Published in Crop Sci. 43:2234-2239 (2003).
© 2003 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
GENOMICS, MOLECULAR GENETICS & BIOTECHNOLOGY
Pyramiding and Validation of Quantitative Trait Locus (QTL) Alleles Determining Resistance to Barley Stripe Rust
Effects on Adult Plant Resistance
Ariel J. Castroa,
Xianming Chenb,
Ann Coreyc,
Tanya Filichkinac,
Patrick M. Hayes*,c,
Christopher Mundtd,
Kelley Richardsonc,
Sergio Sandoval-Islase and
Hugo Vivarf
a Departamento de Producción Vegetal, Est. Exp. "Dr. Mario A. Cassinoni", Facultad de Agronomía, Universidad de la República, Ruta 3 Km.373, Paysandú 60000, Uruguay
b U.S. Department of Agriculture, Agricultural Research Service, Washington State University, Pullman, WA 99164-6430, USA
c Department of Crop and Soil Science, Oregon State University, Corvallis, OR 97331-3002, USA
d Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331, USA
e Instituto de Fitosanidad, Colegio de Postgraduados, Chapingo, Mexico
f ICARDA/CIMMYT, Apdo. 6-641, 06600 Mexico, DF, Mexico
* Corresponding author (Patrick.M.Hayes{at}orst.edu).
 |
ABSTRACT
|
|---|
The use of molecular and quantitative trait locus (QTL) analysis tools initially lent support to the idea that relatively few genetic factors were the principal determinants of complex traits, including quantitative resistance (QR) to plant diseases. However, there are concerns regarding bias in QTL estimation and reproducibility of QTL effects in different genetic backgrounds. We are interested in mapping determinants of QR, and pyramiding resistance alleles at QTL loci may lead to durable resistance as well as provide independent validation of QTL effects and estimation of QTL interactions. We used molecular marker information to validate effects of resistance alleles at three QTL conferring QR to barley stripe rust (caused by Puccinia striiformis West. f. sp. hordei). Two of the QTL [one on chromosome 4(4H) and one on chromosome 7(5H)] trace to one parent, while another QTL on chromosome 5(1H) traces to a different parent. The pyramids of these QR alleles provide independent estimates of QTL effects, influence of genetic background on QTL effects, QTL x QTL interaction, and QTL x environment interaction. Our results validate QTL effect estimates, showing that a small number of QTL explained 94% of the genetic variation in trait expression in a new genetic background. Original QTL estimates were quantitatively biased, but that did not preclude the achievement of selection responses. We also confirmed the additive effects of the QTL alleles, as well as the consistent effects of QTL alleles across environments.
|
INTRODUCTION
|
|---|
GENETIC RESISTANCE is the most economical and environmentally appropriate strategy for disease control in plants. In the case of qualitative resistance—where genes of large effect clearly interact on a gene-for-gene basis with the pathogen—there is evidence that pathogen virulence can evolve more quickly than plant breeders can incorporate single resistance genes into new cultivars (Parlevliet, 1977). Quantitative Resistance (QR) shows continuous variation and it is usually incomplete in expression but has a higher probability of durability (Parlevliet, 1989). Before the advent of QTL analysis, estimates of gene number for QR were derived from statistical analyses of phenotypic data collected from segregating crosses and estimates of the number of "effective factors" determining QR ranged from two to 10 (Geiger and Heun, 1989). QTL approaches to estimating the number of genes determining QR have revealed that in some cases, a significant proportion of the total variance in QR is attributable to one locus or a few loci. However, small population sizes and model assumptions may cause estimates of gene number to be downwardly biased for both statistical analyses of phenotypic data (Geiger and Heun, 1989) and for QTL analyses (Young, 1996; Kover and Caicedo, 2001).
Stripe rust (caused by P. striiformis f. sp. hordei) is an important disease of barley worldwide. We initiated a collaborative effort (reviewed in Hayes et al., 2001) to map and introgress QR to barley stripe rust (BSR) from germplasm that has remained resistant to the spectrum of virulence encountered in Mexico, South America, and the USA (Sandoval-Islas et al., 1998). We have mapped BSR resistance QTL in multiple populations derived from crosses of resistance sources crossed with susceptible cultivars adapted to western North America (Chen et al., 1994, Toojinda et al., 2000). Since QTL from different sources were not coincident, in a previous paper (Castro et al., 2003), we proposed the development of QTL pyramids as a way to increase the resistance level of new germplasm. These pyramids of QTL alleles also provide a tool for an independent validation of QTL effects and for estimation of QTL interactions. Marker assisted selection (MAS) is necessary for pyramiding known QR genes because differential isolates of the pathogen cannot be used to determine the resistance allele architecture of the host, if the resistance is nonrace specific.
