; GEI, genotype by environment interaction; MABC, marker-assisted backcross; MARS, marker-assisted recurrent selection; MAS, marker-assisted selection; RFLP, restriction fragment length polymorphism; QTL, quantitative trait locus.
|
References |
|---|
|
TopAbstractIntroductionThe drought adaptation MABC...The MABC approachMARS: a viable alternative?Other alternatives to MABCConclusionsReferences |
|---|
Basten CJ, Weir BS, Zeng Z-B. (1994) Zmap—a QTL cartographer. In Smith C, Gavora JS, Benkel B, Chesnais J, Fairfull W, Gibson JP, Kennedy BW, Burnside EB (Eds.). Proceedings of the 5th World Congress on Genetics Applied to Livestock Production: Computing Strategies and SoftwareGuelph, Ontario University of Guelph Vol. 22 pp. 65–66.
Botstein D, White RL, Skolnick M, Davis RW. (1980) Construction of genetic linkage map in human using restriction fragment length polymorphisms. American. Journal of Human Genetics 32 314–331.
Blum A. (2006) Drought adaptation in cereal crops—a prologue. In Ribaut J-M (Ed.). Drought tolerance in cerealsBinghamtown, NY The Haworth Press Inc pp. 10–18.
Bolaños J and Edmeades GO. (1993) Eight cycles of selection for drought tolerance in lowland tropical maize. II. Responses in reproductive behaviour. Field Crops Research 31 253–268.[CrossRef]
Bouchez A, Hospital F, Causse M, Gallais A, Charcosset A. (2002) Marker-assisted introgression of favorable alleles at quantitative trait loci between maize elite lines. Genetics 162 1945–1959.
Buckler ES and Thornsberry JM. (2002) Plant molecular diversity and applications to genomics. Current Opinion in Plant Biology 5 107–111.[CrossRef][ISI][Medline]
Charmet G, Robert N, Perretant MR, Gay G, Sourdille P, Groos C, Bernard S, Bernard M. (1999) Marker-assisted recurrent selection for cumulating additive and interactive QTLs in recombinant inbred lines. Theoretical and Applied Genetics 99 1143–1148.[CrossRef]
Crosbie TM, Eathington SR, Johnson GR, et al. (2006) Plant breeding: past, present, and future. In Lamkey KR and Lee M (Eds.). Plant breeding: the Arnel R. Hallauer international symposiumAmes, IA Blackwell Publishing.
Dekkers JCM and Hospital F. (2002) The use of molecular genetics in the improvement of agricultural populations. Nature Reviews 3 22–32.
Eathington S. (2005) Practical applications of molecular technology in the development of commercial maize hybrids. Proceedings of the 60th Annual Corn and Sorghum Seed Research ConferencesWashington, DC American Seed Trade Association.
Flint-Garcia SA, Darrah L, McMullen M. (2003b) Phenotypic versus marker-assisted selection for stalk strength and second-generation European corn borer resistance in maize. Theoretical and Applied Genetics 107 1331–1336.[CrossRef][ISI][Medline]
Flint-Garcia SA, Thornsberry JM, Buckler ES. (2003a) Structure of linkage disequilibrium in plants. Annual Review of Plant Biology 54 357–374.[CrossRef][Medline]
Frisch M, Bohn M, Melchinger AE. (2000) PLABSIM: software for simulations of marker-assisted backcrossing. Journal of Heredity 91 86–87.
Gleeson AC. (1997) Spatial analysis. In Kempton RA and Fox PN (Eds.). Statistical methods for plant variety evaluationLondon Chapman and Hall pp. 68–85.
Haley CS and Knott SA. (1992) A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69 315–324.[ISI][Medline]
Hallauer AR and Miranda JB. (1981) Quantitative genetics in maize breedingAmes, IA Iowa State University Press.
Helentjaris T, Slocum M, Wright S, Schasfer A, Nienhuis J. (1986) Construction of genetic linkage maps in maize and tomato using restriction fragment length polymorphisms. Theoretical and Applied Genetics 72 761–769.[ISI]
Ho JC, McCouch SR, Smith ME. (2002) Improvement of hybrid yield by advanced backcross QTL analysis in elite maize. Theoretical and Applied Genetics 105 440–448.[CrossRef][ISI][Medline]
Hospital F, Moreau L, Lacoudre F, Charcosset A, Gallais A. (1997) More on the efficiency of marker-assisted selection. Theoretical and Applied Genetics 95 1181–1189.[CrossRef]
IFPRI. (2000) 2020 projectionsWashington, DC IFPRI publication.
