Published in Crop Sci. 44:1231-1235 (2004).
© 2004 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
CROP BREEDING, GENETICS & CYTOLOGY
Primitive Accession Derived Germplasm by Cultivar Crosses as Sources for Cotton Improvement
II. Genetic Effects and Genotypic Values
Jack C. McCartya,*,
Johnie N. Jenkinsa and
Jixiang Wub
a USDA-ARS, Crop Science Research Laboratory, P.O. Box 5367, Mississippi State, MS 39762
b Department of Plant and Soil Sciences, Mississippi State University, Mississippi State, MS 39762
* Corresponding author (JMcCarty{at}msa-msstate.ars.usda.gov).
 |
ABSTRACT
|
|---|
Primitive accession derived germplasm of cotton, Gossypium hirsutum L., may provide useful traits for cultivar improvement. The ability to predict advanced generation performance when crossed with commercial cultivars would enhance their utility and encourage their use in breeding programs. Our objective for this study was to predict genetic effects for day-neutral derived lines derived from primitive accessions and crossed to cultivars using a mixed linear model approach. Parents and F2 populations were grown at two field locations in 1998 and 1999 and parents and F3s were grown at two locations in 2000. Lint yield, yield components, and fiber quality traits were evaluated. An additive-dominance additive x additive (ADAA) model was used for genetic analysis. A mixed linear model, minimum norm quadratic unbiased estimation (MINQUE) was used to predict genetic effects and genotype values. Generally, the female cultivar parents had higher additive effects for lint yield and lint percentage; however, these females generally had lower additive effects for fiber strength. Significant AA effects widely existed among parents and F2 populations for lint percentage, boll weight, and fiber strength. The correlation coefficients between observed values and predicted values were mainly high among traits and environments. These data suggest that it is appropriate to use the ADAA genetic model to predict genetic effects and hybrid genotypic values for advanced generations. Our study showed that fiber strength may be significantly improved over that of the female parents, while the lint yield was slightly but not significantly predicted to be less than their female parents. This study suggested that day-neutral primitive germplasm accessions provided a valuable gene resource for selecting high yielding lines with significantly improved fiber strength.
Abbreviations: AA, additive x additive effect • AD, additive dominance model • ADAA, additive-dominance, additive x additive model • MINQUE, minimum norm quadratic unbiased estimation
|
INTRODUCTION
|
|---|
IMPROVING COTTON FIBER QUALITY and lint yield remains challenging for cotton breeders. Many of the current high-yielding, commercial upland cultivars do not possess the fiber quality desired by the textile industry. In developing new cultivars, it is important to utilize variability from diverse plant genetic resources. This can limit vulnerability to pests and diseases, while providing useful variation that can be used to form new favorable genetic combinations. However, Van Esbroeck and Bowman (1998) observed that parental genetic diversity, as estimated by coefficient of parentage, was not imperative for cotton improvement. To improve breeding efficiency when using diverse germplasm in developing high-yielding and acceptable fiber-quality cultivars, it is important to understand the genetic effects of these traits.
Primitive accessions of cotton have been reported to have useful genetic variability (Percival, 1987; McCarty et al., 1995, 1998a, 1998b). Their use has been limited because most require short days to initiate flowers and produce fruit. This requires crosses to be made in a greenhouse or in a tropical nursery. A backcross breeding program has been in place for a number of years to incorporate day-neutral genes into the primitive accessions so they will flower and can be crossed in most cotton breeding nurseries (McCarty et al., 1979; McCarty and Jenkins, 1993). Day-neutral F5, BC1F5, BC2F5, BC3F5, and BC4F5 progenies were evaluated for several agronomic and fiber traits and useful genetic variability for these traits were reported by McCarty et al. (1995)( 1998a, 1998b).
One of the ways to improve fiber quality and cotton yield is to transfer genes into high yielding cultivars from primitive accession germplasm which possess variability for these traits. In this research, crosses were made between five high yielding cultivars and 14 day-neutral derived germplasm lines at Mississippi State, MS (McCarty et al., 2000, 2003). The objective of this study was to predict genetic effects and genotypic values in different generations for agronomic and fiber traits. This research provides insight into how to better utilize these accessions in breeding programs.
|
MATERIALS AND METHODS
|
|---|
Plant Material and Experimental Design
Day-neutral lines have been developed from photoperiodic primitive accessions. Lines have been selected that have good fiber strength and contain variability for other traits. Fourteen primitive, accession-derived lines were crossed to five cultivars in 1997. Parents and their F2 or F3 populations were grown in field plots in 1998 through 2000 and evaluated for yield and fiber quality (McCarty et al., 2000, 2003).