Evidence from the wheat stripe rust pathosystem shows that the growth stage at which resistance is expressed has important implications related to QR (Qayoum and Line, 1985). In this context, "seedling resistance" defines the resistance of a plant that if exposed to the same virulence throughout its life cycle shows resistance at all growth stages, and "adult plant resistance" defines the resistance that is only expressed at later growth stages. The latter, since it can be temperature dependent, has also been described as high temperature adult plant (HTAP) resistance (Chen and Line, 1995).
Resistance at the seedling stage is usually measured in terms of a binomially distributed variable, e.g., susceptible vs. resistant. To map determinants of seedling stage resistance, we used logistic regression techniques and found QTL with complementary gene action. In the Calicuchima-sib/Bowman population, QTL on chromosomes 4(4H) and 6(6H) were required for resistance at the seedling stage (Castro et al., 2002b), while in the Shyri/Galena population QTL on chromosomes 5(1H) and 6(6H) were required for resistance at the seedling stage (Castro et al., 2002a). Coincident QTL for resistance at the adult plant stage, with additive effects, were mapped in all cases (Chen et al., 1994; Toojinda et al., 2000). Additional adult plant stage-specific resistance QTL were mapped in both populations.
As described by Castro et al. (2003) we combined, in the same genetic background, two BSR resistance QTL alleles from the accession "Calicuchima-sib" on chromosomes 4(4H) and 7(5H) (Chen et al., 1994) (hereafter called QTL4 and QTL7), and a BSR resistance QTL allele from the cultivar "Shyri" on chromosome 5(1H) (Toojinda et al., 2000) (hereafter called QTL5). As shown by Castro et al. (2003) QTL effects on seedling resistance on chromosomes 4(4H) and 5(1H), and the epistatic interaction between them, were validated. The presence of resistance alleles at both QTL was necessary to recover the resistance phenotype. No QTL x race interaction was detected.
In this report, we describe the results of mapping determinants of resistance at the adult plant stage in the same germplasm described by Castro et al. (2003). Resistance at the adult plant stage, in this germplasm, is a quantitatively distributed trait, and this analysis provides independent estimates of QTL effects, influence of genetic background on QTL effects, QTL x QTL interaction, and QTL x environment interaction. Accordingly, the objectives of this paper were to: (i) validate the effects of QTL4, QTL5, and QTL7 on resistance at the adult plant stage in a new genetic background; (ii) study the interactions between QTL and their interaction with the environment; and (iii) determine the relationships between resistance to stripe rust at the seedling and adult plant stages.
|
MATERIALS AND METHODS
|
|---|
Plant Material
A population of 115 doubled haploid (DH) lines was developed from the cross Harrington*2/Orca/2/D1-72 as described by Castro et al. (2003). Orca (Calicuchima-sib/Bowman) is a spring feed barley cultivar developed by Oregon State University that has resistance alleles at two BSR resistance QTL, one on chromosome 4(4H) and another on chromosome 7(5H) (Hayes et al., 2000). D1-72 is an experimental line from the Shyri/Galena mapping population that carries a BSR resistance QTL on chromosome 5(1H) (Toojinda et al., 2000). Harrington, a malting cultivar developed by the University of Saskatchewan, is the malting quality standard for two-rowed cultivars in North America. It has no known BSR resistance genes. Backcross-1 (BC1) generation plants from Harrington*2/Orca were selected for the presence of Orca alleles by means of restriction fragment length polymorphism (RFLP) loci flanking the QTL regions on chromosomes 4(4H) and 7(5H). Four BC1 plants with Orca alleles at marker loci flanking QTL4 and QTL7 were crossed with D1-72. DH lines were derived from the F1 plants of these crosses, by the Hordeum bulbosum technique, as described by Chen and Hayes (1989).
