| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS, Northern Crop Sci. Laboratory, Fargo, ND 58105 USA
steven.j.knapp{at}orst.edu
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
Abbreviations: B lines, sterility maintainer lines • R lines, fertility restorer lines • P1 and P2, parent inbred lines • PD, donor inbred lines • P1 x P2, single-cross hybrid to be enhanced • +, favorable allele • -, unfavorable allele
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
The percentage of crosses that lead to new parent inbred lines for single-cross hybrids is low (Hallauer, 1990), so any method for increasing the probability of selecting the most promising crosses (donors) should increase breeding efficiency. Dudley (1984, 1987) tackled this problem by developing a method for estimating the number of favorable dominant alleles in a donor inbred line that are not present in either parent of an elite single-cross (µG). Several statistics in addition to µG have been developed to compare and rank donor inbred lines as sources of new favorable alleles for traits where directional dominance (heterosis) is important (Dudley 1984, 1987; Gerloff and Smith, 1988a, 1988b; Bernardo, 1990; Metz, 1994). Bernardo (1990) proposed a statistic for selecting donor lines called the net gain of favorable alleles (NG). NG estimates the number of alleles that can be gained minus the number of alleles that can be lost during selection when PD is crossed to P1 (NG1) or P2 (NG2). Metz (1994) proposed a statistic called the probability of a net gain of favorable alleles (PNG). PNG estimates the number of loci where favorable alleles can be gained as a proportion of the number of loci where favorable alleles can be gained or lost during selection when PDD is crossed to P1 (PNG1) or P2 (PNG2). Finally, Dudley (1987) proposed statistics for selecting the parent to be crossed to the donor to enhance hybrid performance and to ascertain whether or not the elite parent should be backcrossed to the donor prior to selfing and selection.
Most of the experimental work on the donor selection problem has been done in maize (Zea mays L.) (Gerloff and Smith, 1988a, 1988b; Misevic, 1989a, 1989b; Zanoni and Dudley, 1989; Bernardo, 1990; Metz, 1994; Cartea et al., 1996; Malvar et al., 1998). The methods proposed by Dudley (1987) gave maize breeders a way to systematically screen donor germplasm and pinpoint sources of favorable alleles that are not present in the parents of elite hybrids. Although the rankings of donors are sometimes different with different statistics, the correlations between various donor statistics (µG, NG, and PNG) are usually positive and significant. More importantly, field tests of progeny from crosses between parent and donor inbred lines produced new parent lines superior to the original parent lines of the target hybrids (Misevic, 1989a, 1989b; Zanoni and Dudley, 1989; Hogan and Dudley, 1991).
Sunflower germplasm has not been systematically screened for class-G alleles using the statistics proposed by Dudley (1987) and others. One of the widely held tenets in sunflower research is that the genetic diversity of the elite gene pool is "narrow". Despite this, substantial seed yield increases might be produced by introgressing class-G alleles from elite donors into the parents of outstanding hybrids. Steady seed yield increases have been produced in sunflower by selecting among elite intraspecific progeny, and there are no data to suggest that additional gains cannot be made. The objectives of this study were (i) to assess the merits of a sample of public inbred lines as donors for increasing the seed yields of a few key single-cross hybrids and (ii) to propose donors and crosses for introgressing new favorable alleles from donor to parent inbred lines.
|
Materials and methods |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
Three B x R hybrids (HA383 x RHA373, HA372 x RHA377, and HA89 x RHA373) and one B x B hybrid (HA372 x HA383) were chosen as the hybrids to be enhanced (P1 x P2). Statistics for all possible B-line donors (10 for HA383 x RHA373, seven for HA372 x RHA377, and 10 for HA89 x RHA373) were estimated for each B x R hybrid. R line x R line crosses were not tested; thus, we only assessed B lines as donors for B x R hybrids. Statistics for three R-line donors and 10 B-line donors were estimated for HA372 x HA383.