An additive-dominance additive x additive (ADAA) and G x E interaction genetic model was employed for data analysis (Zhu, 1994). A mixed model, minimum norm quadratic unbiased estimation (MINQUE) approach was used to estimate genetic variance components. The genetic effects and genotype values were predicted by the Adjusted Unbiased Prediction (AUP) method (Zhu, 1993; Zhu and Weir, 1996). The prediction equations for parents and hybrids are listed as follows:
At a specific environment for a parent:
Over environments for a parent:
At a specific environment for predicted hybrid at generation n (Fn):
Over environments for predicted hybrid at generation n (Fn):
All effects are defined in McCarty et al. (2004).
Jackknifing over blocks within environments was used to estimate standard errors and the predicted effects (Miller, 1974). The degrees of freedom were 19 and t tests were used for testing significance of each parameter studied. The data set was analyzed by a program written in C++.
|
RESULTS AND DISCUSSION
|
|---|
Genetic Effects
On the basis of the ADAA genetic model, additive, dominance, additive x additive epistasis, and genotype x environment (GxE) genetic effects were predicted by the AUP method for agronomic traits, micronaire reading, elongation, and fiber strength. Since cotton breeders are more interested in additive and additive x additive epistasis effects, which are important for pure line improvement, only these genetic effects will be addressed in this study. The predicted additive effects for agronomic and fiber traits are summarized in Table 1 and additive x additive epistasis effects are summarized in Table 2.
View this table:
[in this window]
[in a new window]
|
Table 1. Predicted additive effects for lint yield (LYLD, kg ha–1), lint percent, boll weight (BW, g), micronaire reading (Mic), elongation (E1, %), strength (T1, kNm kg–1), 50% span length (SL50, mm) and 2.5% span length (SL2.5, mm).
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Predicted A*A epistasis effects for lint yield (LYLD, kg ha–1), lint percent, boll weight (BW, g), micronaire (Mic), elongation (E1, %), and strength (T1, kNm kg–1) .
|
|
Generally, the female commercial cultivar parents, P1-P5 had higher lint yield, and lint percentage than the primitive derived male parents, P6-P19; however, these female parents generally had fewer positive additive effects for fiber strength (Table 1). Results indicate that P1, P2, P3, and P5 can be used as parents for improving lint yield, while the use of P8, P9, P11, P12, P13, P17, P18, and P19 as parents will decrease lint yield. These results indicate that to improve yield, commercial cultivars are still better sources for the short term; however, for improving yield components that may lead to future yield improvements, primitive derived germplasm can be a valuable source. For example an additive effect for lint percentage was significantly detected for all parents except P9. In addition to the five commercial cultivars, P18 and P19 can be used for lint percentage improvement. Thirteen out of 19 parents had significant additive effects for boll weight. P1, P4, P10, P11, P12, P13, and P19 can be used for developing larger boll lines, while P6, P9, P17, and P18 can be used to develop smaller boll lines. For improving fiber strength, parents 8, 9, 11, 12, 14, 15, 16, and 17 are good choices. For improving 2.5% span length, parents 1, 4, 5, 14, 17, and 19 are good choices. For improving fiber elongation, parents 1, 2, 5, 6, and 18 are good choices.
Just as additive effects are transferable to the offspring progenies, so are additive x additive epistasis effects. Thus, additive x additive epistasis effects are also important for selection of pure lines in a breeding program. All agronomic traits and fiber traits, except fiber span length, expressed additive x additive epistasis effects. One parent (P15) expressed significant negative AA effects for several different traits, while 12 crosses expressed significant AA effects for the same trait. Significant AA effects widely exist among parents and F2 populations for lint percentage, boll weight, and fiber strength (Table 2). For example, 19 parents and 31 crosses expressed significant AA effects for lint percentage; 12 parents and 32 crosses expressed significant AA effects for boll weight; 15 parents and 22 crosses expressed significant AA effects for fiber strength. Not all significant effects were desirable because some move the trait in an undesirable direction.