Genotyping
The DH lines were genotyped in the regions defining the BSR QR QTL on chromosomes 4(4H), 5(1H), and 7(5H) by means of simple sequence repeat (SSR) markers. We first screened markers of known map position on Orca, Harrington, and D1-72 for polymorphisms. Since our interest was to define the regions of chromosomes 4(4H) and 7(5H) that were introgressed from Orca, map positions of markers in these regions were confirmed in the Calicuchima-sib x Bowman mapping population. In chromosome 5(1H), our interest was in the region tracing to Shyri (via D1-72) with Shyri x Galena as the reference population for confirming map position of polymorphic markers. We used four SSR markers (GMS21, Bmac213, Bmac399, and Bmac90) for genotyping the chromosome 5(1H) QTL region, three SSRs (EBmac788, HVMLO3, and HvAmyB) for the chromosome 4(4H) QTL region, and six SSRs (Bmac303, Bmac337, HVM30, EBmac970, Bmac113, and Bmac223) for the chromosome 7(5H) QTL region. On the basis of marker allele genotypes, we were able to determine the presence or absence of the corresponding QR loci alleles in each of the lines. PCR assays and fragment analysis were performed as described by Castro et al. (2003).
Phenotyping
The population was tested in seven experiments during 1996 (one planting date), 1998 (three planting dates), and 1999 (three planting dates) at the Centro Internacional de Mejoramiento de Maiz y Trigo (CIMMYT) facilities at Toluca, Mexico. Disease severity was measured on the basis of visual assessment of the percentage of diseased foliage, on a plot basis. The plant material was evaluated in two-row, 3-m plots. Spreader rows, planted at 5.25-m intervals and consisting of a mixture of 15 susceptible genotypes were inoculated twice with infected plants placed in the foliage, and with applications of spores suspended in oil. Infected plants and spores were collected locally. The race composition of this inoculum was not determined. Disease severity readings were taken at the end of the cycle in 1996 and at weekly intervals through grain filling, starting at the beginning of infection, in 1998 and 1999. Stripe rust epidemics were very consistent and intense. Data from multiple growth stages allowed for the calculation of the area under the disease progress curve (AUDPC) and the epidemic infection rate. AUDPC was calculated as the integral of the disease progress curve. For the calculation of infection rate, we used the Gompertz transformation {-log[-log (disease severity)]} (Berger, 1981; Fleming, 1983).
Data Analysis
Individual effects were estimated by an approach analogous to candidate gene analysis, where the genotypes at the QTL region (determined through the genotype of the markers at that region) are used as independent variables. Therefore, the independent variables had two levels each, with one level corresponding to the resistance allele (tracing to Orca in chromosomes 4 (4H) and 7 (5H), and tracing to D1-72 in chromosome 5 (1H)) and the other level corresponding to the alternative alleles at these QTL. There were no lines recombinant for the markers used to characterize QTL 4 or QTL 7 and as a consequence all DH lines were included in analyses of these QTL. One DH line was recombinant for QTL5 markers; accordingly, analysis of this QTL was based on 114 lines. The treatment design was a 2 x n factorial, where n is the number of genome regions considered. The difference between parental marker class means estimates the additive effect of the QTL flanked by the markers. Double crossovers between the QTL and marker loci downwardly bias estimates of the effects. Thus, differences between parental marker genotype means are conservative estimates of the effects of QTL residing in the n chromosomal regions. Statistical analyses were performed using the GLM procedure of SAS 7.1 (SAS Inst., 2001). Disease severity analysis was performed using data from seven environments, while AUDPC and infection rate analyses were performed using data from six environments (multiple disease readings were not taken in the seventh environment). Infection rate was determined by adjusting a multiple regression model for all of the data. The dependent variable was the infected area transformed with the Gompertz transformation y = -log[-log(infected area)]. The independent variables were individual experiments and QTL (dummy variables), QTL x QTL interactions, date of observation (expressed as days after emergence), QTL x date interactions, and experiment x date interactions. The model was selected using backward selection.
Individual allele frequencies were tested against expectations by a chi-square test. Orca alleles in the chromosome 4(4H) QTL were significantly higher in frequency than expected. Once observed allele frequencies were considered to calculate the expected genotypic frequencies, the observed genotypic frequencies were not significantly different from expected. The numbers of lines in each allele class are not the same. To determine the effect of this imbalance in sample size (i.e., more lines representing each of the single QTL resistance allele classes, fewer lines in the two resistance allele classes, and the fewest in the three resistance allele class), we randomly sampled in each class to generate subsamples with the same number of lines per class (n = 10). The results of the analysis of subsamples were the same, in terms of order and magnitude of differences of phenotypes and in declaration of significant differences, as the analysis based on the full population. Accordingly, we present data from the full population in this report. Components of variance were calculated from the datasets of the original mapping reports as well as the present dataset by the VARCOMP procedure of SAS 7.1 (Chen et al., 1994; Toojinda et al., 2000; SAS Inst., 2001). Total genetic variation, and components of variance for each QTL, were calculated to determine the amount of genetic variance explained by each QTL.