Jansen RC. (1993) Interval mapping of multiple quantitative trait loci by using molecular markers. Theoretical and Applied Genetics 85 252–260.
Jiang C and Zeng ZB. (1995) Multiple trait analysis of genetic mapping for quantitative trait loci. Genetics 140 1111–1127.[Abstract]
Johnson GR. (2004) Marker assisted selection. In Janick J (Ed.). Plant breeding reviewsHoboken, NJ John Wiley & Sons Vol. 24 pp. 293–310 part 1.
Jourjon M-F, Jasson S, Marcel J, Ngom B, Mangin B. (2005) MCQTL: multi-allelic QTL mapping in multi-cross design. Bioinformatics 21 128–130.
Knapp SJ. (1998) Marker-assisted selection as a strategy for increasing the probability of selecting superior genotypes. Crop Science 38 1164–1174.
Lande R and Thompson R. (1990) Efficiency of marker-assisted selection in the improvement of quantitative traits. Genetics 124 743–756.[Abstract]
Lander ES, Green P, Abrahamson J, Barlow A, Daly MJ, Lincoln SE, Newburg L. (1987) MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1 174–181.[CrossRef][Medline]
Lander ES and Schork NJ. (1994) Genetic dissection of complex traits. Science 265 2037–2048.
Lawrence CJ, Dong Q, Polacco ML, Seigfried TE, Brendel V. (2004) MaizeGDB, the community database for maize genetics and genomics. Nucleic Acids Research 32 D393–D397.
Moreau L, Charcosset A, Gallais A. (2004) Experimental evaluation of several cycles of marker-assisted selection in maize. Euphytica 137 111–118.[CrossRef][ISI]
Moreau L, Charcosset A, Hospital F, Gallais A. (1998) Marker-assisted selection efficiency in populations of finite size. Genetics 148 1353–1365.
Morris M, Dreher K, Ribaut J-M, Khairallah M. (2003) Money matters (II): costs of maize inbred line conversion schemes at CIMMYT using conventional and marker-assisted selection. Molecular Breeding 11 235–247.[CrossRef]
Openshaw S and Frascaroli E. (1997) QTL detection and marker-assisted selection for complex traits in maize. In: Proceedings of the 52nd Annual Corn and Sorghum Research ConferenceWashington, DC American Seed Trade Association pp. 44–53.
Peleman JD, Van Der Voort JR. (2003) Breeding by design. Trends in Plant Science 7 330–334.
Ragot M, Biasiolli M, Delbut MF, Dell'Orco A, Malgarini L, Thevenin P, Vernoy J, Vivant J, Zimmermann R, Gay G. (1995) Marker-assisted backcrossing: a practical example. In Bervillé A and Tersac M (Eds.). Les Colloques, no 72, techniques et utilisations des marqueurs moléculairesParis INRA pp. 45–56.
Ragot M, Gay G, Muller J-P, Durovray J. (2000) Efficient selection for the adaptation to the environment through QTL mapping and manipulation in maize. In Ribaut J-M and Poland D (Eds.). Molecular approaches for the genetic improvement of cereals for stable production in water-limited environments. MéxicoDF CIMMYT pp. 128–130.
Ribaut J-M, Bänziger M, Setter T, Edmeades G, Hoisington D. (2004) Genetic dissection of drought tolerance in maize: a case study. In Nguyen H and Blum A (Eds.). Physiology and biotechnology integration for plant breedingNew York Marcel Dekker Inc. pp. 571–611.
Ribaut J-M and Hoisington D. (1998) Marker-assisted selection: new tools and strategies. Trends in Plant Science 3 236–239.[CrossRef][ISI]
Ribaut J-M, Hoisington DA, Deutsch JA, Jiang C, González-de-León D. (1996) Identification of quantitative trait loci under drought conditions in tropical maize. 1. Flowering parameters and the anthesis–silking interval. Theoretical and Applied Genetics 92 905–914.[CrossRef]
Ribaut J-M, Hu X, Hoisington D, González-de-León D. (1997b) Use of STSs and SSRs as rapid and reliable preselection tools in a marker assisted selection backcross scheme. Plant Molecular Biology Reporter 15 154–162.