Several statistics were estimated for each donor and target hybrid: (i) predicted three-way hybrid mean (PTC) (Sprague and Eberhart, 1977); (ii) the relative number of favorable dominant alleles present in PD that are not present in P1 or P2 (µG) (Dudley, 1987); (iii) the minimum upper bound on µG (UBND) (Gerloff and Smith, 1988a); (iv) the net gain of favorable alleles (NG) (Bernardo, 1990); and (v) the probability of a net gain of favorable alleles (PNG) (Metz, 1994). PTC was estimated by
, where 
1D is the least square mean for P1 x PD and 2D is the least square mean for P2 x PD. UBND was estimated by the minimum of 1D - 1 and 2D - 2), where 1 is the least square mean for P1 and 2 is the least square mean for P2 (Gerloff and Smith, 1988a). µG was estimated using the methods proposed by Dudley (1987), where µ is half the difference between genotypes fixed for favorable (++) and unfavorable (--) alleles and G is the number of loci for which PD is homozygous for favorable alleles and P1 and P2 are homozygous for unfavorable alleles.
The net gain of favorable dominant alleles for the cross between P1 and PD (NG1) was estimated by
, and the net gain of favorable dominant alleles for the cross between P2 and PD (NG2) was estimated by
, where 12 is the least square mean for the target hybrid (P1 x P2), D is the number of loci for which P1 is homozygous for favorable alleles and P2 and PD are homozygous for unfavorable alleles, and F is the number of loci for which P2 is homozygous for favorable alleles and P1 and PD are homozygous for unfavorable alleles (Bernardo, 1990). If 1D > 2D, then P2 should be crossed to the donor and P1 should be used as the tester and vice versa when 1D < 2D.
The probability of net gain of favorable dominant alleles for the cross between P1 and PD (PNG1) was estimated by µG/(µG + µD) and the probability of a net gain of favorable dominant alleles for the cross between P2 and PD (PNG2) was estimated by µG/(µG + µF) (Metz, 1994). If PNG1 > PNG2, then the cross between P1 and PD should be superior to the cross between P2 and PD. If PNG1 > PNG2, then P1 should be crossed to the donor and P2 should be used as the tester and vice versa when PNG1 < PNG2.
Two additional statistics (µC + µF and µD + µE) were estimated for each hybrid and donor inbred line combination by using the estimators proposed by Dudley (1987), where C is the number of loci for which P1 and PD are homozygous for favorable alleles and P2 is homozygous for unfavorable alleles, and E is the number of loci for which P2 and PD are homozygous for favorable alleles and P1 is homozygous for unfavorable alleles. If µC + µF > µD + µE, then the donor should be more closely related to P1 and should be crossed to P1 to develop new inbreds using P2 as the tester or vice versa if µC + µF < µD + µE (Dudley, 1984). Simple correlations were estimated between µG, UBND, PTC, the maximum of NG1 and NG2, the maximum of PNG1 and PNG2, and donor inbred performance per se for seed yield and plant height for the target B x R hybrids.
A similar statistical analysis was done for plant height for each of the B x R hybrids. The goal of this analysis was to select donors to decrease plant height or to hold plant height constant. When the goal is to select donors to decrease mean performance, class-G alleles are unfavorable (++) in donor inbreds, whereas class-D alleles are favorable (--) in donor inbreds when PD is to be crossed to P1 and class-F alleles are favorable (--) in donor inbreds when PD is to be crossed to P2 (Zanoni and Dudley, 1989). The probability of extracting a shorter line from P1 x PD was estimated by µD - µG, whereas the probability of extracting a shorter line from P2 x PD was estimated by µF - µG for each donor.
|
Results and discussion |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
|
The parent lines of the selected hybrids stood out among the inbreds tested (Cheres et al., 1999). HA383 and HA372 produced the first and second highest ranking hybrids with R-line testers and RHA373 and RHA274 produced the first and second highest ranking hybrids with B-line testers (Cheres et al., 1999) (Table 1). Among the ten highest-yielding B x R hybrids, two had HA383, three had HA372, and two had HA384 as the female parent, while six had RHA373 and three had RHA274 as the male parent. Although none of the hybrids we selected had HA384 or RHA274 as a parent, the seed yields of several B x R and B x B hybrids with HA384 or RHA274 and the selected parents were close to the seed yields of the selected hybrids. These hybrids were not chosen as target hybrids because they duplicated the heterotic patterns of the three target hybrids.
