| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Plant Agriculture, Crop Science Division, Univ. of Guelph, Guelph, ON N1G 2W1
b Agriculture and Agri-Food Canada, 2585 Highway 20 East, Harrow, ON N0R 1G0
c Ridgetown College, Univ. of Guelph, Ridgetown, ON N0P 2C0
d Guelph Center for Functional Foods, Laboratory Services Division, Univ. of Guelph, 95 Stone Road West, Guelph, ON N1H 8J7
e Agriculture and Agri-Food Canada, 1391 Sandford Street, London, ON N5V 4T3
* Corresponding author (irajcan{at}uoguelph.ca)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Breeding soybean for high or low isoflavone content in the seed is challenging since isoflavone content is heavily influenced by environment (Eldridge and Kwolek, 1983; Wang and Murphy, 1994; Carrao-Panizzi and Kitamura, 1995; Hoeck et al., 2000; Lee et al., 2002). Since molecular markers are not affected by environment, the genetic dissection and characterization of many quantitatively inherited agronomic and seed quality traits in soybean is possible. A limited number of researchers have attempted to map quantitative trait loci (QTL) associated with isoflavone content in soybean seeds. Currently, QTL associated with this trait have been reported on LGs A1, B1, B2, D1a+Q, H, K, and N in the mapping population Essex x Forrest (Njiti et al., 1999; Meksem et al., 2001; Kassem et al., 2004). Our mapping population was derived from germplasm adapted to southern Ontario environments, which could assist in identifying additional QTL and perhaps confirm some of the previously identified QTL. The use of these QTL through MAS could be advantageous in the design of an efficient and cost-effective breeding strategy for developing high or low isoflavone soybean varieties.
Epistasis is also thought to contribute to quantitative variation (Cheverud and Routman, 1995). Most mapping studies have had little success detecting epistasis because of small population size, type of population, limited analysis software, and use of an appropriate threshold level for declaring epistasis significant (Lukens and Doebley, 1999; McMullen et al., 2001). However, significant epistatic interactions have been detected in soybean for the traits height (Lark et al., 1995) and yield (Orf et al., 1999) and in other crops such as oat, Avena sativa L. (Holland et al., 1997; Zhu et al., 2004), rice (Li et al., 1997), and maize, Zea mays L. (McMullen et al., 2001). Currently, there is no information on epistatic interactions associated with isoflavone content in soybean seeds.
The objective of this study was to identify and characterize QTL affecting daidzein, genistein, glycitein, and total isoflavone content by means of a large recombinant inbred line (RIL) population derived from the cross AC756 x RCAT Angora grown in two environments. The possibility of epistatic interactions between pairs of loci affecting isoflavone accumulation was also investigated.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1417 µg g–1) was considered the low and RCAT Angora (2246 µg g–1) was considered the high isoflavone parent on the basis of 3 yr data (1995 to 1997) (Drs. Chung-Ja Jackson and Vaino Poysa, personal communication, 2000). The F1 plants were self-pollinated and F2 seeds were advanced for two generations by single-seed descent (SSD) to the F4 population in the growth room. In 2000, the F4 population was grown in the field at Harrow, ON, and F5 recombinant inbred lines (RILs) established by harvesting each plant (n = 207) individually. The RIL population was sent to a winter nursery at Los Andes, Chile, for seed increase. In 2001, the F4:6 seeds together with parents and five maturity check varieties (OAC Kent, OAC Walton, OAC Oxford, OAC Arthur, and RCAT Staples) were planted at Harrow and Talbotville, ON, Canada as a 16 x 16 simple lattice design with two replications. The plots at Harrow consisted of two rows, each trimmed to 4 m long, with 60 cm between rows within the plots and 60 cm between adjacent plots. Plots were sprayed with Dual II Magnum (Syngenta Crop Protection Canada, Guelph, ON) at 1.75 L ha–1, Pursuit (BASF Canada, Toronto, ON) at 0.310 L ha–1, and Sencor (Bayer CropScience, Guelph, ON) at 0.6 L ha–1. The soil type at Harrow was Fox sandy loam and was fertilized with 0–0–60 (N–P–K), applied at 120 kg ha–1. The plots at Talbotville consisted of two rows, each trimmed to 5.3 m long, with 35 cm between rows within plots and 40 cm between adjacent plots. Plots were sprayed with Pursuit at 0.312 L ha–1, Basagran Forte (BASF Canada, Toronto, ON) at 1.75 L ha–1, and Assist (BASF, Toronto, ON) at 1L per 1000 L of water on June 18. On June 28, plots were sprayed with Pinnacle (Dupont Canada, Mississauga, ON) at 6 g ha–1 and Basagran Forte at 1.75 L ha–1. The soil type at Talbotville was Tavistock/Gobles loamy.
