| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a 310 Plant Sciences Bldg., Dep. of Plant and Soil Sciences, Univ. of Kentucky, Lexington, KY 40546-0312
b Dep. of Plant and Soil Sciences, Dep. of Statistics, Univ. of Kentucky, Lexington, KY, 40546
* Corresponding author (dli2{at}uky.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: BIL, backcross inbred lines • CIM, composite interval mapping • LG, linkage group • LRS, likelihood ratio statistic • MR, marker regression • PCR, polymerase chain reaction • QTL, quantitative trait loci • RIL, recombinant inbred line • SIM, simple interval mapping • SSR, simple sequence repeat
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The soybean germplasm collection may be a rich reservoir of alternative alleles. There are 18,688 accessions of G. max and 1166 G. soja accessions in the GRIN USDA soybean database (USDA-ARS, National Genetic Resources Program, 2007). A germplasm source that is collected but never used contributes as little to plant breeding progress as germplasm that is never collected. In the past the germplasm collections have been primarily used to find single gene sources of resistance to diseases and insects or tolerance to abiotic stresses. This has been true of all our major crops (Shands and Wiesner, 1991). But, as Tanksley and McCouch (1997) pointed out, the germplasm collections have been vastly underutilized as reservoirs of genes for improvement of quantitative agronomic traits such as yield.
Most traits important to agriculture, such as yield, are controlled by polygenes. It has been shown that wild species may have alleles of genes which positively influence agronomic traits (Tanksley and McCouch, 1997). Such favorable alleles might be beneficial if introduced into elite cultivars lacking such alleles.
The traditional approach to the utilization of exotic germplasm is to screen entries from a gene bank for a clearly defined character recognizable in the phenotype (Tanksley and McCouch, 1997). Ertl and Fehr (1985) developed two BC5 populations from G. max x G. soja matings. The F2 plants from the backcross populations were selected for the absence of deleterious phenotypes. Progeny of selected plants were evaluated for seed yield, maturity, lodging, and height in four environments during 2 yr. This method of introgression of G. soja alleles into G. max did not effectively increase yield potential.
Molecular markers are a powerful tool for breeders to search for new sources of variation and to investigate genetic factors controlling quantitatively inherited traits. Genetic linkage maps based on molecular markers have now been developed for most major crop species, including soybean (Song et al., 2004). Molecular linkage maps have made it possible to identify, map, and study the effects of the individual loci that control a quantitatively inherited trait.
The identification of favorable G. soja alleles is difficult for a quantitatively inherited trait (e.g., yield), because many undesirable and deleterious alleles will also segregate in populations derived from G. max x G. soja matings. The use of advanced backcross populations has been proposed as a means to solve this problem (Tanksley and Nelson, 1996). This method couples molecular marker analysis and phenotypic evaluation in BC2 or BC3 populations. The marker data are then used to select lines from the next generation, isolating the quantitative trait loci (QTL) region while eliminating or minimizing the contribution of the wild species in all other chromosome regions.
A number of researchers have evaluated G. soja as a source of useful genes. A G. soja allele with increased seed yield from PI 407305 was identified in a BC2 population developed by the advanced backcross method (Concibido et al., 2003). That G. soja allele was associated with a 9.3% yield increase across testing environments. The QTL was backcrossed into six elite soybean genetic backgrounds and retested. A significant (P < 0.05) positive QTL effect on yield was observed in two of the six backgrounds.
The objective of this study was to identify, locate, and validate, in an elite soybean genetic background, yield-enhancing QTL derived from a G. soja accession, PI 245331.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Field Trials and Trait Evaluation
Performance trials of the 295 BC2F4 BILs were conducted in 2003 and 2004 at Lexington and Princeton, KY (03Lex, 04Lex, 03Prin, and 04Prin). The soil types were Maury silt loam (fine, mixed, semiactive, mesic Typic Paleudalf) at Lexington, and Crider silt loam (fine-silty, mixed, active, mesic Typic Paleudalf) at Princeton. Lines from Populations A and B were each planted according to a split plot design repeated over locations with a one-way treatment classification (genotype), three maturity sets per environment, and two replications within maturity set at each of the two locations in both years. In Population A, there were 49 lines in each of the three maturity sets. In Population B, there were 49 lines in maturity set 1 and 2, and 50 lines in maturity set 3. Plot size was two 6-m rows with 0.76 m between rows. The plot was end trimmed to 5 m before harvest.