Ribaut J-M, Jiang C, González-de-León D, Edmeades GO, Hoisington DA. (1997a) Identification of quantitative trait loci under drought conditions in tropical maize. 2. Yield components and marker-assisted selection strategies. Theoretical and Applied Genetics 94 887–896.[CrossRef]
Ribaut J-M, Jiang C, Hoisington D. (2002) Efficiency of a gene introgression experiment by backcrossing. Crop Science 42 557–565.
Risch N and Merikangas K. (1996) The future of genetic studies of complex human diseases. Science 273 1516–1517.[ISI][Medline]
Saiki RK, Gelfand DH, Stoffel S, Scharf S, Higuchi RH, Horn GT, Mullis KB, Erlich HA. (1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239 487–491.
Sawkins M, de Meyer J, Ribaut J-M. (2006) Drought adaptation in cereal crops—drought adaptation in maize. In Ribaut J-M (Ed.). Drought tolerance in cerealsBinghamtown, NY The Haworth Press Inc. pp. 356–387.
Schaeffer ML, Sanchez-Villeda H, McMullen M, Coe E Jr. (2006) IBM2 2005 neighbors—45 000 locus resource for maize. Plant and Animal Genome Conference Proceedings XIV 200.
Stam P. (1995) Marker-assisted breeding. In Van Ooijen JW and Jansen J (Eds.). Biometrics in plant breeding: applications of molecular markers. Proceedings of the Ninth Meeting of the EUCARPIA Section Biometrics in Plant BreedingWageningen, The Netherlands CPRO-DLO pp. 32–44.
Stuber CW and Moll RH. (1972) Frequency changes of isozyme alleles in a selection experiment for grain yield in maize (Zea mays L.). Crop Science 12 337–340.
Tanksley SD, Young ND, Patterson AH, Bonierbale MW. (1989) RFLP mapping in plant breeding: new tools for an old science. BioTechnology 7 257–264.[CrossRef]
Thornsberry JM, Goodman MM, Doebley J, Kresovich D, Nielsen D, Buckler ES IV. (2001) Dwarf8 polymorphisms associate with variation in flowering time. Nature Genetics 28 286–289.[CrossRef][ISI][Medline]
Tuberosa R, Salvi S, Sanguineti MC, Landi P, MacCaferri M, Conti S. (2002) Mapping QTLs regulating morpho-physiological traits and yield: case studies, shortcomings and perspectives in drought-stressed maize. Annals of Botany 89 941–963.
Van Berloo R and Stam P. (1998) Marker-assisted selection in autogamous TIL populations: a simulation study. Theoretical and Applied Genetics 96 147–154.[CrossRef]
Van Berloo R and Stam P. (2001) Simultaneous marker-assisted selection for multiple traits in autogamous crops. Theoretical and Applied Genetics 102 1107–1112.[CrossRef]
Vargas M, van Eeuwijk F, Crossa J, Ribaut J-M. (2006) Mapping QTLs and QTLxenvironment interaction for CIMMYT maize drought stress program using factorial regression and partial least squares methods. Theoretical and Applied Genetics 112 1009–1023.[CrossRef][ISI][Medline]
Willcox M, Khairallah M, Bergvinson D, et al. (2002) Selection for resitance to southwaestern corn borer using marker-assisted and conventional backcrossing. Crop Science 42 1516–1528.
Zeng Z-B. (1994) Precision mapping of quantitative trait loci. Genetics 136 1457–1468.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. P. Moose and R. H. Mumm Molecular Plant Breeding as the Foundation for 21st Century Crop Improvement Plant Physiology, July 1, 2008; 147(3): 969 - 977. [Full Text] [PDF]
|
|
|
|
|
|
N. C. Collins, F. Tardieu, and R. Tuberosa Quantitative Trait Loci and Crop Performance under Abiotic Stress: Where Do We Stand? Plant Physiology, June 1, 2008; 147(2): 469 - 486. [Full Text] [PDF]
|
|
|
|
 |
|
 R. Tuberosa, S. Salvi, S. Giuliani, M. C. Sanguineti, M. Bellotti, S. Conti, and P. Landi Genome-wide Approaches to Investigate and Improve Maize Response to Drought Crop Sci., December 18, 2007; 47(Supplement_3): S-120 - S-141. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||