|
RESULTS
|
|---|
We detected significant main effects of the three QTL under study on the different measures of BSR infection considered (disease severity, AUDPC, and infection rate) (Tables 1 and 2). In all cases, the presence of resistance alleles at the three QTL regions (from Orca at QTL4 and QTL7 and from D1-72 at QTL5) was significantly associated with lower disease levels (Table 3). This provides evidence that these alleles have an effect on reducing BSR symptoms in a genetic background different from those in which the alleles were originally discovered and thus validating the effects of these alleles. For the phenotypes of disease severity and AUDPC, the BSR resistance alleles reduced disease in an additive fashion, except for one case of significant QTL x QTL interaction. The QTL resistance allele on chromosome 4(4H), contributed by the cultivar Orca, showed an interaction with the QTL resistance allele contributed by D1-72 on chromosome 5(1H).
View this table:
[in this window]
[in a new window]
|
Table 1. ANOVA table of the candidate gene analysis for disease severity and AUDPC. Only QTL main effects and QTL x QTL interactions are presented (all interactions involving experiments were nonsignificant).
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. Least squares means of disease severity and area under the disease progress curve (AUDPC), and least squares estimates of the infection rate according to the presence or absence of resistance alleles in the QTL regions under study on chromosomes 4(H), 5(1H) and 7(5H). Values in the same column followed by the same letter are not significantly different (p < 0.05) on the basis of pairwise comparisons (t-test).
|
|
The presence of two QTL resistance alleles led to greater resistance than the presence of either resistance allele alone, but the reduction in disease conferred by the two resistance alleles did not reduce disease symptoms as much as would be expected by the sum of the individual effects of the two alleles (Table 3). For the infection rate, two other QTL x QTL interactions were also detected (QTL4 x QTL7, and the three-way QTL interaction). As in the previous case, they were magnitude interactions, resulting in smaller that expected reductions of infection rates (Table 3).
The estimates of the combination of resistance alleles on chromosomes 4(4H) and 7(5H) are somewhat lower than expected, while the estimates of resistance alleles in all three chromosomes appear higher than expected. The accuracy of both groups of estimates is likely constrained by the limited number of DH lines representing the higher order combinations of alleles. We found that the same QTL were associated with disease severity and AUDPC.
QTL effects were confirmed in each experiment, regardless of epidemic intensity. No QTL x experiment interaction was detected. At the level of resolution afforded by our disease measurements, resistance alleles at the three QTL showed quantitative differences in the magnitude of their effects. Figure 1 shows results from the experiments in which the greatest disease pressure was observed (Third planting date in 1998, first planting date in 1999).
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 1. Average disease progress curves for doubled haploid lines with different combinations of resistance QTL alleles in two of the six experiments where this phenotype was measured. Figures on the left (98) correspond to the third planting date in 1998. Figures on the right (99) correspond to the first planting date in 1999. Figures in the first panel show results based on untransformed data; figures in the center show the parental lines; and figures in the lower panel show the results from the model adjusted with the Gompertz transformation, which was used to calculate the infection rate.
|
|
A comparison of QTL effects, as estimated in the source mapping populations and in the DH lines included in this experiment, reveals marked changes in magnitude of effect (Table 4). In the original QTL mapping report, the chromosome 4(4H) QTL had a very small effect on reducing disease severity, while the effect of the chromosome 7(5H) allele was much greater (Chen et al., 1994). We found the reverse in the current study. In the original mapping report, regarding QTL5, alleles at this locus were the main determinants of quantitative resistance to BSR, while in the current study the effects of alleles at this locus were less than those at QTL 4 but greater than those at QTL7 (Toojinda et al., 2000).
View this table:
[in this window]
[in a new window]
|
Table 4. Comparison of the amount of genotypic variance for disease severity  explained by the QTL effects in the original datasets (Chen et al., 1994; Toojinda et al., 2000), and in the MAS derived QR pyramid population, calculated using the same procedures in all cases. The heritabilities in the original mapping experiments were 0.77 (Shyri/Galena) and 0.74 (Cali/Bowman). The heritability in the present experiment was 0.56.