Data Collection
The following agronomic traits were measured on all plots at each location: seed yield, days to maturity, plant lodging, and plant height. Seed quality, seed weight, oil content, and protein content were measured on an entry mean basis at each location by bulking approximately 150 g of seed from each plot. Seed yield was converted to kg ha–1 and adjusted to 130 g kg–1 (13%) moisture. Days to maturity were recorded when 95% of the pods matured. Lodging was scored from 1 (all plants in a plot erect) to 5 (all plants in a plot prostrate). Plant height was estimated as the distance from the soil surface to the tip of the main stem for a representative plant. Seed weight was recorded as the weight of 100 randomly selected seeds from a bulk at each location. Seed quality was rated from 1 (seed surface smooth with no evidence of shriveling, disease-free) to 5 (seed very shriveled, cracked seeds, discoloration, evidence of disease). Approximately 300 g of seed was used to measure oil and protein contents by using near-infrared reflectance (NIR) on a GrainSpec B1126 from FOSS North America Incorporated (Brampton, ON).
Isoflavone Extraction and HPLC Analysis
The 2001 soybean seed samples (10 g) were ground in a coffee grinder and sent to Laboratory Services (University of Guelph, Guelph, ON, Canada) for daidzein, genistein, glycitein, and total isoflavone analysis. Isoflavone concentrations were determined by HPLC as described by Vyn et al. (2002). Briefly, 0.5-g samples were mixed with 10 mL of ethanol and 2 mL of concentrated HCl. The mixtures were hydrolyzed by heating at 125°C for 2 h in a sand bath. The samples were allowed to cool down at room temperature and then centrifuged at 1500 x g for 10 min. The clear aliquot was filtered through a 0.45-µm PTFE filter. The samples were analyzed for isoflavone content on an HPLC under the following instrumental conditions; mobile phases were solvent A (4% v/v aqueous acetic acid) and solvent B (100% methanol); solvent system (% solvent A/% solvent B), 0 min (70/30), 12.5 min (65/35), 13 min (50/50), 15 min (30/70), 22.5 min (25/75), and 23 min (70/30); flow rate, 1.5 mL min–1; and injection volume, 5 µL.
DNA Extraction
Remnant F5–derived RIL seeds were used to extract genomic DNA from the 207 RILs. A seed from each RIL was planted in the growth room. Leaf discs were obtained from a single-hole punch modified to fit a 1.5 mL screw-cap tube and were stored at –80°C. Genomic DNA was extracted from five leaf discs of each parent and RIL with the FastDNA Kit (BIO 101, Carlsbad, CA). The DNA samples used in PCR reactions were diluted 1:100 (1 µL DNA in 99 µL distilled and deionized water) and stored at 4°C.
Molecular Analysis
PCR reactions were adapted from the guidelines developed for the soybean SSR collection (Cregan et al., 1999). The 15-µL PCR reaction contained 3 µL of diluted genomic DNA template, 0.2 units of Platinum Taq DNA Polymerase (Invitrogen Canada Inc., Burlington, ON), 1.5 µL of 10x PCR buffer [20 mM Tris-HCl (pH 8.4), 50 mM KCl], 1.5 µL of 2.5 mM MgCl2, 0.6 µL of 5 mM dNTP (Amersham Pharmacia Biotech, Piscataway, NJ), 2 µL each of 2.25 µmol of the forward and reverse SSR primers, and 4.2 µL of water. SSR primers were synthesized by Laboratory Services (University of Guelph, Guelph, ON) and Sigma Genosys (Oakville, ON).