Six agronomic traits were measured: yield, seed size, height, lodging, seed filling period, and maturity. Height (cm) was measured as the distance from the soil to the tip of the main stem at maturity. Duration of seed filling was measured at Lexington as the number of days from R5 (beginning seed fill) date to the date of R7 (physiological maturity) (Fehr and Caviness, 1977). The growth stages were determined on 10 consecutive plants in one row of the two-row plots, and the plot was considered to be at a given growth stage when five or more of the 10 plants reached that stage. The 10 plants were identified before growth stage R5, and all measurements were made on the same 10 plants. The plots were observed once a week before R5 and three times a week between R6 to R7 (Egli et al., 1984). The R5 date was determined by interpolation of the weekly growth stage data collected before R5. Harvest maturity date was recorded as days from 31 August to the date when 95% of the pods had reached the final pod color. Lodging was rated on a 1 to 5 scale with 1 designated as plants standing erect and 5 as plants prostrate. Seed yield (kg ha–1) adjusted to 13% seed water content was measured. A 100-seed sample was drawn and used to obtain a measurement of 100-seed weight (g).
Marker Analysis
Leaf tissue DNA was extracted from 10 greenhouse-grown seedlings per BC2F4 line according to Stefaniak et al. (2006). More than 500 simple sequence repeat (SSR) markers (USDA-ARS and Iowa State University, 2003) covering the 20 chromosomes of the soybean genome were screened against the parents. The polymerase chain reaction (PCR) was conducted as defined on SoyBase (USDA-ARS and Iowa State University, 2003). Amplified PCR fragments were separated by either metaphor agarose or polyacrylamide gel electrophoresis, depending on the size of the polymorphism between the two parents. In the end, 147 Population A BILs were genotyped with 120 SSRs. The PI 245331 parent has black pods (L1L1) and 7499 has brown pods (l1l1). This phenotypic marker was also recorded in Population A. The observed G. max/G. soja allele segregation ratios were evaluated for each of the 121 individual markers and for all markers on each of the 20 linkage groups (LGs) using the chi-square test with 
= 0.05.
Statistical Analysis
Analysis of variance (ANOVA) was conducted on each measured trait, using the PROC MIXED functions of SAS (Statistical Analysis System version 8.0, SAS Institute, Cary, NC). Locations and years were treated as fixed effects; replication and BILs as random effects.
In 2003 at Lexington, the plots of the QTL mapping (A) population were infected with Tobacco ringspot virus producing bud blight infection with an accompanying pod abortion. Extensive yield loss was observed on many plots, and a total yield loss was observed on 42 of the 295 plots. The reliability of the Lexington 2003 yield data was tested using a correlation analysis of the line means between environment 03Lex and the other three environments: 04Lex, 03Prin, and 04Prin. Pearson correlation coefficients for yield were calculated between environment pairs.
To improve the reliability of the seed filling period and maturity data collected in 03Lex, plots with severe bud blight disease were discarded. Disease was most serious on end plants in a row. Healthy plants were chosen from the middle of the row, and seed filling period and maturity notes were taken only on the healthy plants. To test the reliability of the 03Lex seed filling period and maturity data, a correlation analysis was conducted on the line means between environment 03Lex and 04Lex.
QTL Analysis
Combined environment QTL analysis was done using the marker regression (MR), simple interval mapping (SIM), and composite interval mapping (CIM) functions of Map Manager QTXb.20 software (Meer et al., 2003). For single environment QTL analysis, the MR function was used.
A genomewise likelihood ratio statistic (LRS) criterion for evaluating the statistical significance of QTL effects for each trait was estimated for MR, SIM, and CIM by permutation (Churchill and Doerge, 1994). For the BC2F4 population, this LRS criterion was determined to be about 13.8 [i.e., LOD score = LRS(13.8) x 0.217 = 3.0] for the MR analyses, based on N = 1000 permutation tests conducted with the yield, height, lodging, seed weight, seed filling period, and maturity data. Quantitative trait loci with 2.0 < LOD < 3.0 were used as background markers in the CIM analysis. For seed yield QTL identification with CIM analysis, the background markers used were Satt050 on LG A1 and Satt285 on LG J. For 100 seed weight, the background markers used were Satt304 on LG B2, Satt324 and Satt594 on LG G, Satt192 on LG H, and Satt549 on LG N. For seed filling period, the background markers used were Satt334 on LG F, Satt284 on LG L, and Satt584 on LG N. For lodging, the background markers used were Satt594 on LG G, Satt330 on LG I, Satt652 on LG L, and Satt549 on LG N. For the SIM and CIM QTL analysis with combined data, 1000 permutation tests were used to define an empirical detection threshold at an experiment-wise P value <0.01 (corresponding to a SIM mapping LOD >3.6 or a CIM LOD >3.8). The mean map position of the significant QTL was established by the map position of the peak LOD score in the interval between two flanking markers. A one-LOD fall-off (from the QTL peak) method was used to estimate the left and right flanking map positions of a confidence interval surrounding the mean QTL map position (Chaky, 2003).