Donor Inbreds for Enhancing HA383 x RHA373
The most promising donors for enhancing the seed yield of HA383 x RHA373 were HA822, HA851, and HA372 (Table 2)
. These donors had the three highest µG and UBND estimates and were ranked the same using these two statistics (one, two, and three, respectively). PTC produced different ranks for these donors. HA372 was ranked first, HA822 was ranked fourth, and HA851 was ranked fifth. NG and PNG produced the same ranks as µG and UNBD for these three donors (Table 3)
. NG2 was greater than NG1 for HA822 and HA851 and slightly less than NG1 for HA372. Similarly, PNG2 was greater than PNG1 for HA822 and HA851 and equal to PNG1 for HA372. This suggests that the greatest progress in developing hybrids superior to HA383 x RHA373 can be made by extracting new inbreds from the cross of HA822 or HA851 to RHA373 and the cross of HA372 to either parent with the other parent as the tester. These data show that there are promising donors for enhancing both sides of this hybrid pedigree. The proposed crosses seem to have a nearly equal chance of producing new inbred lines superior to HA383 (with HA372 as the donor) or RHA373 (with any of the three as the donor) (Table 3).
|
|
Dudley (1984) proposed using (µC + µF) – (µD + µE) to assess whether a donor was more closely related to the female or male parent of a target hybrid. When µC + µF > µD + µE, PD is deemed to be more closely related to P1 than P2 and the heterotic pattern is preserved by crossing PD to P1 with P2 as the tester, and vice versa when µC + µF < µD + µE. µC + µF was greater than µD + µE for all of the donors tested for HA383 x RHA373 (Table 3). This suggests that all of the donors are more closely related to the female than the male parent of this hybrid and, using this as the sole criteria, that new inbred lines should be developed by crossing the selected donors to HA383 with RHA373 as the tester. These results have practical ramifications in sunflower breeding.
First, B lines as a whole tend to be more closely related to other B lines than R lines (Berry et al., 1994; Gentzbittel et al., 1994). The B lines we tested seem to be more closely related to each other than to R lines (Hongtrakul et al., 1997; Cheres and Knapp, 1998). However, as Cheres et al. (1999) showed, there are some discrepancies in the heterotic group assignments produced by analyses of pedigree, DNA fingerprint, and hybrid performance data. Some lines cannot be rigidly assigned to heterotic groups, and heterotic groups in sunflower tend to have greater within-group genetic distances than those described in maize.
Second, developing new female inbreds is simpler from B x B crosses because progeny from B x R crosses segregate for branching and fertility restorer genes. These factors must be weighed against the prospect of developing inbreds superior to one or the other parent. The differences between NG1 and NG2 or PNG1 and PNG2 were not great for the promising donors (Table 3). Even though this analysis suggests that progress can be made on either side of the hybrid pedigree, the most rapid and efficient strategy for developing new inbred lines is to select among progeny from donor x HA383 (B x B) as opposed to donor x RHA373 (B x R) crosses.
The greatest progress in developing superior inbred lines might be made by backcrossing to HA383 before selfing. Because µD and µF were greater than µG for all of the donors, one backcross to HA383 would increase the probability of retaining + alleles for class-D loci that could be lost in donor x HA383 crosses (Dudley, 1984). Similarly, one backcross to RHA373 would increase the probability of retaining + alleles for class-F loci that could be lost in donor x RHA373 crosses.
We assessed the merits of donors for reducing plant height. Seed yield and plant height were genetically correlated
among the inbred lines and hybrids we tested. Thus, selecting for increased seed yield may increase plant height in some crosses. This typically varies among crosses in sunflower (Miller et al., 1980) and could vary among the hybrids and donors we studied. Although height differences can be tolerated within a certain range, lodging susceptibility tends to increase as plant height increases. When other traits are equal, shorter hybrids are usually superior.
There were significant correlations between µG, PTC, UBND, NI, and PNG estimates for plant height and donor inbred line plant height per se (Table 4) . This suggests that height per se can be used to select donors for reducing plant height. Although heterosis was significant for plant height, most of the genetic variance for plant height was additive (Cheres et al., 1999). This has been reported in previous studies in sunflower (Miller et al., 1980). The µD - µG and µF - µG estimates for plant height were positive for most of the donors for HA383 x RHA373 (Table 5) . The µD - µG and µF - µG estimates were greatest for the shortest inbred (HA370), which was ranked seventh out of 10 donors for seed yield. The top three donors for seed yield (HA822, HA851, and HA372) (Tables 2 and 3) had negative µF - µG estimates for plant height (Table 5), whereas two of the top three donors for seed yield (HA851 and HA372) had negative µD - µG estimates for plant height. However, the most outstanding donor for seed yield (HA822) had a positive µD - µG estimate for plant height. Thus, HA822 seems to be an excellent donor for enhancing HA383 (increasing seed yield while decreasing plant height or holding plant height constant).
