HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dudley, J. W. Search for Related Content PubMed Articles by Dudley, J. W. Agricola Articles by Dudley, J. W. Related Collections Maize Crop Genetics

Epistatic Interactions in Crosses of Illinois High Oil x Illinois Low Oil and of Illinois High Protein x Illinois Low Protein Corn Strains

Dep. of Crop Sci. Univ. of Illinois, S112 Turner Hall, 1102 S. Goodwin Ave., Urbana, IL 61801. This research was supported by the Illinois AES and a grant from Renessen, LLC

^* Corresponding author (jdudley{at}uiuc.edu).

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

 
Epistasis has been proposed as a possible explanation for the long continued progress from selection in the Illinois long-term corn (Zea mays L.) selection strains. From the crosses of Illinois High Oil (IHO) x Illinois Low Oil (ILO) and of Illinois High Protein (IHP) x Illinois Low Protein (ILP), 500 S₂ lines were developed. Each IHOxILO line was genotyped for 479 single nucleotide polymorphism (SNP) markers, and each IHPxILP line was genotyped for 499 SNP markers. Per se and testcross progenies of each S₂ line were evaluated for oil, protein, and starch. A large number of QTL were found to control these traits. The objective of this paper is to report the results of analysis of these crosses for two-way epistatic interactions. More epistatic interactions were significant than expected by chance. The proportion of significant interactions that involved a marker significant in a single marker analysis (SMA) was no greater than expected by chance. The number of markers associated only with significant epistatic effects ranged from 46.3 to 72.2% of the total number of markers significant for either an interaction effect or from SMA. The number of QTL controlling a trait is much greater than will be found by analyzing for significant QTL main effects. Thus, epistasis could contribute to the long continued response to selection in the Illinois long-term selection strains and also may help explain the continued success of commercial corn breeding.

Abbreviations: BLUP, best linear unbiased predictor • IHO, Illinois High Oil • IHP, Illinois High Protein • ILO, Illinois Low Oil • ILP, Illinois Low Protein • QTL, quantitative trait loci • SMA, single marker analysis • SNP, single nucleotide polymorphism

	NOTES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Received for publication July 23, 2007.

Epistatic Interactions in Crosses of Illinois High Oil x Illinois Low Oil and of Illinois High Protein x Illinois Low Protein Corn Strains

Dep. of Crop Sci. Univ. of Illinois, S112 Turner Hall, 1102 S. Goodwin Ave., Urbana, IL 61801. This research was supported by the Illinois AES and a grant from Renessen, LLC

^* Corresponding author (jdudley{at}uiuc.edu).

Epistasis has been proposed as a possible explanation for the long continued progress from selection in the Illinois long-term corn (Zea mays L.) selection strains. From the crosses of Illinois High Oil (IHO) x Illinois Low Oil (ILO) and of Illinois High Protein (IHP) x Illinois Low Protein (ILP), 500 S₂ lines were developed. Each IHOxILO line was genotyped for 479 single nucleotide polymorphism (SNP) markers, and each IHPxILP line was genotyped for 499 SNP markers. Per se and testcross progenies of each S₂ line were evaluated for oil, protein, and starch. A large number of QTL were found to control these traits. The objective of this paper is to report the results of analysis of these crosses for two-way epistatic interactions. More epistatic interactions were significant than expected by chance. The proportion of significant interactions that involved a marker significant in a single marker analysis (SMA) was no greater than expected by chance. The number of markers associated only with significant epistatic effects ranged from 46.3 to 72.2% of the total number of markers significant for either an interaction effect or from SMA. The number of QTL controlling a trait is much greater than will be found by analyzing for significant QTL main effects. Thus, epistasis could contribute to the long continued response to selection in the Illinois long-term selection strains and also may help explain the continued success of commercial corn breeding.

Abbreviations: BLUP, best linear unbiased predictor • IHO, Illinois High Oil • IHP, Illinois High Protein • ILO, Illinois Low Oil • ILP, Illinois Low Protein • QTL, quantitative trait loci • SMA, single marker analysis • SNP, single nucleotide polymorphism

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

 
SELECTION FOR HIGH OIL and high protein concentration in the corn (Zea mays L.) kernel has been effective for over 100 generations (Dudley, 2007). Four hypotheses have been advanced to explain the continued progress in long-term selection experiments: (i) large numbers of loci with low frequency of favorable alleles in the original population (Dudley, 1977), (ii) epistasis (Goodnight, 2004; Eitan and Soller, 2004), (iii) changes in environment (Dudley et al., 1974; Eitan and Soller, 2004), and (iv) mutation (Walsh, 2004; Keightley, 2004).

