Published online 1 July 2008
Published in Crop Sci 48:1419-1424 (2008)
© 2008 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
Microsatellite Markers for Kernel Color Genes in Wheat
J. D. Shermana,
E. Souzab,
D. Seec and
L. E. Talberta,*
a Dep. of Plant Sciences and Plant Pathology, Montana State Univ., Bozeman MT 59717
b USDA-ARS, Soft Wheat Quality Research, Williams Hall, 1680 Madison Ave., Wooster, OH 44691
c USDA-ARS, Western Regional Small Grain Genotyping Lab., 209 Johnson Hall, Washington State Univ., Pullman, WA 99164–6420
* Corresponding author (usslt{at}montana.edu).
 |
ABSTRACT
|
|---|
The establishment of hard white wheat (Triticum aestivum L.) as a viable alternative for growers has been impeded by several factors, one of which is that new hard white wheat cultivars may not be competitive with hard red cultivars. This is due to the fact that most breeding programs devote more resources to hard red wheat, and that the genetics of kernel color makes rapid conversion of red to white kernel color problematic. Three homoeologous loci control kernel color, with red being dominant. A single red allele is sufficient to cause the kernel to be classified as red. A rapid conversion of red-seeded genotypes to white-seeded types would be facilitated by the use of molecular markers to help select for the recessive alleles for white color. In this experiment, we developed three populations that were segregating at only one of each of three respective color loci, being homozygous recessive at the other two. F2 plants or recombinant inbred lines were assayed for kernel color and screened with a series of microsatellite markers. Linked microsatellite markers were identified for all loci. A validation experiment was established with a population of 1786 F2 plants from a cross between a red-seeded line and white-seeded line. The markers were 100% diagnostic for the white-seeded phenotype in this population. Thus, the markers will find utility in backcrossing programs to convert red-seeded wheat to white.
Abbreviations: PCR, polymerase chain reaction
|
INTRODUCTION
|
|---|
SIX CLASSES OF WHEAT, hard red winter, hard red spring, soft white, soft red, durum, and hard white, are primarily based on simply inherited genetic characteristics of seed color, vernalization requirement, and seed hardness. Red and white refer to the color of the seed pericarp, which is controlled by three independent, homoeologous genes, located on chromosomes 3A, 3B, and 3D. The alleles conferring red color are denoted as R-A1b, R-B1b, and R-D1b, and alleles conferring white color as R-A1a, R-B1a, and R-D1a (McIntosh et al., 1998). Red color is dominant to white, and a single locus containing the dominant allele is sufficient to result in red color (Metzger and Silbaugh, 1970). The degree of red color is additive, with genotypes homozygous dominant at all three loci (R-A1b, R-B1b, and R-D1b) having the darkest red color, and only those homozygous recessive at all three genes being white (R-A1a, R-B1a, and R-D1a).
Hard white wheat is the newest class of wheat recognized in the United States (Lin and Vocke, 2004) and may offer advantages for certain end-use applications. Many types of Asian noodles are made with hard white wheat, primarily due to superior color characteristics (Miskelly, 1984). In addition, domestic markets have an interest in hard white wheat due to possible advantages in milling and in end-use quality (Taylor et al., 2005). In some regions, breeding hard white wheat cultivars is generally the responsibility of the same breeding programs that develop hard red spring wheat cultivars. Due to the long history of breeding in hard red wheat, and current demand by wheat growers for new hard red varieties, hard white wheat has not benefited to the same degree in genetic improvement as has hard red wheat. Thus, in many regions, the selection of hard white wheat cultivars is limited, and agronomic performance may not be as good as with the available hard red wheat lines.
The fact that three homozygous recessive loci are need to give white seed color makes conversion of red-seeded types to white-seeded cumbersome. For instance, in a cross between a "three-gene" red line (homozygous for R-A1b, R-B1b, and R-D1b) and a white-seeded line (homozygous for R-A1a, R-B1a, and R-D1a), only 1/64 of the F2 progeny are expected to be homozygous recessive at all three loci. Furthermore, since seed color is determined by the maternal parent, homozygous recessive F2 individuals can be detected only by observing F3 seed. The issue of few white lines may be somewhat alleviated by inbreeding, as inbred lines derived from a cross between a three-gene red-seeded genotype and a white-seeded genotype are expected to be homozygous recessive at all three loci at a frequency of 1/8. However, the inbreeding process is time intensive and impedes the ability of breeders to rapidly incorporate the agronomic properties of superior hard red genotypes into white-seeded types.