Validation Strategy
To confirm in a second test the identified QTL associated with PI 245331 alleles found in Population A, another 148 BC2F4 lines (Population B) were used. Population B and Population A are related with one line in each population tracing back to the same BC2 plant. Markers identified as significant in Population A were analyzed in Population B. The ANOVAs were conducted in Population B to calculate the QTL effects and QTL x environment interaction effects. After a positive QTL was identified in Population A, it was verified in Population B by using a less stringent criterion (P < 0.25), because instead of identifying an unknown QTL in the whole soybean genome the identified QTL in Population A were tested for significance in Population B.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A correlation analysis was conducted on the line means between environment 03Lex and 04Lex for seed filling period and maturity. Correlations between 03Lex and 04Lex were r = 0.34 (P < 0.0001) for seed filling period and r = 0.66 (P < 0.0001) for maturity. Based on the highly significant positive correlations, we concluded that the seed filling period and maturity data from 03Lex could be used in the QTL discovery analysis.
There was significant (P < 0.01) genetic variance among BILs for yield, 100 seed weight, plant height, and plant lodging in all single environments and the combined environment data (Table 1 ). Maturity set effects were not significant for any of these four traits at single and combined environments except plant height in 04Lex. Environment had a significant effect on all the traits except seed yield. Variance among lines was significant for both maturity and seed filling period at combined environments. Maturity set effects were significant for maturity at P < 0.01 level and significant for seed filling period at P < 0.05 level for the combined data.
|
= 0.05). None of the chi-square values were significant ( = 0.05) for the 20 LGs. The constructed genetic map for this study contained 121 markers, for a total length of 1506 cM (Haldane) or 50% of the soybean genome (estimated to be approximately 3000 cM), with an average interval size of 12.5 cM. These 121 markers coalesced into 20 different LGs, leaving no markers in the unlinked class. The LGs were designated with names corresponding to the integrated public soybean genetic map (Cregan et al., 1999; USDA-ARS and Iowa State University, 2003).
Calculated marker orders for the SSRs were identical to the published maps for nine of the LGs: B1, D1b, D2, G, H, I, M, N, and O. The mapping order of the SSR markers for the other 11 LGs was in large part similar to the public soybean genetic map. Three kinds of differences, however, were observed: (i) the inversion of two markers closely linked to one another, (ii) the inversion of two marker groups when a large gap exists between these two marker groups, (iii) simultaneous occurrence of both inversions. Possible reasons for this are first, the QTL mapping population size is not very large; second, an advanced backcross population is not the ideal population for constructing a genetic map; and third, the number of markers on each LG was too few to allow highly accurate mapping.
Seven gaps of 25 cM or more existed between pairs of markers in this population. The maximum distance separating two markers was approximately 68 cM on LG A2.
Quantitative trait loci mapping analysis was conducted using the orders and genetic distances from these calculated linkage maps and also from the composite linkage maps (Cregan et al., 1999, USDA-ARS and Iowa State University, 2003). Since QTL identification was marker order neutral, the SSR marker alignments in each of the 20 LGs equivalent to the soybean map of LGs published by Cregan et al. (1999) and SoyBase (USDA-ARS and Iowa State University, 2003) were used. By adjusting the marker orders for LGs A1, A2, B2, C1, C2, D1a, E, F, J, K, and L, the Map Manager software calculated a new genetic distance when gene orders were changed.