The results from classic quantitative genetic studies (Dudley, 1977) and quantitative trait locus (QTL) studies (Laurie et al., 2004; Clark et al., 2006; Dudley et al., 2007) suggest the presence of a large number (a minimum of 40) of segregating QTL affecting oil, protein, and starch concentration in the original Burr's White cultivar. The large number of QTL and estimates of low allelic frequency in the original population support the hypothesis advanced by Dudley (1977) that the long continued response to selection could be explained by the presence of a large number of QTL with relatively low frequency of favorable alleles segregating in the original Burr's White.

Until now, data supporting the presence of epistasis in long-term selection experiments have been minimal. Eitan and Soller (2004) suggested epistasis might be responsible for long-term response to selection in broiler chickens. Carlborg and Haley (2004) suggested epistasis had been neglected in complex trait studies. Carlborg et al. (2003) and Carlborg et al. (2004) presented data from QTL analysis in chickens that identified epistatic interactions in crosses between long-term selection lines of broiler chickens. Carlborg et al. (2006) provided an illustration of how epistasis could contribute to increased response to selection. Epistasis was shown to be important in QTL studies in rice (Oryza sativa, L.) by Li et al. (1997), Li et al. (2001), and Hua et al. (2003). Eitan and Soller (2004) suggested the presence of negative heterosis in crosses of an advanced generation of a selected strain to the original generation would be evidence for epistasis. Dudley (2007) presented evidence for negative heterosis in crosses involving different strains of the Illinois long-term selection experiment. However, there have been no reports of interactions between molecular marker associated effects in crosses involving the long-term selection strains. The objectives of this paper are to report the results of analyzing data from the crosses of Illinois High Oil (IHO) x Illinois Low Oil (ILO) and Illinois High Protein (IHP) x Illinois Low Protein (ILP) for epistatic interactions and to discuss the implications of these results for response to long-term selection and to corn breeding.

	MATERIALS AND METHODS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

 
Details of the development of progenies and markers used in this study have been reported, as have the experimental designs used (Laurie et al., 2004; Clark et al., 2006; Dudley et al., 2007). Briefly, crosses were made between plants from generation 70 of IHO and ILO and between plants from generation 70 of IHP and ILP. The IHOxILO cross was random mated 10 times, while the IHPxILP cross was random mated 7 times. Five hundred S₂ lines were developed from each cross. Each line was evaluated as a line per se and in testcrosses to a Monsanto tester. The IHOxILO lines per se and testcross progenies were evaluated at three locations for 2 yr. The IHPxILP lines were evaluated as lines per se for 2 yr at three locations, while the testcrosses were evaluated for only 1 yr in three locations. In each year–location combination, an (0,1) design with two replications and 50 blocks of 10 lines per replication was used. Plots consisted of 15 plants in a row 5.3 m long with 0.76 m between rows. Starch, protein, and oil concentrations were measured on a sample of grain from each plot using near infrared transmittance.

For each S₂ line, DNA was extracted from seedling tissue germinated from a bulk of about 50 seeds and single nucleotide polymorphism (SNP) genotypes obtained using procedures described by Laurie et al. (2004). Markers were chosen largely on the basis of allele frequency difference between the parents of the cross, as estimated by genotyping a random sample of 24 individuals each from IHP, ILP, IHO, and ILO as described by Laurie et al. (2004) for IHO and ILO. For IHOxILO, 479 SNP markers were used, while 499 SNP markers were used for the IHPxILP progenies. Map locations of QTL involved in significant interactions are based on a proprietary composite map constructed by Monsanto.

The phenotypic data were analyzed using the PROC MIXED algorithm of SAS software (version 9.00; SAS Institute, Cary, NC) to estimate components of variance and best linear unbiased predictors (BLUPs) for each line. The lines and all environmental factors were considered random. Using variance component estimates, heritabilities were calculated. In addition, using the variance component estimates and expected mean squares, mean squares were constructed and F tests made and used to determine exact confidence intervals (Knapp et al., 1985).