|
|
The NG estimates were negative and the PNG estimates were moderate and <0.5 for all of the donors, but were greater for HA821 and HA384 than the other donors (Table 3). µC + µF was greater than µD + µE for HA821 and HA384 (Table 4); thus, these donors seem to be more closely related to the female than the male parent. The most promising strategy for developing superior inbred lines should be to cross the selected donors (HA821 and HA384) to the female parent (HA372), particularly since the differences between NG1 and NG2 or PNG1 and PNG2 were negligible for these donors (Table 3). NG2 was much greater than NG1, PNG2 was slightly greater than PNG1, and µC + µF was slightly less than µD + µE for HA89. These statistics suggest that crossing HA89 to RHA377 has more merit than the reverse. This possibility should be explored because both sides of the HA372 x RHA377 could be enhanced (the male side using RHA377 x HA89 and the female side using HA821 x HA372 or HA384 x HA372). Because µD was greater than µG for all of the donors, one backcross to the female parent should increase the probability of retaining + alleles for class-D loci that could be lost in donor x HA372 crosses.
HA821 and HA384, the top two donors for increasing the seed yield of HA372 x RHA377, should not adversely affect plant height when crossed to HA372 (Table 5). µD - µG estimates for HA821 and HA384 were positive for plant height (Table 5). By contrast, µF - µG estimates for HA821 and HA384 were negative for plant height and were the only two negative estimates, suggesting that enhanced inbreds arising from crosses between these donors and RHA377 could produce hybrids taller than HA372 x RHA377. HA821 and HA384 are excellent candidates for increasing the seed yield of HA372 x RHA377 without increasing plant height.
Donor Inbreds for Enhancing HA89 x RHA373
The range of µG estimates for HA89 x RHA373 donors was wider than for the other two target hybrids (Table 2). This can be attributed to the lower seed yield of this hybrid (Table 1). µG, PTC, NG, and PNG ranked the top three donors (HA383, HA384, and HA821) the same, while UNBD ranked HA384 first, HA821 second, and HA383 third and donor seed yield ranked HA384 first, HA821 fourth, and HA383 fifth for enhancing the seed yield of HA89 x RHA373 (Tables 2 and 3). NG1 was substantially greater than NG2, PNG1 was substantially greater than PNG2, and µC + µF was greater than µD + µE for HA383. The NG1 and PNG1 estimates for this donor were greater than for the other donors. NG1 was greater than NG2, PNG1 was greater than PNG2, and µC + µF was greater than µD + µE for HA384 and HA821. These estimates predict that the greatest progress can be made by crossing the donors to the female parent and that the greatest progress can be made by crossing HA383 to HA89 (Table 3).
µG was much greater than µD for all three donors and much greater than µD for HA383. These data predict that the greatest gains can be made by selfing the crosses without backcrossing to HA89. Backcrossing to the parent line is usually ill-advised when the parent line is "inferior" to the donor (Dudley, 1984). The seed yield of HA89 x RHA373, for example, was significantly less than the seed yield of HA383 x RHA373; thus, HA89 is inferior to HA383 when crossed to the RHA373 tester. The analysis of HA89 x RHA373 as a target hybrid was done to cover a historically important and novel heterotic pattern. If the target hybrids had been selected on the basis of seed yield alone, HA89 x RHA373 would not have been chosen as one of the hybrids to be enhanced.
Most of the µD - µG and µF - µG estimates for plant height were negative for the top three donors (HA383, HA384, and HA821) for HA89 x RHA373 (Table 5). The only exception was the µF - µG estimate for HA821. Although enhanced inbreds arising from HA821 x RHA373 crosses would not necessarily increase plant height, this cross was not chosen to increase seed yield (Tables 2 and 3). HA89 x RHA373 was 5 cm shorter than HA383 x RHA373. New inbreds arising from the most promising cross for increasing seed yield (HA89 x HA383) would probably not produce hybrids taller than HA383 x RHA373.