Molecular markers for alternative alleles at each of the R loci have the potential to facilitate the conversion of superior red-seeded genotypes into white-seeded derivatives by a backcrossing program. The advantage to this approach is that white-seeded germplasm could be quickly improved to the level of existing hard red cultivars and elite lines. Thus, for this project, we developed three populations, each segregating for the dominant allele at only one R locus. Closely linked microsatellite markers were identified at each of the three homoeologous loci, and the markers were used as proof of concept by selecting white lines in an F2 population segregating for all three genes.
|
MATERIALS AND METHODS
|
|---|
Plant Materials
Populations were established to allow identification of polymerase chain reaction (PCR)-based markers that cosegregate with each of the kernel-color genes on homoeologous Group 3 chromosomes as follows:
‘Rio Blanco’ (PI 531244)/Idaho 444 (PI 578278)
Rio Blanco is a hard white winter cultivar, while Idaho 444 is a hard red winter experimental line (Windes et al., 1995). Recombinant inbred lines were created by F2 single-seed descent with derivation at F7. The complete population consisted of over 300 lines. Observation of recombinant inbred lines from this cross suggested a 1:1 segregation for red and white seed color, indicating that a single locus was segregating (presumed genotype R-A1b, R-B1a, R-D1a; see Results) (Souza, unpublished data, 2007). To test for cosegregation of seed color with PCR markers, DNA was isolated from 46 red- and 46 white-seeded lines that were randomly selected from the complete population.
‘Dollar’ (PI592117)/MTHW9904
Dollar is a hard red spring wheat cultivar reported as being R-A1a, R-B1b, R-D1a (McIntosh et al., 1998). MTHW9904 is a white-seeded experimental line. Two hundred F3 seeds from the cross were phenotyped as described below, resulting in 145 red-seeded and 55 white-seeded plants, which is not significantly different from a 3:1 ratio (x2 = 0.67, p > 0.3). DNA was isolated from equal numbers of red- and white-seeded F2 plants for PCR analysis.
‘Chinese Spring’ (CItr 14108)/MTHW9904
Chinese Spring has been reported as being R-A1a, R-B1a, R-D1b (McIntosh et al., 1998). Phenotypic analysis of F3 seed found 143 red individuals and 51 white out of 194 screened, which is not significantly different from a 3:1 ratio (x2 = 0.17, p > 0.50). DNA was extracted from 141 F2 plants that produced red-seeded progeny, and 47 F2 individuals that produced white-seeded progeny.
Validation Population
To identify a good population for marker validation we screened several red and white Montana lines to determine if the markers were polymorphic, including ‘Choteau’ (C) and ‘Vida’ (V), as well as two Montana white-seeded experimental lines MTHW0202 (02) and MTHW0471 (04) (Fig. 1
). Vida (PI 642366), a hard red spring wheat, and MTHW0471 a hard white spring experimental line, were used to create a set of material to validate the utility of the PCR markers identified in the previous three populations. We developed a population of 1786 F2 individuals segregating for all three color genes from a cross of Vida by MTHW0471. DNA was isolated from leaf tissue from this population for high-throughput marker analysis (see below). F3 seed from each plant was screened for kernel color.
View larger version (80K):
[in this window]
[in a new window]
|
Figure 1. Screening gels for the three microsatellite markers (xgwm155-3A, xgwm4010-3B, and xgwm4306-3D) linked to the seed-color loci in wheat in two red-seeded varieties, ‘Choteau’ (C) and ‘Vida’ (V), as well as two Montana white-seeded experimental lines, MTHW0202 (02) and MTHW0471 (04). Each 15% polyacrylamide gel has a marker in the far right lane with fragments of 240 and 670 base pairs.