|
|
|
DISCUSSION
|
|---|
Our results validate the effects of resistance alleles at three QTL regions on disease severity and AUDPC, proving that they are significant determinants of adult plant stripe rust resistance. We also showed that, in the population under study, these QTL explain 94% of the genetic variation of the trait expression (Table 4). In a simulation study, Bernardo (2001) compared the selection efficiency of state-of-the-art estimates of the genetic value based on phenotypic data vs. those obtained with genotypic information and concluded that in most cases genotypic information did not improve selection response. In some cases, selection response decreased with the use of genotypic information. One of the cases where Bernardo (2001) did conclude that genotypic information would be valuable is when only a few genes control the trait. Our results, based on actual phenotypic and genotypic information, support this contention: three QTL explained 94% of genotypic variation in trait expression. In this case, despite biases in QTL effect estimation, the approach has generated information that should be useful for the development of cultivars with QR to BSR.
The effects of the three QTL were quite similar for two measures of the QR phenotype: disease severity and AUDPC. From a breeder's perspective, this finding is of great practical utility: disease severity is based on a single point reading, whereas AUDPC requires multiple observations.
Our present results provide some interesting perspectives on the growth stage specificity of disease resistance. The results from the analysis of the pyramid allele population at the seedling stage (Castro et al., 2003) validated the original mapping population reports; and the results of mapping at the adult plant stage also validate the original reports (Castro et al., 2002a, 2002b). QTL for seedling and adult plant resistance coincide at QTL4 and QTL5, but QTL 7 was not significant at the seedling stage. Likewise, QTL effects were epistatic at the seedling stage and additive at the adult plant stage.
The current level of resolution does not allow us to determine if both QTL effects at the seedling and adult plant stage correspond to the action of the same gene or genes. In that regard there is evidence in the literature of clustering of resistance genes in certain regions of plant genomes (Michelmore, 1995, Hulbert et al., 2001). The large physical sizes corresponding to regions certainly make it plausible that multiple genes could be found within the QTL confidence intervals. We do not have the data at this point to determine if QR is due to the alternative alleles at loci where other loci determine qualitative resistance, per the Robertsonian hypothesis. However, our results do not support the infinitesimal hypothesis that an infinite number of loci are the primary determinants of QR in this germplasm.
Considering our data in view of the wheat stripe rust pathosystem (Qayoum and Line, 1985) QTL4 and QTL5 may be determinants of "seedling resistance," while QTL7 may be a determinant of "adult plant" resistance. However, we have not yet conducted the necessary experiments to determine if QTL 7 is temperature sensitive. Our data suggest that an optimum strategy for developing BSR-resistant cultivars will involve QTL detection at the seedling and adult plant stages, followed by MAS for resistant alleles at multiple loci.
Our estimates of QTL effect in derived lines differ from those in the original mapping populations (Table 4). This could be due to bias in estimation of QTL effects in the source mapping populations or the validation population, or it could be due to uncharacterized interactions of the resistance QTL alleles with the new genetic background. The introgression of QTL alleles into a new genetic background shows that the resistance alleles at QTL4 have a higher "breeding value" than those at QTL7. This reversal of estimates of value indicates that, insofar as resources allow, even QTL of small effect should be target for MAS.
In summary, we validated the expression of QTL that significantly affect stripe rust infection in a different genetic background. We also showed that, in the germplasm we studied, a relatively small number of QTL explained a significant portion of the total genetic variation of the trait in trait expression. Although our original estimates of QTL effect may have been biased, we believe that our data justify the use of QTL mapping strategies for detecting and manipulating the genetic determinants of QR.
|
ACKNOWLEDGMENTS
|
|---|
This research was supported by the North American Barley Genome Project, USDA-ARS, the Oregon Grains Commission, the Idaho Barley Commission, and the Washington Barley Commission.
|
NOTES
|
|---|
Oregon Agricultural Experiment Station paper No. 11896.
Received for publication November 28, 2002.
|
REFERENCES
|
|---|
- Berger, R.D. 1981. Application of epidemiological principles to achieve plant disease control. Phytopathology 71:716–719.
- Bernardo, R. 2001. What if we knew all the genes for a quantitative trait in hybrid crops? Crop Sci. 41:1–4.[Abstract/Free Full Text]
- Castro, A., X.M. Chen, P.M. Hayes, and M. Johnson. 2003. Pyramiding of quantitative trait locus (QTL) alleles determining resistance to barley stripe rust: Effects on seedling resistance. Crop Sci. 43:651–659.[Abstract/Free Full Text]
- Castro, A., X. Chen, P.M. Hayes, S.J. Knapp, R.F. Line, T. Toojinda, and H. Vivar. 2002a. Coincident QTL that determine seedling and adult plant resistance to stripe rust in barley. Crop Sci. 42:1701–1708.[Abstract/Free Full Text]
- Castro, A., P.M. Hayes, T. Fillichkin, and C. Rossi. 2002b. Update of barley stripe rust resistance QTL in the Calicuchima-sib x Bowman mapping population. Barley Genet. Newsl. 32:1–12; also available online at http://wheat.pw.usda.gov/ggpages/bgn/; verified 17 June 2003.