Proc GLM in SAS was used to analyze for two-way epistatic interactions between markers. The model used was Trait = M_i M_j M_i*M_j where M_i and M_j are markers M_i and M_j, respectively. Line BLUPs were used for analysis. Markers and the marker interaction were considered fixed. Only pairs of markers for which there were representatives in each two locus genotypic class were used in the final interpretation of the data. An interaction was declared significant if the p value in the F test was significant at the 0.001 probability level. For comparative purposes, single marker analyses (SMAs) were declared significant at the 0.001 probability. Permutation analyses were not done because of the extreme amount of computer time needed. Chi-square analysis was used to determine the probability of the number of observed significant interactions being greater than the number expected by chance. In addition, contingency table Chi-square analysis adjusted for continuity (Snedecor and Cochran, 1989) was used to compare the relative frequency of markers significant from the SMA, the interaction analysis, and from both in the per se and testcross progenies within the IHOxILO and IHPxILP crosses for each trait. Contingency table Chi-square analysis was also used to compare the relative frequency of significant markers in the two crosses for each trait.

	RESULTS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

 
Marker Interactions
Heritabilities were high enough and the coefficients of variation low enough to make these experiments appropriate for QTL analysis in both crosses and both types of progenies (Table 1 ). Detailed discussion of these results is found in Clark et al. (2006) and Dudley et al. (2007). Similarly, detailed discussion of QTL analysis of the experiments is found in these references and is not repeated here. The objective of this paper is to evaluate the importance of two-way epistatic interactions for oil, protein, and starch in the IHOxILO and IHPxILP crosses. For the IHOxILO cross there were 111,796 two-way interactions for which there were data in each of the nine genotypic classes; for the IHPxILP cross, there were 120,962 such interactions. Thus, approximately 97% of all possible pairs of markers were represented in each cross. An interaction was considered significant if the p value in the F test for interaction in the PROC GLM analysis was significant at the 0.001 probability level. Because of the large number of interactions tested, many of those found significant were expected to be significant by chance. At the 0.001 probability level used, 120.96 interactions were expected to be significant by chance in the IHPxILP cross and 111.80 in the IHOxILO cross. Benjamini and Hochberg (1995) presented a statistic called the false discovery rate, which measures the proportion of observed significant hypothesis tests expected by chance. The Chi-square test used here to test for deviations of observed numbers of significant interactions from the number expected is essentially a test for deviation of the false discovery rate from 1. Based on this Chi-square analysis, all progeny trait combinations except for the IHOxILO testcrosses for protein showed significant deviations from the expected, and in every case, the number of significant interactions observed was larger than expected (Table 2 ). Thus, epistasis occurs for oil, protein, and starch in at least one progeny type in both the IHOxILO and IHPxILP crosses.

Except for protein in both per se and testcross progenies in IHPxILP and starch in the per se progenies in IHPxILP, over half the markers found significant in either the SMA or the interaction analysis were significant as part of an epistastic interaction (Table 3 ). This result, along with the results of Carlborg et al. (2004) in chickens and Li et al. (2001) in rice, suggests that any estimates of number of QTL affecting a trait that do not take into account the QTL involved in significant interactions among regions showing no significant main effects will be a gross underestimate of the total number of QTL affecting the trait.

QTL Interactions
Discussion so far has been based on individual markers. Because of the large number of markers, several markers showing a significant interaction map to the same chromosome region. These markers likely are associated with the same QTL. To more closely identify the location of interacting QTL and to measure the number of interactions between QTL (as opposed to interactions between markers), markers on each chromosome of a significant pair were grouped so that for a given chromosome pair all markers within a 10-cM region on chromosome A, for example, that interacted with markers on chromosome B within a 10-cM region were grouped together. The location of the QTL on each chromosome was taken to be the mean location of the markers in a region. Marker locations were those derived from the Monsanto consensus map (Clark et al., 2006). Use of this map has limitations because it is based on F₂ data from crosses not used in this study. However, it is the only map that is in common between the two crosses. Because the progenies used in this study are from random-mated generations, crossovers between markers that mapped to the same location in the F₂ may have occurred and the genotypic patterns for a pair of markers that were identical in the F₂ may no longer be identical. Thus, map locations given should be considered only as approximations of the location of a given QTL involved in an interaction.

For protein, no significant QTL interactions involved chromosomes 5, 7, or 9 in either IHOxILO or IHPxILP (Table 5 ). In addition, no significant QTL interactions involved chromosome 10 in IHOxILO. In counting QTL interactions for pairs of chromosomes, a QTL interaction was considered significant if it was significant in either the per se or testcross progenies or both. Pairs of chromosomes for which there were significant QTL interactions in both crosses were 2 with 2, 2 with 8, 3 with 3, 3 with 4, and 4 with 6 (Table 5). The QTL interactions involving chromosome 1 with 6 were significant only in IHOxILO. By contrast, QTL interactions involving chromosomes 1 with 4, 2 with 4, 2 with 6, and 6 with 10 were significant only in IHPxILP. As expected based on the history of the parents of the crosses, the total number of QTL interactions in IHPxILP (82 = sum of 57 for per se + 44 for testcross – 19 for both) (Table 6 ) for protein was over 2.5 times larger than the total for IHOxILO (31).