If the AA interaction effect between two parents is greater than the mean AA interaction effect within two parents [AAij – (AAii + Ajj)/2 > 0], the genotypic value for the hybrid between the two parents in later generations will be potentially greater than the mid-parent genotypic values. Our results showed that AAij is numerically greater than (AAii + Ajj)/2 for most crosses (66 out of 70) for lint yield. Therefore, cotton yield heterosis for these crosses can be fixed at later generation without selection. Additionally, these F2 populations can be used for pure line selection in early generations. Other crosses showing AAij effects greater than zero relative to the mean of (AAii + Ajj)/2 included, 14 crosses for lint percentage, 55 for boll weight, 41 for micronaire, 31 for elongation, and 37 for fiber strength. The existence of large AAij effects indicate there is a genetic basis for improving yield and fiber quality by fixing heterosis in advanced generations. For example, crosses 1 x 15, 1 x 19, 4 x 9, 5 x 6, and 5 x 10 could be used to increase yield by exploiting heterosis and/or selecting high yielding lines (Table 2); while crosses 2 x 11, 3 x 7, 3 x 14, and 4 x 17 could be used for selecting high strength lines.
Predicted Hybrid Genotypic Values
The predicted genetic effects were used to calculate hybrid values from F1 through F6 for each environment based on the ADAA model for all traits except fiber span length (AD model) under the assumption of no selection pressure. To detect the efficiency of prediction, we compared the predicted F2 genotypic values to observed F2 values for Environments 2, 3, and 4 for agronomic traits, and for Environments 1, 2, 3, and 4 for fiber traits. The predicted F3 genotypic values were compared to observed F3 values for Environments 5 and 6 for agronomic traits, and at Environment 5 for fiber traits. The correlation coefficients between observed and predicted values are summarized in Table 3. The results indicated that predicted hybrid values were in good agreement with observed hybrid values for all traits except micronaire reading and 50% span length. Since no additive by additive epistasis effects and small AA effects were previously detected for 50% span length and 2.5% span length using the ADAA model, it was not appropriate to use this model to predict these two traits, so a reduced model (AD model) was used to provide estimates for these two traits; however, our findings suggest that the use of the ADAA model was appropriate for all other traits.
View this table:
[in this window]
[in a new window]
|
Table 3. Correlation coefficients between predicted F2 and F3 genotypic values and F2 and F3 observed values for certain environments.
|
|
The results indicate that we may be able to use predicted genotypic values based on statistical analyses to make decisions in our breeding program in early generations. The predicted F6 genotype values with their mid-parent values for lint yield and fiber strength are provided for each cross in Table 4. There was no cross better than the high parent for lint yield (P5) or for fiber strength (P8); however, 66 crosses were numerically greater than mid-parent values for lint yield, 38 crosses were 10% greater than mid-parent values, and 13 crosses were 20% greater than mid-parent values for the lint yield. Most predicted F6 crosses were significantly greater than their corresponding male parents and were not significantly less than their corresponding female parents (commercial cultivars or better parents) with respect to their predicted lint yield (Tables 4 and 5). On the other hand, most crosses were significantly greater than their corresponding female parents with respect to the predicted fiber strength genotype values. For example, in the predicted F6 generation, crosses 5 x 7, 5 x 10, 5 x 15, and 5 x 16 made high lint yields, which were not significantly less than the female parent 5, while the fiber strength for these crosses was significantly improved compared with their female Parent 5. Similar examples can be also found in Tables 4 and 5. Our study suggested that fiber strength can be significantly improved from their female parents, while lint yields were not significantly less than their female parents for many crosses. This study provided strong evidence that crosses between commercial cultivars and primitive derived lines can produce progeny with improved fiber strength and good yields.
View this table:
[in this window]
[in a new window]
|
Table 4. Predicted genotype values for lint yield (LYLD, kg ha–1) and fiber strength (T1, kNm kg–1) at F6 generation for 70 crosses (5 females by 14 males).
|
|
View this table:
[in this window]
[in a new window]
|
Table 5. Predicted parent values and standard errors across environments for lint yield (LYLD, kg ha–1) and fiber strength (T1, kNm kg–1).
|
|
|
NOTES
|
|---|
Contribution of the USDA-ARS in cooperation with the Mississippi Agric. and Forestry Exp. Stn.