PCR reactions were performed in a 96-well RoboCycler (Stratagene, La Jolla, CA) under the following thermal sequence: 2 min at 95°C, followed by 40 cycles of 45 s denaturation at 92°C, 45 s annealing at 47°C, and 45 s extension at 68°C. A final extension at 72°C for 5 min completed the reaction. Amplified PCR products were separated by electrophoresis on 5% (w/v) MetaPhor agarose (Bio Whittaker Molecular Applications, Rockland, ME) gels with a Sunrise 96 Horizontal Electrophoresis Apparatus (Gibco BRL, Life Technologies, Carlsbad, CA) with 115–118 mA of current supplied by an EC105 Electrophoresis Power Supply (ThermoEC, Holbrook, NY). DNA bands were visualized under UV light (365 nm) by staining with ethidium bromide. Additional screening was performed when needed by polyacrylamide gel electrophoresis using a 6% (w/v) gel in a Dual Adjustable Slab Gel System (C.B.S. Scientific, Del Mar, CA) following the method described by Wang et al. (2003a).
The parental lines were screened for polymorphisms with 458 SSR primer pairs that were selected for coverage of all soybean linkage groups using the public soybean genetic map (Cregan et al., 1999).
Data Analysis
Entry means were adjusted for lattice block effects for daidzein, glycitein, genistein, and total isoflavone content in the Harrow trials by PROC MIXED procedure in SAS version 8.2 (SAS Institute, 1988). In addition, maturity was used as a covariate to adjust entry means for daidzein, genistein, and total isoflavone content but not glycitein. Neither adjustment was done for the Talbotville trials since entries were bulked from two reps for trait analysis. Broad-sense heritabilities were estimated by conducting a linear regression analysis of the mean progeny value (n = 207) from individual environments with the parental values (RILs grown at Harrow, ON, 2000).
Chi-square analysis was used to detect significant (P < 0.01) deviation of genotypic classes from the expected 1:1 Mendelian segregation ratio. A linkage map, based on the 207 RILs and 99 SSR markers, was generated by MAPMAKER/EXP version 3.0 (Lander et al., 1987). The commands "group," "map," "sequence," "lod table," "try," and "compare" were used for building the linkage groups. The error detection ratio was set at 1%. The Kosambi mapping function was used with a minimum LOD score of 3.0 and a maximum distance of 50 cM.
QTL were identified by single-factor analysis of variance (PROC GLM, SAS Institute, 1988) on individual environment values. Locus main effects were considered for linear additive models without epistasis if they were significant at P < 0.01. Significant loci on the same LG were tested by two-factor analysis of variance without interactions. If both loci were significant at P < 0.05 in the two-factor model, they would both be considered for linear additive models. Otherwise, the locus with the larger individual R2 value was chosen to represent the effect of the putative QTL on the LG. If two loci remained significant in a two-factor model, both loci were considered for linear additive models.
Significant (P < 0.001) two-way epistatic interactions were identified by running EPISTACY 2.0 (Holland, 1998) with SAS version 8.2 (SAS Institute, 1988) on individual environment values. Input of data was performed according to Holland (1998). Significance level was chosen on the basis of previous reports and a manageable number of interactions. Significant interactions were considered for linear additive models. Trait data and marker map were simultaneously analyzed with MAPMAKER/QTL version 1.1 (Lander et al., 1987) and Windows QTL Cartographer version 2.0 (Wang et al., 2003b) to identify the approximate position of QTL within intervals. A LOD score of 2.0 was used as a minimum to declare the presence of a QTL in a particular genomic region.
Multiple Locus Model Analysis
Main effect loci significant at the P < 0.01 level were considered for multiple locus models excluding epistasis. Backward regression analysis was used to eliminate loci individually, using all possible combinations with other markers until all remaining loci in the model were significant at P < 0.01. The model with the highest R2 was selected as the best model without epistasis.