A Strategy for Screening and Selecting Restorer-Line Donors
This study was originally designed to screen B lines as donors for enhancing the performance of sunflower hybrids. As such, the only branched lines we tested were the four R lines (all of the crosses were unbranched). This was done purposefully to eliminate the confounding effects of branching on seed yield. Crosses between donor and parent inbred lines had to be produced for this study. Thus, to test R lines as donors for B x R hybrids, we would have had to have produced and tested R x R hybrids. There are two problems with this. First, producing seed from R x R crosses is more difficult than from B x R or B x B crosses (B x R and B x B hybrids were made in this study using A lines as the females), even when genetic male-sterile R lines are used as females. Second, most R x R hybrids are branched and produce significantly less seed than B x R and B x B hybrids (B x R and B x B hybrids are unbranched). The morphological and seed yield differences between branched and unbranched hybrids would have produced misleading estimates of the merits of the donors and prevented us from accurately assessing the merits of R lines as donors for B x R hybrids. Despite this technical difficulty, a strategy for assessing R lines as donors emerged from this study.
We did not consider selecting B x B single-crosses as hybrids to be enhanced in the original experiment design because such hybrids cannot be commercially produced. We produced and tested B x B crosses because they were essential for the statistical analyses, but breeders would not normally produce B x B hybrids for hybrid yield testing. However, if a B x B single-cross were to be used as the hybrid to be enhanced, then R lines could be tested as donors because all of the crosses necessary for the donor analysis would be unbranched and the seed yields would not be confounded by the effects of branching. Several of the B x B hybrids we tested were among the top performers in Oregon and North Dakota and no difference was observed in the mean seed yield of B x B and B x R hybrids in this study (Cheres et al., 1999) (Table 1). There were, for example, six B x B hybrids among the top 10 yielding hybrids overall. We selected one (HA372 x HA383) as a target hybrid.
HA372 x HA383 (a B x B hybrid) was the highest-yielding hybrid in North Dakota (2032 kg ha-1) and ranked fifth for yield in Oregon (4072 kg ha-1) among all the hybrids (B x R and B x B) tested (Cheres et al., 1999). Even though this hybrid cannot be commercially produced and the parents are the females of the top two B x R hybrids we studied, this hybrid and other outstanding B x B hybrids can be used to screen R-line donors. We estimated statistics for all possible donors (B or R) for this hybrid. Three donors (RHA373, RHA274, and HA850) stood out among the three R-line and 10 B-line donors for HA372 x HA383 (Tables 2 and 3). PTC and µG ranked RHA373 first, RHA850 second, and RHA274 third. UNBD ranked these hybrids differently, but still identified them as the top three prospects. All three donors seem to be sources of class-G alleles. NG and PNG further differentiated the top three donors. RHA373 was the only donor with a positive NG estimate for HA372 x HA383 and had the highest PNG estimate among the donors tested (Table 3). NG1 was greater than NG2 and PNG1 was greater than PNG2 for this donor. Similarly, µC + µF was much greater than µD + µE for RHA373. These results suggest: (i) RHA373 is more closely related to HA372 than HA383; (ii) the prospects for developing a superior inbred line are greater from HA372 x RHA373 than HA383 x RHA373; and (iii) the seed yield of HA372 x HA383 can be increased by introgressing class-G alleles from RHA373 to HA372 (using HA383 as a tester). This process would entail selecting for branching and fertility restorer alleles in the segregating progeny. HA850 would not be a useful donor for this hybrid because all three lines (P1, P2, and PD) are females.
|
Summary and conclusions |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
This study concentrated on screening and selecting B-line donors (we sampled a small number of R lines). More work is needed to assess the merits of elite R lines (intraspecific R-line diversity) as donors of class-G alleles. This can be accomplished in practice by selecting a B x B single-cross as the hybrid to be enhanced. HA372 x HA383 is one of several outstanding B x B single-crosses that could be used for such an analysis. The R lines seem to be less diverse than B lines and thus are excellent targets for increasing seed and oil yields in sunflower.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
Received for publication August 31, 1998.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. M. Bertoia, R. Burak, and M. Torrecillas Identifying Inbred Lines Capable of Improving Ear and Stover Yield and Quality of Superior Silage Maize Hybrids Crop Sci., March 1, 2002; 42(2): 365 - 372. [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 | |||
, single-crosses between donor and parent inbred lines 
, and three target hybrids 
(HA383 x RHA373, HA372 x RHA377, and HA89 x RHA373) grown at Corvallis, OR and Casselton, ND in 1996 and 1997