QTL Analysis
Quantitative trait loci were identified for all six traits in this study (Tables 2
, 3
, 4
). A total of 11 QTL were detected above an empirically determined experiment-wise significance threshold equivalent to P < 0.0001 (corresponding to LOD > 3.0 for marker regression) and P < 0.01 (corresponding to a SIM LOD > 3.6 or a CIM LOD > 3.8) for the combined environments. An experiment-wise significance threshold of P < 0.05 (LOD > 3.0 for SIM analysis) was used to declare a QTL significant for single environments. One more QTL was identified when single environment data was analyzed.
|
|
|
|
100-Seed Weight
Two QTL for seed size were identified from the combined data by MR analysis. These QTL were located at Satt304 on LG B2 (LOD = 3.3) and at Satt541 on LG H (LOD = 4.6) (Table 2). The G. soja allele decreased seed weight for both QTL. The more significant QTL was mapped to LG H, between markers Satt192 and Satt541, with a mean position of 34 cM, and an additive effect –1.2 g (LOD = 4.2) on the substitution of a G. soja allele for G. max allele (Table 3). The phenotypic variance accounted for by this QTL was 12%. One more QTL for seed size was identified from the combined data by SIM analysis: the QTL on LG N between markers Satt549 and Satt339, with a mean position of 41 cM and an additive effect of –1.0 g (Table 3). The significant 100-seed weight QTL on LG B2 had CIM-generated LOD scores that were no longer significant. In contrast, the 100-seed weight QTL on LG H remained significant (LOD = 5.0) in the CIM analysis (Table 4). However, the QTL mean position for this QTL shifted rightward compared to the position determined by SIM. The seed size QTL on LG H was significant in each individual environment. No environment specific seed size QTL were identified.
Seed Filling Period
One seed filling period QTL was identified from the combined data by MR analysis: the QTL at Satt335 on LG F (LOD = 4.0) (Table 2). It was mapped between markers Satt335 and Satt334, with a mean position of 73 cM, and an additive seed filling period effect of 1.3 d on the substitution of a G. soja allele for a G. max allele (Table 3). The phenotypic variance accounted for by this QTL was 11%. The significant seed filling period QTL on LG F remained significant in the CIM analysis, with a similar map position and increased additive effects (Table 4). This QTL was found to be significant only in 03Lex (LOD = 3.9), but positive additive effects of this QTL were also observed in 04Lex (Table 5).
Maturity
Two QTL were detected for maturity by MR analysis, with the G. soja allele increasing maturity at Satt335 on LG F and decreasing maturity at Satt584 on LG N (Table 2). The QTL at Satt335 has a positive additive effect of 1.4 d (LOD = 3.3) with R2 = 10%. The QTL at Satt584 has a negative additive effect of –1.4 d (LOD = 3.2) with R2 = 9%. Single environment MR analysis of maturity data identified the QTL at Satt335 on LG F with positive additive effects in both environments, but it was significant only in 04Lex (LOD = 4.0) (Table 5). Similarly, the QTL at Satt584 exhibited negative additive effects in both environments, significant only in 04Lex (Table 5). No significant maturity QTL were identified by SIM and CIM analysis for the combined environments.
Lodging
Three lodging QTL were identified for the combined data by MR analysis. For all of these QTL the G. soja allele increased lodging. The QTL with the largest effect, Satt284 on LG L (LOD = 5.2), remained significant in SIM and CIM analysis (Table 3 and 4). This QTL was significant in each individual environment (Table 5).
Height
One height QTL was mapped to LG L by SIM analysis for the combined data between markers Satt652 and Satt284, with a mean position of 5 cM, and an additive height effect of 5.2 cm (LOD = 3.6) on the substitution of a G. soja allele for a G. max allele (Table 3). The phenotypic variance accounted for by this QTL was 11%. This marker allele was also associated with an increase in lodging scores (Table 3). The high lodging score is likely caused by the tall plants, but more research is needed to dissect whether these multiple trait associations are the result of pleiotropy or genetic linkage.
QTL Confirmation
To validate the QTL effects found from PI 245331 in Population A, another 148 BC2F4 lines (Population B) were used as a confirmation population. Population B and Population A are related with one line in each population tracing back to the same BC2 plant. Each line in Population B started with 75% of its genotype identical with the genotype of a line in Population A. Genotypic differences were possible at the 25% of the BC2 genome heterozygous at the time of line divergence. Thus Population B was similar but not identical to Population A.