For starch, no QTL involved in significant interactions were found in either cross on chromosomes 1, 4, 7, 8, or 9, but QTL involved in significant interactions were found in both crosses on chromosomes 2, 3, 5, 6, and 10. Significant QTL interactions of chromosomes 2 with 2 and 10, 3 with 3, 5 with 5, 6 with 6, and 6 with 10 were found in both crosses. Significant interactions of QTL on chromosome 5 with QTL on chromosome 6 were found only in the IHPxILP cross. The total number of significant QTL interactions (30 for IHPxILP and 23 for IHOxILO) was similar for the two crosses and lower than for either oil or protein.

Comparison of Traits
Chromosome pairs were classified according to the traits that had significant QTL interactions. Chromosome pairs showing significant QTL interactions only for oil were 1 with 5, 4 with 7, 6 with 8, and 9 with 10 (Table 5). Those showing significant QTL interactions only for protein were 1 with 4, 2 with 4, 6, with 8, 3 with 3, 3 with 4, and 4 with 6. Those showing significant QTL interactions only for starch were 2 with 10, 5 with 5, and 6 with 6. The only chromosome pair showing significant QTL interactions for both oil and protein was 1 with 6, for both oil and starch 5 with 6, and for protein and starch 6 with 10. Only QTL interactions involving chromosome 2 with 2 were significant for all three traits. In the IHPxILP cross, 6 of the 8 chromosome pairs showing significant QTL interactions were significant for both the per se and testcross progenies and 2 were significant for only the per se progenies.

Locations of QTL Involved in Interactions
There were no interactions involving both protein and oil controlled by QTL in the same regions on a pair of chromosomes in either cross (Table A1). For oil and starch, a QTL on chromosome 2 at location 35.9 interacted with a QTL on the same chromosome at position 134.9 and a QTL at location 139.8 interacted with one at location 164.2 for both traits in the IHOxILO cross.

For starch and protein, QTL on chromosome 2 interacting with other QTL on chromosome 2 for both traits were located at positions 9.9 and 36.0, 76.8 and 87.5 in the IHPxILP cross, and 78.2 and 105.0 in the IHOxILO cross (Table A1). On chromosome 3, QTL showing interactions with other QTL on chromosome 3 were located at positions 8.1 and 143.0, 19.7 and 59.7, 56.0 and 142.0, 111.9 and 142.0, and 123.7 and 142.0 in IHPxILP and 59.0 and 68.0 in IHOxILO. A QTL located near position 142.0 was involved in all the interactions found in IHPxILP. Interactions involving chromosomes 6 and 10 were found near positions 18.0 on chromosome 6 and 58.0 on chromosome 10, and between 117.8 on chromosome 6 and 55.6 on chromosome 10 in the IHPxILP cross. Both of these interactions apparently involve a QTL located near position 56.0 on chromosome 10.

	DISCUSSION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

 
The implications of these results differ regarding attempts to understand the genetic mechanisms controlling oil, protein, and starch and regarding changes in breeding methods. The presence of a large number of QTL involved in epistasis that were not identified by SMA suggests that many more QTL are involved in control of protein, oil, and starch concentration in the corn kernel than would be suggested by results of SMA. Thus, understanding the genetic mechanisms controlling kernel quality will require an understanding of gene-by-gene interactions as well as the effects of individual genes. To best use this information, it may be necessary to identify the important interactions and the metabolic pathways involved. Future research in this area may involve gene expression experiments to identify the genes and pathways involved.

The presence of epistasis has major implications for plant breeding. It means there is much more variability potentially available for selection than is often thought. Although the results from the IHPxILP and IHOxILO generation 70 crosses do not measure the epistasis available for selection in the original Burr's White or in previous generations, the finding of significant epistatic effects controlling oil and protein in these crosses suggests that epistasis could have contributed to genetic variability for oil and protein during the selection process. Thus, epistasis must be considered a possible contributor to the long continued progress from selection for oil and protein in the Illinois long-term selection strains. More important, the presence of epistasis may help explain the continued success of commercial corn breeding programs even though selection continues to be within a narrow range of germplasm.