Main effect loci that remained significant in the multiple locus models without epistasis were considered for models including epistasis. All significant interaction terms (including their main effects) were included in the model. Backward regression analysis was used to eliminate interaction terms or main effect loci until all factors in the model including epistasis remained significant at P < 0.01.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
An analysis of variance of isoflavone content indicated that there was significant (P < 0.0001) genotypic variation for daidzein, genistein, glycitein, and total isoflavone content among the RILs. Significant variation was also detected between environments (P < 0.0001), indicating that the data collected for the two environments could not be pooled and analyzed as a single data set. In fact, isoflavone contents were significantly higher at Harrow than at Talbotville. Therefore, the QTL mapping analyses were performed on entry means for each environment.
Correlations involving isoflavone content and agronomic and seed quality traits are presented by environment in Table 2. Significant and positive correlations were found between all isoflavone groups. Seed yield, maturity, plant height, and lodging were all significantly and positively correlated with daidzein, genistein, and total isoflavone. Glycitein was negatively correlated with maturity and plant height at Talbotville but was positively correlated with plant height at Harrow. Genistein and glycitein were positively correlated with oil content at Talbotville. The negative correlations between protein content and all isoflavones were significant in both environments. Glycitein was negatively correlated with seed weight in both environments, whereas daidzein, genistein, and total isoflavone were positively correlated with the trait at Talbotville. Seed quality was not significantly correlated with any of the traits.
|
The relative mapping order of the SSR markers within LGs was in large part similar to the public soybean genetic map; however, two major differences were observed. First, the distance between markers Satt387 and Satt521 on LG N was relatively large on our linkage map (50 cM), in comparison to the public soybean genetic linkage map (12 cM). Segregation distortion, which can cause changes in map distances (Lu et al., 2002) was observed for markers Satt387 and Satt521. The second major difference was related to markers Satt512 and AT21 on LG E, which are currently not included on the public soybean map. This increases the total coverage of the public soybean genetic linkage map. Gijzen et al. (2003) and Reinprecht (2003) also mapped marker AT21 to LG E.
QTL Associated with Individual and Total Isoflavone Content
QTL associated with individual and total isoflavone content were identified by one-way analysis of variance (ANOVA) (Table 3). Results from the one-way ANOVA were used to describe QTL identified for individual traits because it captured all QTL including those associated with unlinked markers. Environment affected the magnitude of the allelic effects but not the direction, and for all isoflavone traits, both parents contributed favorable alleles.
|
A large number of the QTL with the smallest effects became nonsignificant when included in multiple locus models (Table 3). Models generally included QTL with the largest effects; however, some models did include QTL with small effects. Only one significant locus was detected for glycitein and total isoflavone at Talbotville, and therefore, multiple locus models were not available. Multiple locus models included two to five significant loci depending on the trait and environment. The total explained phenotypic variation for multiple locus models ranged from 7.5 to 25.0%.
IM and CIM also identified several marker associations (LOD score >2) with daidzein, genistein, glycitein, and total isoflavone content as demonstrated through one-way ANOVA (Table 3). However, IM and CIM were not able to confirm unlinked loci significantly associated with individual and total isoflavone content since these two methods are dependent on at least two loci being linked. IM and CIM both identified QTL associated with isoflavones on LGs H, J, and M (Fig. 2a and 2b) . In addition, the two methods identified two distinct peaks on LG M for each of daidzein, genistein, and total isoflavone suggesting that two different QTL might be located on this LG. CIM identified QTL locations on LGs A1, C2, D1a, F, and G (Fig. 2b), which were not identified by IM. IM identified a QTL region on LG K, which was not identified by CIM. QTL identified by IM and CIM require further validation and should be treated as probable regions.
|
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
50% of the SSR markers listed on the public soybean genetic map (Cregan et al., 1999) and to the 13 unlinked markers. Mapping additional SSR markers should improve genome coverage and increase the ability to identify more QTL and epistatic interactions associated with isoflavone content. Of the 99 markers tested, 17 showed highly significant associations (P < 0.01) with isoflavone content (Table 3). These 17 QTL were located on 11 independent linkage groups (A1, A2, C2, D1a, F, G, H, J, K, M, and N). Individually, these markers explained from 3.5 to 10.5% of the phenotypic variation for specific isoflavone content in this population. The low level of variation explained by these QTL was indicative of the quantitative nature of isoflavone inheritance in soybean seeds. However, by combining the individual effects in multiple locus models (Table 3), more of the phenotypic variation was explained (R2 ranged from 7.5–25.0). The QTL identified in our study might be useful in a MAS program but require further testing in several environments and years, as well as other genetic backgrounds, before implementing MAS strategies. Considerable attention should be given to QTL on LGs J and M because they were detected across environments, and remained significant in multiple locus models with and without epistasis.