|
|
Phenotyping
Initial phenotyping was based on visual observation of untreated seed. In that seed color is determined by the maternal genotype, the phenotype of F3 seed from F2 plants is controlled by the F2 genotype. Kernel color phenotypes were determined by analysis of F3 seed harvested from individual F2 plants. The alkali test to detect hard white/red wheat kernels (USDA, 2003) was used as follows: A single seed was removed from each head and placed in a flat-bottomed 96-well plate; 200 µL of a fresh solution of 15 g KOH and 40 mL household bleach (6% Na hypochlorite) was added to each well, and the plate was allowed to shake for 5 min on an orbital shaker. The seeds were removed from the solution and visually examined immediately. Although this technique improved our ability to differentiate red and white types, we still had difficulty distinguishing very light red types from white types. This resulted in a decision to select individuals of contrasting phenotype, where possible, for mapping populations. This likely reduced the number of heterozygous individuals in F2 populations. Additionally, our assumption is that there may be a small amount of phenotypic misclassification in the populations.
Molecular Techniques
Genomic DNA was isolated from leaf tissue as described by Riede and Anderson (1996).
Previous maps were referenced to identify candidate microsatellite markers using the GrainGenes database (http://wheat.pw.usda.gov/cgi-bin/graingenes/browse.cgi). Approximately 50 ng of DNA was used in PCR as described by Roder et al. (1998). Annealing temperatures were used as suggested on GrainGenes (http://wheat.pw.usda.gov/GG2/quickquery.shtml, 2006). Potential linked markers were screened on parents. To increase the likelihood of identifying tightly linked markers, the Institute of Plant Genetics and Crop Plant Research (http://www.ipk-gatersleben.de/Internet) and TraitGenetics (http://traitgenetics.de/ie/home3.html) were contacted for permission to use proprietary markers. Associated markers are listed in Table 1
. The primers for the three most tightly linked markers amplified well using an annealing temperature of 60°C, and all three markers were resolved on 15% polyacrylamide gels stained with ethidium bromide and photographed using a charged-coupled device camera and a thermal printer (Fig. 1).
View this table:
[in this window]
[in a new window]
|
Table 1. Markers found to cosegregate with seed color in crosses of red-seeded wheat genotypes having red alleles at a single locus and white-seeded genotypes.
|
|
High-Throughput Genotyping
For high-throughput marker detection, the addition of a tail sequence was added to the forward primer of the flanking microsatellite markers (Boutin et al., 1997). Tailed PCR was done in a 12-µL reaction which contained 100 ng gDNA, 1x reaction buffer (New England Biolabs), 3.0 mM MgCl2, 3 pmol reverse primer, 2.4 pmol tailed primer with attached fluorophore (M13 CACGACGTTGTAAAACGAC), 0.6 pmol forward-tailed primer, and 0.6 U DNA polymerase (New England Biolabs). Thermocyclic conditions consisted of 3 min at 94°C followed by 41 cycles of 94°C 1 min, specific annealing temp 1 min, 72°C 1 min, and finishing with one step at 72°C for 10 min. PCR products were diluted in H2O added to formamide, denatured, then loaded onto a 3130 x l DNA sequencer. The tailed primers allowed cost-effective incorporation of four different fluorochromes that provided the ability to pool PCR products and detect the polymorphic products on an ABI 3130 x l DNA sequencer. The results from the analysis were screened using software from SoftGenetics (SoftGenetics, State College, PA).
Mapping
The mapping analysis was conducted by MAPMAKER 3.0b (Lander et al., 1987). Linkage was determined by the group, compare, and map commands using the Kosambi function. Best orders were tested with ripple. Maps were drawn using the draw map command, generating a postscript file.
|
RESULTS
|
|---|
White-seeded individuals were observed in the Rio Blanco/Idaho 444 cross with approximate 1:1 ratio, suggesting that the population was segregating at a single seed-color locus. We screened the recombinant inbred lines with a total of 60 markers from 3A (29), 3B (14), and 3D (17), including the markers that would be determined to be most tightly linked to seed color on 3B and 3D. The only markers that showed linkage were those on chromosome 3A. Of the 3A markers screened, 18 were polymorphic and 6 showed linkage (Table 1). The most tightly linked marker was xgwm155, where only one recombinant was observed in 94 individuals.
The Dollar/MTHW9904 population was screened with 84 microsatellite markers all previously mapped to 3B. Seventeen were observed to have adequate polymorphism to be discerned on a 15% acrylamide gel. Six of the linked markers are listed in Table 1, with xgwm4010 being the most tightly linked, as three apparent recombinants were observed in 96 individuals (Table 1).