- Chen, F.Q., and P.M. Hayes. 1989. A comparison of Hordeum bulbosum- mediated haploid production efficiency in barley using in vitro floret and tiller culture. Theor. Appl. Genet. 77:701–704.
- Chen F.Q., D. Prehn, P.M. Hayes, D. Mulroney, A. Corey, and H.E. Vivar. 1994. Mapping genes for resistance to barley stripe rust (Puccinia striiformis f.sp. hordei). Theor. Appl. Genet. 88: 215–219.
- Chen, X.M., and R.F. Line. 1995. Gene action in wheat cultivars for durable, high temperature, adult-plant resistance and interaction with race-specific, seedling resistance to Puccinia striiformis. Phytopathology 85:633–637.
- Fleming, R.A. 1983. Development of a simple mechanistic model of cereal rust progress. Phytopathology 73:308–312.
- Geiger, H.H., and M. Heun. 1989. Genetics of quantitative resistance to fungal diseases. Annu. Rev. Phytopathol. 27:317–341.[ISI]
- Hayes, P.M., A.E. Corey, R. Dovel, R. Karow, C. Mundt, K. Rhinart, and H. Vivar. 2000. Registration of Orca barley. Crop Sci. 40:849–851.
- Hayes, P.M., A. Castro, A. Corey, T. Filichkin, M. Johnson, C. Rossi, S. Sandoval, I. Vales, H.E. Vivar, and J. Von Zitzewitz. 2001. p. 47–60. In H.E. Vivar and A. McNab (ed.) Breeding barley in the new millennium. CIMMYT, Mexico.
- Hulbert, S.H., C.A. Webb, S.M. Smith, and Q. Sun. 2001. Resistance gene complexes: Evolution and utilization. Annu. Rev. Phytopathol. 39:285–312.[ISI][Medline]
- Kover, P.X., and A.L. Caicedo. 2001. The genetic architecture of disease resistance in plants and the maintenance of recombination by parasites. Mol. Ecol. 10:1–16.[Medline]
- Michelmore, R.W. 1995. Molecular approaches to manipulation of disease resistance genes. Annu. Rev. Phytopathol. 33:393–428.
- Parlevliet, J.E. 1977. Evidence of differential interaction in the polygenic Hordeum vulgare–Puccinia hordei relation during epidemic development. Phytopathology 67:776–778.
- Parlevliet, J.E. 1989. Identification and evaluation of quantitative resistance. p. 215–248 In K.J. Leonard, W.E. Fry (eds), Plant disease epidemiology, Vol. 2. McGraw-Hill, New York.
- Qayoum, A., and R.F. Line. 1985. High temperature, adult plant resistance to stripe rust of wheat. Phytopathology 75:1121–1125.
- SAS Inst. 2001. Statistical analysis system online documentation. Cary, NC.
- Sandoval-Islas, S., L.H.M. Broers, H.E. Vivar, and K.S. Osada. 1998. Evaluation of quantitative resistance to yellow rust (Puccinia striiformis f. sp. hordei) in the ICARDA/CIMMYT barley breeding program. Plant Breed. 117:127–130.
- Toojinda, T., L.H. Broers, X.M. Chen, P.M. Hayes, A. Kleinhofs, J. Korte, D. Kudrna, H. Leung, R.F. Line, W. Powell, L. Ramsay, H.E. Vivar, and R. Waugh. 2000. Mapping quantitative and qualitative disease resistance genes in a doubled haploid population of barley (Hordeum vulgare). Theor. Appl. Genet. 101:580–589.
- Young, N.D. 1996. QTL mapping and quantitative disease resistance in plants. Annu. Rev. Phytopathol. 34:479–501.[ISI][Medline]
This article has been cited by other articles:
|
|
|
|
 
T. Ishii and K. Yonezawa
Optimization of the Marker-Based Procedures for Pyramiding Genes from Multiple Donor Lines: I. Schedule of Crossing between the Donor Lines
Crop Sci.,
March 1, 2007;
47(2):
537 - 546.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