The importance of epistasis in designing marker-assisted selection programs will depend on whether models including epistatic effects have greater power to predict the success of a given selection program than models not including such effects. Many questions remain to be answered before the implications to design of breeding programs are known. How many lines need to be tested to develop a useful prediction model? Can prediction models developed in one cross be used to predict performance in a related cross? Are there methods of using marker data to predict performance that do not depend on identifying individual epistatic effects? In a practical sense, the identification of epistatic effects will require progeny sizes considerably larger than those normally used in most corn breeding programs to identify all possible two-locus genotypes with large enough numbers of individuals for each genotype to allow sufficiently reliable estimates of genotype means to improve precision of selection. Thus, evaluation of the potential of methods such as partial least squares regression to predict performance using marker and phenotypic data may provide useful information. One unknown is whether fewer progeny will allow useful predictive power using partial least squares than are required for identifying epistatic effects. Results from the present study suggest that epistasis is important, which then opens the door to questions about how to use epistasis in corn breeding.

	APPENDIX

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

 

Table A1. Location of epistatic quantitative trait locus (QTL) regions for oil, protein, and starch in the Illinois High Protein (IHP) x Illinois Low Protein (ILP) and Illinois High Oil (IHO) x Illinois Low Oil (ILO) crosses. An epistatic QTL region was identified by grouping markers on each chromosome of a significant pair so that for a given chromosome pair all significant markers within a 10 cM region on chromosome A, e.g., which interacted with markers on chromosome B within a 10 cM region were grouped together. The location of the QTL region on each chromosome was taken to be the mean location of the markers in a region.