Analysis of variances were conducted in Population B at each single environment and combined environments to validate the QTL1 (Satt511) and QTL2 (Satt529) effects. Three of four locations showed significant (P < 0.25) QTL effects at Satt511 (QTL1). The yield increase for 03Lex, 04Lex, and 04Prin was 6.9, 7.6, and 12.8%, respectively (Table 6 ). Environment 03Prin had a yield decrease of –1.2%, but it was not significant (Table 6). For the combined environments, lines that were homozygous for the PI 245331 allele at the QTL locus Satt511 had a significant effect (P = 0.08) (Table 7 ). It demonstrated a 6.3% yield increase (P = 0.037) over lines that were homozygous for the G. max 7499 allele (Table 6). Two out of four environments showed significant (P < 0.25) QTL effects at Satt529 on LG J (QTL2). For the combined analysis, the QTL effect was significant (P < 0.25) (Table 7). The genotype mean for lines that were homozygous for the PI 245331 allele at the QTL locus Satt529 showed a significant yield increase (P = 0.03) only in environment 03Lex (Table 6). This QTL also increased yield in environments 03Prin and 04Prin, but the increases were not significant. For the combined analysis, the genotype mean for lines with the G. soja allele demonstrated a 3.7% yield increase, but it was not significant (Table 6).
|
|
Genotype means and class differences for combined genotypes at Satt511 and Satt529 were calculated across the testing environments in Population B (Table 8 ). Their genotype means for the four combined classes mm, ms, sm, and ss, were 3055, 2734, 3209, and 3402 kg ha–1, respectively. The genotype ss had the maximum yield compared with the other three genotypes. The class differences of mm-sm, mm-ss, ms-ss, and sm-ss were all significantly different (P < 0.25).
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this study, permutation tests defined an empirical detection threshold at an experiment-wise P value <0.0001 for MR analysis and P value <0.01 for the combined SIM and CIM QTL analysis. After a positive QTL was identified in Population A, the QTL was verified in Population B by using a less stringent criterion (P < 0.25). This was because, instead of identifying an unknown QTL in the whole soybean genome, the QTL identified in Population A were tested for functionality in Population B. Once a QTL was discovered in Population A we were conservative about rejecting it in Population B.
A limited population size in QTL detection experiments may lead to underestimation of QTL number (Beavis, 1994). Another advanced backcross introgression population reported by Concibido et al. (2003) was 70% larger than the present 147 line BC2F4 population. The original goal was to produce a 300-line population, but this was not successful due to the difficulties of backcrossing.
Quantitative trait loci discovery in advanced backcross populations has both negative and positive features (Kaeppler, 1997). Negatively, the recurrent parent marker class mean becomes more precisely estimated as the number of backcrosses increases, but the donor parent marker class mean becomes less precisely estimated due to the increasing imbalance of the number of members in each group. With soybean, it is difficult to produce a large quantity of BC1 and BC2 seed and an adequately sized mapping population needed in the advanced backcross method. Positively, increased similarity within the population from backcrossing may reduce error variance and improve QTL detection. Assessing the wild species alleles in the elite parent genetic background reduces the impact of epistasis among wild species alleles and improves individual QTL characterization in an elite genetic background.
By using an advanced backcross population we identified two positive yield QTL from G. soja. One was found in the combined environments and one in the environment 04Prin. In this population segregation was not skewed for any of the 121 markers. By delaying QTL analysis until an advanced generation, one would be less likely to detect QTL with epistatic effects. Also, associated deleterious effects due to linkage drag are less likely to be observed (Tanksley and Nelson, 1996). In this study, the two yield QTL marked by Satt511 (QTL1) and Satt529 (QTL2) appear to have an interactive effect in addition to the additive effect (Table 7). As shown in Table 8, genotype mean differences were not significant for mm-ms and ms-sm in either Population A or Population B. In both populations lines with sm yielded significantly more than mm lines, and the genotype mean of ss is significantly greater than mm, ms, and sm in both populations. QTL1 and QTL2 not only have additive effects, but the interaction effect also exists between them (Table 7). Putting these two QTL together is better than each single QTL alone even when the G. soja allele at Satt529 does not have a significant effect.
Concibido et al. (2003) used SSR and amplified fragment length polymorphism markers and the advanced backcross method of QTL mapping to identify a yield QTL that was associated with increased seed yield from G. soja PI 407305. The yield QTL marked by AFLPU3944 on LG B2 provided a significant 9.3% yield increase across the test environments. More markers were assayed in this B2 chromosome region, and the QTL mapped closest to marker Satt066. To further test the QTL effect, lines containing the QTL were backcrossed into six genetic backgrounds. In two out of the six backgrounds, significant (P < 0.05) positive effects on yield were observed for the QTL allele from G. soja. This suggests the potential of using exotic germplasm to improve soybean yield.