Isoflavone QTL have been previously identified in a mapping population derived from the cross Essex x Forest in which the linkage map encompassed 2990 cM and contained 201 SSR markers (Njiti et al., 1999; Meksem et al., 2001; Kassem et al., 2004). Twelve of the 17 QTL detected in this study were located in genomic regions different from the other mapping studies which suggested that AC756 and RCAT Angora contain isoflavone alleles different from Essex and Forest (Table 6). Interestingly, five of the QTL were identified in genomic regions on LGs A1, D1a, H, K, and N, similar to the other mapping studies. Soybean breeders should also focus their efforts on these five QTL because they were identified in different genetic backgrounds and environments, suggesting their usefulness in breeding programs aimed at increasing isoflavone content in soybean seeds.
|
Individual and total Isoflavone contents were significantly (P < 0.0001) higher at Harrow than at Talbotville, a result which is consistent with the large environmental interaction generally associated with isoflavone content in soybean seeds (Eldridge and Kwolek, 1983; Kitamura et al., 1991; Wang and Murphy, 1994; Carrao-Panizzi and Kitamura, 1995; Hoeck et al., 2000). Detection of QTL for isoflavones in soybean seeds was a reflection of this environmental effect as 11 and three QTL were identified at Harrow and Talbotville, respectively. Inconsistent detection of QTL is common in mapping studies (Paterson et al., 1991; Veldboom and Lee, 1996a; Veldboom and Lee, 1996b; Lee et al., 1996; Chapman et al., 2003), which Beavis (1994) proposed as QTL being expressed only under certain environmental conditions. If this is true, the QTL identified in this study could be useful for MAS under a particular environment.
Measurement of the additive effects of the QTL detected in this study was possible because the RIL were bred to near homozygosity, minimizing potential dominant gene action, and leaving only additive gene action and their interaction. For all isoflavones, QTL with opposite effects were observed. Combining alleles with effects in the same direction and from different parental sources would allow trait values to exceed parental values. This complimentary gene action would explain RIL with isoflavone values exceeding the parental values. In comparing the additive effects of the QTL detected in the mean environment and individual environments, the parental contribution was always the same. However, the magnitude of the additive effects changed likely because of genotype x environment interactions. In plant breeding programs, it is desirable to use QTL involving changes in magnitude rather than effect because one can still produce gains from selection.
Traits that are correlated may have loci in common manifested through linkage or pleiotropy (Aastveit and Aastveit, 1993). Genetic correlations observed among isoflavones and agronomic and seed quality traits (Table 2) were investigated further by comparing our QTL mapping results for isoflavones as well as agronomic (seed yield, maturity, plant height, and lodging) and seed quality (oil, protein, and seed weight) traits (data not shown). All traits shared a common genomic region on LG M. It is possible that different QTL conditioning these traits are inherited in clusters as tightly linked loci. Alternatively, it is known that in soybean, these traits are interrelated through maturity, and that the region on LG M could represent a major maturity gene. Fine mapping could resolve this issue.
Genomic regions shared by isoflavones were also identified on LGs A1, C2, and H, which could be explained by a key step that interrelates the synthesis of these constituents in soybean seed. Isoflavones also shared a common locus with the traits seed yield and plant height on LG A1. Furthermore, the alleles that were significantly associated with an increase in isoflavone content were always associated with an increase in seed yield and plant height, which is consistent with previous studies (Wang et al., 2000; Meksem et al., 2001; Kassem et al., 2004; Primomo et al., 2005).