The Chinese Spring/MTHW9904 population was screened with 18 chromosome 3D microsatellites. Nine of these were polymorphic. The most tightly linked marker in this population was xgwm4306, with 10 apparent recombinant individuals observed among 197 individuals.
Figure 1 shows screening gels involving two white-seeded and two red-seeded genotypes, using the most closely linked microsatellite markers. In most cases, the red and white lines were polymorphic with all three markers (Fig. 1). However, screening of additional red-seeded spring and winter wheat lines has shown that some red-seeded genotypes were not distinguishable from white-seeded types for some of the markers (Talbert, Bruckner, and Sherman, unpublished, 2007).
Maps were generated for the most tightly linked markers and are shown in Fig. 2
. Map orders were consistent with existing maps (http://wheat.pw.usda.gov/cgi-bin/graingenes/browse.cgi).
View larger version (11K):
[in this window]
[in a new window]
|
Figure 2. Genetic maps of markers linked to red grain color (R-A1, R-B1, and R-D1 on chromosomes 3A, 3B, and 3D, respectively). Maps are oriented with markers that have been previously mapped toward the centromere, at top. Distances between markers are reported in centimorgans. Maps were generated from three separate crosses, each segregating for only one color locus (3A, ‘Rio Blanco’ [PI 531244]/Idaho 444; 3B, ‘Dollar’ [PI592117]/MTHW9904; 3D, ‘Chinese Spring’ [CItr 14108]/MTHW9904).
|
|
A validation population of 1786 F2 lines was established from the cross ‘Vida’ (PI 642366)/MTHW0471. Vida is a hard red spring wheat that appears from previous crosses to have all three dominant red alleles (Sherman and Talbert, unpublished, 2006), while MTHW0471 is an experimental hard white spring wheat. The F2 progeny were screened with the closest marker for each gene (xgwm155-3A, xgwm4010-3B and xgwm4306-3D, respectively) by the Western USDA-ARS genotyping laboratory, which has high-throughput capabilities. The PCR amplicon sizes, including tail, are as follows: gwm155: MTHW0471 145 bp, Vida 165 bp; gwm4010: MTHW0471 189 bp, Vida 213 bp; gwm4306: MTHW0471 253 bp, Vida 237 bp. These size estimates are more accurate than the size estimates obtained from polyacrlyamide gels (Fig. 1). However, amplicons without the tail are 19 base pairs smaller. Each individual was classified as homozygous for red or white parent markers or as a heterozygote. The different classes of homozygotes were expected to be equally represented (each with about 30 individuals). However, classes homozygous for the white parent D genome allele were underrepresented, while those with the red parent D genome allele were overrepresented (Table 2
). For example, the genotyping lab reported 16 individuals that were homozygous recessive at the three loci. The genotyping results (Table 2) were significantly different than expected (x2 = 27.5, 7 df; p < 0.001). The observation of fewer individuals with white parent banding patterns seems to be due to the poor amplification of the white allele using the tailed gwm4306. To test this hypothesis, using the DNA isolated by the genotyping lab with the nontailed gwm4306 primer, we reran 78 of the individuals, including 11 individuals previously determined to be homozygous for xgwm155 and xgwm4010. Through this process we identified 8 additional individuals for a total of 24 that were homozygous for all three white markers and confirmed the genotyping labs results for the other markers. All genotyped individuals had the expected phenotypes (Table 2). In other words, those individuals with red parent markers were red, and those 24 with only white parent markers produced white seed.