Kabelka et al. (2004) evaluated two soybean PIs as sources of alleles for the enhancement of seed yield in North American cultivars. They identified 15 QTL for seed yield. At 9 of the 15 QTL the PI alleles increased seed yield. One positive yield QTL (LOD = 2.6) from the PI was identified within their maturity set one marked by Satt225 on LG A1 (95.2 cM) (Table 9 ). This QTL explained 14% of the phenotypic variation and increased seed yield 60 kg ha–1 (2.1%) across 12 environments. This Satt225 marker allele also was associated with an increase in days to maturity and with taller plants. The correlation coefficients between seed yield and these traits were small, and seed yield was not significantly different between maturity sets. Here, we report a positive yield QTL from G. soja, marked by Satt511 on LG A1 (94.2 cM) with an additive yield effect of 191 to 235 kg ha–1 depending on the QTL analysis method, accounting for 12 to 13% of the phenotypic variance of the combined data. This genetic region around Satt511 was not associated with maturity, and there was no obvious segregation for plant height or any other traits with this allele in our population. Even though these two QTL (Satt225, Kabelka et al. (2004); Satt511, this report) have different expressions for maturity and height, crossing will be necessary to determine if they segregate separately (are different QTL) or if the same QTL has been identified in both studies.
|
Some QTL were significant for more than one trait. For example the marker allele Satt335 on LG F was associated with an increase in days to maturity and also was associated with an increase in days of seed filling period (Table 2). The Pearson correlation coefficient between seed maturity and seed filling period in this experiment is 0.39 (P < 0.01). It is likely that the longer seed filling period is caused by the later maturity. Further genetic dissection of the region containing this QTL would be needed to distinguish between pleiotropy or gene linkage.
Examples have been reported that QTL identification and subsequent assessment of marker-assisted selection occurred successfully in different genetic backgrounds and similar environments (Concibido et al., 2003; Sebolt et al., 2000). However, the value of these yield QTL in a broad array of genetic backgrounds and environments is not known. Reyna and Sneller (2001) assessed the value of three yield QTL from the northern United States, cultivar Archer for increased yield in southern U.S. environments, and genetic backgrounds. But, following backcrossing none of the marker effects was significant for any of the three QTL for any trait. The authors concluded that it may be difficult to capture the value assigned to QTL alleles derived from diverse parents when the alleles are introgressed into different genetic backgrounds and subsequently tested in different environments.
It is also important to show that the G. soja alleles are unique and not present in the G. max germplasm. Sebolt et al. (2000) tried to evaluate the effect of one G. soja protein QTL in three genetic backgrounds. The test populations were developed by crossing a line from the initial mapping population with the cultivars Parker and Kenwood and the experimental high protein line C1914. The cultivars are high yielding and are representative of cultivars in commercial production. The protein QTL allele from G. soja was associated with an increase in protein concentration in the Parker and Kenwood populations but not in the C1914 population. The authors concluded that there may be a gene conferring increased protein concentration in C1914 but not in Parker or Kenwood that is allelic with the G. soja gene.
To confirm the yield QTL identified in this study in other genetic backgrounds, BILs containing these QTL should be crossed to soybean cultivars that are high yielding and of diverse parentage. Using cultivars in commercial production will ensure confirmation in the best current genetic material and will also put the confirmed alleles into a useful germplasm source. These BILs should also be crossed with the G. soja yield allele containing HS1 line deposited in the American Type Collection (Concibido et al., 2003) to determine whether the alleles are additive.
More experiments will be needed to test these yield QTL in a broader array of genetic backgrounds and in different environments. Also, closely linked markers or flanking markers to the QTL are needed in future research, to avoid recombination between the markers and the QTL. If recombination occurs between a marker and a QTL, then selection based on the marker will not be effective. Nearby markers are available for the yield QTL on LG A1. Polymorphism and linkage will be tested in future experiments. At this time favorable alleles have been identified from G. soja. The availability of more SSR markers around the QTL region makes it a good candidate for marker-assisted breeding in the future.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Received for publication June 26, 2007.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