Protein content shared common regions with genistein on LG A1, and with glycitein on LGs F, J, and K. Moreover, alleles associated with an increase in genistein content on LG A1 and with an increase in glycitein content on LGs F and K were associated with a decrease in protein content, suggesting negative correlation between isoflavone and protein content (Chiari et al., 2004; Primomo et al., 2005). In contrast, the allele associated with glycitein increase on LG J was positively associated with protein content, which suggested that it should be possible to breed soybean seeds with increased isoflavone and protein content, and support the findings of Primomo et al. (2005).
The oil content shared a genomic region with genistein on LGA1 and glycitein on LG N, whereas seed weight shared regions with these two isoflavones on LGs A1, C2, and F. A greater map density or a larger mapping population would be necessary to resolve whether those regions containing multiple QTL are controlled by the same gene or by closely linked genes.
The importance of epistasis in determining isoflavone content in soybean seed has not been investigated. Epistasis could play an important role in MAS since QTL effects may change when incorporated into different genetic environments. Furthermore, epistatic interactions could also assist in explaining the genetic control of isoflavone accumulation in soybean seeds, and thus facilitate the identification of additional QTL associated with isoflavone. A total of 23 epistatic interactions were detected for isoflavones (Table 4), of which 14 (61%) did not involve any marker individually associated with isoflavones found in this study. Interactions that involved one or two QTL may require additional analysis in near isogenic lines (NILs) to determine how large of an effect these interactions have on isoflavone content.
Three important observations were made when multiple locus models with epistasis were considered (Table 5). First, QTL that remained significant were generally the ones that were identified by one-way ANOVA as being highly significant and with the largest effects. More specifically, QTL associated with isoflavones on LGs J, K, and M remained significant in models with epistasis. It provided further evidence that these QTL could be incorporated into different genetic backgrounds because epistasis did not seem to have any significant consequences on their effects. Second, the locus Sat_135 was initially nonsignificant by one-way ANOVA; however, it became significant (P < 0.01) when its main effect and interaction with another locus were included in multiple locus models with epistasis. Third, including epistatic terms into multiple locus models explained a greater portion of the phenotypic variation (15.8–35.8%) than without, suggesting that isoflavone content in soybean seeds may also be determined by gene–gene interactions. The variation that remains unexplained may be attributed to QTLs or epistatic effects not detected in this study because of incomplete genome coverage (39%), environment, or weak linkage relationships between marker loci and QTLs.
Marker-assisted selection could be a powerful approach for breeding desirable amounts of individual or total isoflavones in soybean seeds. QTL have been identified for daidzein, genistein, glycitein, and total isoflavone content that show consistent effects in diverse environments. Although some of the genomic regions explained a small portion of the genotypic variation, or were identified only in a specific environment, they could be vital to understanding the genetic control of isoflavone accumulation in soybean seeds. Evaluation of these QTL in distinct environments and in different genetic backgrounds would help to validate the findings of this study. The regions of the soybean genome that were associated with isoflavones are candidates for changing isoflavone profiles in high yielding soybean lines and are likely not influenced by background effects because epistatic interactions did not seem to have significant effects on these regions. Future research should focus on narrowing down genomic regions, verifying QTL across years and in different genetic backgrounds, transferring of these QTL alleles into high yielding soybean varieties through backcrossing, and further analysis of epistatic interactions in NILs.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Received for publication November 22, 2004.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S.K. St. Martin, F.-t. Xie, H.-j. Zhang, W. Zhang, and X.-j. Song Epistasis for Quantitative Traits in Crosses between Soybean Lines from China and the United States Crop Sci., January 28, 2009; 49(1): 20 - 28. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Dhaubhadel, M. Gijzen, P. Moy, and M. Farhangkhoee Transcriptome Analysis Reveals a Critical Role of CHS7 and CHS8 Genes for Isoflavonoid Synthesis in Soybean Seeds Plant Physiology, January 1, 2007; 143(1): 326 - 338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Iwashina, E. R. Benitez, and R. Takahashi Analysis of Flavonoids in Pubescence of Soybean Near-isogenic Lines for Pubescence Color Loci J. Hered., September 1, 2006; 97(5): 438 - 443. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