|
DISCUSSION
|
|---|
We have identified microsatellite markers that can be used to select for white alleles in crosses between red-seeded and white-seeded wheat genotypes. We initiated our search for linked markers by referring to previous maps, identifying flanking markers, and then using GrainGenes (http://wheat.pw.usda.gov) to identify all markers in the region. We then screened parents of mapping population for polymorphisms. Polymorphic markers were then run on complete populations. Our data correlates well with previous reports of the position of the red gene (Bailey et al., 1999; Groos et al., 2002; Osa et al., 2003; Himi et al., 2005). For example, the closest microsatellite markers to R-D1 (xgwm3 and xgwm314) reported by Groos et al. (2002) were also linked in our population, but not as tightly linked as xgwm4306. However, xgwm480, the most tightly linked microsatellite to R-A1 (Groos et al., 2002; Osa et al., 2003), was not polymorphic in our mapping population. A newly identified sequence-tagged sites–PCR marker for RA-1 in tetraploid wheat (Vasu et al., 2008) was also not polymorphic in our populations. The remaining previously reported flanking markers of R-A1 and R-B1 are restriction fragment length polymorphisms, which we did not run. Validation experiments suggest that the markers provide a reliable assay for the presence of the genes conferring kernel color. Thus, a backcrossing effort to convert superior red-seeded genotypes to white-seeded complements seems feasible. The tailed primer for xgwm4306 appeared to preferentially amplify one allele over another. However, we initiated a backcross program using a different white-seeded parent and have been able to select heterozygotes with all three primer pairs, using both tailed and untailed primers (Sherman and Talbert, unpublished, 2007).
There are a few caveats to the use of the identified markers in marker-assisted selection. First, as with all microsatellite markers, they are not polymorphic in all crosses of interest. The converse is also true where the observation of a polymorphism for a microsatellite does not ensure that the lines are segregating for a particular trait (Ellis et al., 2007). However, a range of flanking markers is given (Table 1, Fig. 2) and can be used to identify other previously mapped markers (Somers et al., 2004) that may be polymorphic in a given cross. A second caveat due to the nature of microsatellites is that a specific allele size is not diagnostic for the white allele. For example, all four banding patterns shown in Fig. 1 produced with xgwm155 are different even though two of the parents were white (MTHW0202 and MTHW0471) and two were red (Choteau and Vida), thus, emphasizing the need for empirical determination of polymorphism and banding pattern for each potential red/white parent combination. A third caveat for the use of the markers is that phenotyping for red versus white seed is not unambiguous. That is, individuals that are homozygous recessive at all three loci may appear to have dark kernels in some conditions (Wu et al., 1999; Matus-Cadiz et al., 2003), and some very light red-seeded individuals may be misclassified as white. Therefore, we suspect our estimate of linkage is not completely accurate in that it is likely that some phenotypic misclassification occurred. These caveats were a primary rationale for the validation population. Results from this population suggest that the linked markers worked efficiently for selecting white-seed genotypes.
The existence of three loci controlling kernel color provides a further complication in that red-seeded genotypes may contain one, two, or three loci with red alleles. Since the genotype for most red-seeded phenotypes is unknown, a lack of polymorphism between a red-seeded and white-seeded parent may in fact be due to the presence of the white allele at one of the loci in the red-seeded parent. The only reliable way to determine the genotype is through genetic analysis as described by a number of authors and summarized by McIntosh et al. (1998).
Perfect markers for seed color, where the polymorphism causing the functional difference between the alleles is the marker, would remove the above caveats for using markers to follow seed color. There is a likelihood that the kernel color locus will be cloned in the future (e.g., Himi et al., 2005; Vasu et al., 2008), allowing the production of such markers. However, the linked microsatellites will find utility for marker-assisted selection until that time and will remain useful for haplotyping once perfect markers are developed.
|
ACKNOWLEDGMENTS
|
|---|
The authors wish to acknowledge the support of USDA/CREES-NRICAP award 2006-55606-16629 and the Montana Wheat and Barley Committee.
|
NOTES
|
|---|
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.
Received for publication October 10, 2007.