Received for publication July 14, 2003.
|
REFERENCES
|
|---|
- McCarty, J.C., Jr., and J.N. Jenkins. 1993. Registration of 79 day-neutral primitive cotton germplasm lines. Crop Sci. 33:351.[Free Full Text]
- McCarty, J.C., Jr. J. N., Jenkins, and B. Tang. 1995. Primitive cotton germplasm: Variability for yield and fiber traits. Miss. Agric. and Forestry Exp. Stn. Tech. Bull. 202.
- McCarty, J.C., Jr., J.N. Jenkins, W.L. Parrott, and R.G. Creech. 1979. The conversion of photoperiodic primitive race stocks of cotton to day-neutral stocks. Miss. Agric. and Forestry Exp. Res. Rep. 4(19):4.
- McCarty, J.C., J.N. Jenkins, and J. Wu. 2000. Evaluation of cultivars by race stock F2 hybrids. p. 517–518. In Proc. Beltwide Cotton Res. Conf., San Antonio, TX. 4–8 Jan. 2000. Natl. Cotton Council Am., Memphis, TN.
- McCarty, J.C., J.N. Jenkins, and J. Wu. 2003. Use of primitive accessions of cotton as sources of genes for improving yield components and fiber properties. Mississippi Agric. and For. Exp. Stn. Bull. 1130. (Available on-line at http://msucares.com/pubs/bulletins/b1130.pdf; verified 5 March 2004).
- McCarty, J.C., J.N. Jenkins, and J. Wu. 2004. Primitive accession derived germplasm by cultivar crosses as sources for cotton improvement: I. Phenotypic values and variance components. Crop Sci. 44:1226–1230 (this issue).[Abstract/Free Full Text]
- McCarty, J.C., Jr., J.N. Jenkins, and J. Zhu. 1998a. Introgression of day-neutral genes in primitive cotton accessions: I. Genetic variances and correlations. Crop Sci. 38:1425–1428.[Abstract/Free Full Text]
- McCarty, J.C., Jr., J.N. Jenkins, and J. Zhu. 1998b. Introgression of day-neutral genes in primitive cotton accessions: II. Predicted genetic effects. Crop Sci. 38:1428–1431.[Abstract/Free Full Text]
- Miller, R.G. 1974. The jackknife-A review. Biometrika 61:1–15.[Abstract/Free Full Text]
- Percival, A.E. 1987. The national collection of Gossypium germplasm. So. Coop. Series Bull. 321.
- Van Esbroeck, G., and D.T. Bowman. 1998. Cotton germplasm diversity and its importance to cultivar development. J. Cotton Sci. 2:121–129.
- Zhu, J. 1993. Methods of predicting genotype value and heterosis for offspring of hybrids. J. Biomath. 8(1):32–44.
- Zhu, J. 1994. General genetic models and new analysis methods for quantitative traits. J. Zhejiang Agric. Univ. 20(6):551–559.
- Zhu, J., and B.S. Weir. 1996. Diallel analysis for sex-linked and maternal effects. Theor. Appl. Genet. 92:1–9.
Related articles in Crop Science:
- THIS ISSUE IN CROP SCIENCE
Crop Science 2004 44: 1109-1112.
[Full Text]
This article has been cited by other articles:
|
|
|
|
 
J. C. McCarty, J. Wu, and J. N. Jenkins
Use of Primitive Derived Cotton Accessions for Agronomic and Fiber Traits Improvement: Variance Components and Genetic Effects
Crop Sci.,
January 22, 2007;
47(1):
100 - 110.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J.C. McCarty and J.N. Jenkins
Registration of 14 Primitive Derived Cotton Germplasm Lines with Improved Fiber Strength
Crop Sci.,
October 27, 2005;
45(6):
2668 - 2669.
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. C. McCarty, J. N. Jenkins, and J. Wu
Primitive Accession Derived Germplasm by Cultivar Crosses as Sources for Cotton Improvement: I. Phenotypic Values and Variance Components
Crop Sci.,
July 1, 2004;
44(4):
1226 - 1230.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