|
REFERENCES
|
|---|
- Bailey, P.C., R.S. McKibbin, J.R. Lenton, M.J. Holdsworth, J.E. Flintham, and M.D. Gale. 1999. Genetic map locations for orthologous Vp1 genes in wheat and rice. Theor. Appl. Genet. 98:281–284.[CrossRef][Web of Science]
- Boutin, P., E.H. Hani, F. Vasseur, C. Roche, B. Bailleul, J. Hager, and P. Froguel. 1997. Automated fluorescence-based screening for mutation by SSCP: Use of universal M13 dye primers for labeling and detection. Biotechniques 23:358–362.[Web of Science][Medline]
- Ellis, M.H., A.E. David, G. Bonnett, A.E. Gregory, and J. Rebetzke. 2007. A 192bp allele at the Xgwm261 locus is not always associated with the Rht8 dwarfing gene in wheat (Triticum aestivum L.). Euphytica 157:209–214.[CrossRef][Web of Science]
- Groos, C., G. Gay, M.-R. Perretant, L. Gervais, M. Bernard, F. Dedryver, and G. Charmet. 2002. Study of the relationship between pre-harvest sprouting and grain color by quantitative trait loci analysis in a white x red grain bread-wheat cross. Theor. Appl. Genet. 104:39–47.[CrossRef][Medline]
- Himi, E., J.E. Flintham, and K. Noda. 2005. Red grain color gene (R) of wheat is a Myb-type transcription factor. Euphytica 143:239–242.[CrossRef][Web of Science]
- Lander, E.S., P. Green, J. Abrahamson, A. Barlow, M.J. Daly, S.E. Lincoln, and L. Newburg. 1987. MAPMAKER: An interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174–181.[CrossRef][Medline]
- Lin, W., and G. Vocke. 2004. Electronic output report from the Economic Research Service. Available at http://www.ers.usda.gov/publications/agoutlook/aug1998/ao253e.pdf (verified 10 May 2008). USDA, Washington, DC.
- Matus-Cadiz, M.A., P. Hucl, C.E. Perron, and R.T. Tyler. 2003. Genotype x environment interaction for grain color in hard white spring wheat. Crop Sci. 43:219–226.[Abstract/Free Full Text]
- McIntosh, R.A., G.E. Hart, K.M. Devos, M.D. Gale, and W.J. Rogers. 1998. Catalogue of gene symbols for wheat. p. 117. In Proc. 9th Int. Wheat Genetics Symp. Vol. 5. University Extension Press, University of Saskatchewan, Saskatoon, SK, Canada.
- Metzger, R.J., and B.A. Silbaugh. 1970. Location of genes for seed color in hexaploid wheat (Triticum aestivum L.). Crop Sci. 10:495–496.[Abstract/Free Full Text]
- Miskelly, D.M. 1984. Flour components affecting paste and noodle colour. J. Sci. Food Agric. 35:463–471.[CrossRef][Web of Science]
- Osa, M., K. Kato, M. Mori, C. Shindo, A. Torada, and H. Miura. 2003. Mapping QTLs for seed dormancy and the Vp1 homologue on chromosome 3A in wheat. Theor. Appl. Genet. 106:1491–1496.[Medline]
- Riede, C.R., and J.A. Anderson. 1996. Linkage of RFLP markers to an aluminum tolerance gene in wheat. Crop Sci. 36:905–909.[Abstract/Free Full Text]
- Roder, M.S., V. Korzun, K. Wendehake, J. Plashke, M.H. Tixier, P. Leroy, and M.W. Ganal. 1998. A microsatellite map of wheat. Genetics 149:2007–2023.[Abstract/Free Full Text]
- Somers, D.J., P. Isaac, and K. Edwards. 2004. A high-density microsatellite consensus map for bread wheat (Triticum aestivum L.). Theor. Appl. Genet. 109:1105–1114.[CrossRef][Web of Science][Medline]
- Taylor, M.R., G.W. Brester, and M. Boland. 2005. Hard white wheat and gold medal flour: General Mills contracting program. Rev. Agric. Econ. 27:117–129.[CrossRef]
- USDA. 2003. Alkali test to detect hard white/red wheat kernels. FGIS-PN-03-02. Available at http://archive.gipsa.usda.gov/reference-library/bulletins/pn03-02.pdf (verified 11 Apr. 2008). USDA Grain Inspection, Packers, and Stockyards Administration, Federal Grain Inspection Service, Washington, DC.
- Vasu, K., S. Sood, and B. Gill. 2008. Targeted genomic mapping of a red seed color gene (R-A1) in wheat. Crop Sci. 48:S37–S48.[Web of Science]
- Windes, J.M., E. Souza, D.W. Sunderman, and B. Goates. 1995. Registration of four dwarf bunt resistant wheat germplasm: Idaho 352, Idaho 364, Idaho 443, and Idaho 444. Crop Sci. 35:1239–1240.[Free Full Text]
- Wu, J., B.F. Carver, and C.L. Goad. 1999. Kernel color variability of hard white and hard red winter wheat. Crop Sci. 39:634–638.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