Trait Cross Generation Chromosomes Position in cM Protein IHPxILP PS TC 1 4 22.7 118.7 Protein IHPxILP – TC 1 4 35.6 62.1 Protein IHPxILP PS – 1 4 46.8 81.6 Protein IHPxILP PS TC 1 4 51.5 52.8 Protein IHPxILP PS – 1 4 58.4 127.5 Protein IHPxILP PS – 1 4 61.5 31 Protein IHPxILP PS – 1 4 61.5 80 Protein IHPxILP – TC 1 4 107.3 62.1 Protein IHPxILP PS TC 1 4 120.0 62.0 Protein IHPxILP – TC 1 4 130.7 129.2 Protein IHPxILP PS – 1 4 133.9 58.6 Protein IHPxILP PS TC 1 4 133.9 129.2 Protein IHPxILP PS TC 1 4 149.0 119.2 Protein IHPxILP PS – 1 4 160.4 138.1 Protein IHPxILP PS – 1 4 178.2 119.2 Protein IHPxILP PS – 1 4 190.7 58.6 Protein IHPxILP PS – 1 4 190.7 119.2 Oil IHPxILP PS TC 1 5 58.4 1.6 Oil IHPxILP PS – 1 5 58.4 139.9 Oil IHOxIHO – TC 1 5 60.6 57.7 Oil IHOxILO PS TC 1 5 83.2 57.7 Oil IHOxILO PS – 1 5 90.5 70.8 Oil IHOxILO – TC 1 5 95.0 41.7 Oil IHOxILO PS – 1 5 96.9 1.6 Oil IHOxILO – TC 1 5 128.5 81.2 Oil IHOxILO PS – 1 5 132.1 13.8 Oil IHPxILP PS – 1 5 133.9 70.2 Oil IHOxILO PS TC 1 5 138.5 117.1 Oil IHPxILP PS – 1 5 144.0 69.3 Oil IHOxILO PS TC 1 5 160.5 76.2 Oil IHOxILO PS – 1 5 165.6 71.7 Oil IHOxILO – TC 1 5 200.3 136.0 Oil IHPxILP PS – 1 5 204.6 69.3 Oil IHPxILP PS – 1 6 16.0 53.1 Protein IHOxILO – TC 1 6 30.4 95.8 Protein IHOxILO PS – 1 6 44.0 132.7 Protein IHOxILO PS – 1 6 45.0 53.5 Oil IHPxILP PS – 1 6 50.4 37.3 Protein IHOxILO PS – 1 6 60.3 66.6 Oil IHPxILP PS TC 1 6 62.0 32.8 Protein IHOxILO – TC 1 6 89.5 36.3 Protein IHOxILO PS – 1 6 92.3 53.5 Oil IHPxILP PS – 1 6 95.0 21.2 Oil IHPxILP PS – 1 6 95.0 97.8 Protein IHOxILO PS – 1 6 103.3 35.3 Protein IHOxILO – TC 1 6 116.3 59.9 Oil IHOxILO – TC 1 6 127.6 36.3 Oil IHOxILO – TC 1 6 127.6 53.5 Oil IHOxILO – TC 1 6 128.4 67.5 Protein IHOxILO – TC 1 6 129.5 35.3 Oil IHOxILO – TC 1 6 129.5 53.5 Oil IHOxILO – TC 1 6 138.5 30.8 Protein IHOxILO PS – 1 6 140.0 61.5 Oil IHPxILP PS – 1 6 159.4 58.1 Protein IHOxILO PS – 1 6 159.7 17.3 Oil IHPxILP PS – 1 6 164.6 33.0 Oil IHOxILO – TC 1 6 200.3 19.4 Protein IHPxILP PS – 2 2 9.9 35.7 Starch IHPxILP PS – 2 2 9.9 36.5 Protein IHPxILP PS TC 2 2 13.5 51.7 Starch IHOxILO PS – 2 2 19.5 95.0 Oil IHOxILO – TC 2 2 32.8 164.2 Oil IHOxILO – TC 2 2 35.9 134.9 Starch IHOxILO – TC 2 2 35.9 134.9 Protein IHPxILP – TC 2 2 36.0 82.8 Oil IHOxILO PS TC 2 2 74.8 123.4 Starch IHPxILP PS – 2 2 76.2 76.8 Protein IHPxILP PS TC 2 2 76.8 87.5 Starch IHPxILP PS – 2 2 76.8 87.5 Starch IHOxILO PS TC 2 2 78.2 102.0 Protein IHOxILO – TC 2 2 78.2 107.0 Oil IHOxILO PS – 2 2 99.7 155.3 Oil IHOxILO PS – 2 2 123.4 127.0 Oil IHOxILO – TC 2 2 139.8 164.2 Starch IHOxILO – TC 2 2 139.8 164.2 Protein IHPxILP PS TC 2 4 19.5 21.1 Protein IHPxILP PS TC 2 4 19.5 58.6 Protein IHPxILP – TC 2 4 19.5 129.2 Protein IHPxILP – TC 2 4 30.1 117.7 Protein IHPxILP – TC 2 4 44.6 81.8 Protein IHPxILP – TC 2 4 47.0 8.2 Protein IHPxILP PS TC 2 4 70.4 8.2 Protein IHPxILP – TC 2 4 70.4 118.7 Protein IHPxILP – TC 2 4 76.5 48.0 Protein IHPxILP PS – 2 4 76.5 133.5 Protein IHPxILP – TC 2 4 88.0 32.0 Protein IHPxILP PS TC 2 4 88.5 65.0 Protein IHPxILP PS – 2 4 131.0 115.1 Protein IHPxILP PS TC 2 4 154.9 56.0 Protein IHPxILP PS – 2 4 154.9 119.0 Protein IHPxILP – TC 2 6 9.9 53.1 Protein IHPxILP PS TC 2 6 19.5 7.5 Protein IHPxILP PS – 2 6 30.9 49.1 Protein IHPxILP – TC 2 6 82.8 71.7 Protein IHPxILP PS TC 2 6 89.6 60.8 Protein IHPxILP PS – 2 6 93.7 78.8 Protein IHPxILP – TC 2 6 106.2 6.5 Protein IHPxILP – TC 2 6 106.2 103.4 Protein IHPxILP PS – 2 6 154.9 37.3 Protein IHPxILP PS – 2 6 154.9 61.5 Protein IHPxILP PS TC 2 8 31.5 88.0 Protein IHOxILO – TC 2 8 38.3 64.0 Protein IHPxILP PS – 2 8 48.0 50.0 Protein IHPxILP PS – 2 8 70.4 50.0 Protein IHOxILO – TC 2 8 78.2 65.8 Protein IHPxILP PS – 2 8 87.5 102.1 Protein IHPxILP PS – 2 8 93.7 20.0 Protein IHOxILO – TC 2 8 153.5 8.8 Protein IHOxILO – TC 2 8 164.2 105.5 Starch IHPxILP – TC 2 10 9.6 108.1 Starch IHOxILO – TC 2 10 19.5 55.0 Starch IHOxILO – TC 2 10 32.8 23.9 Starch IHOxILO – TC 2 10 32.8 55.0 Starch IHPxILP – TC 2 10 33.9 73.6 Starch IHPxILP PS – 2 10 80.0 60.1 Starch IHOxILO – TC 2 10 80.0 60.1 Starch IHPxILP – TC 2 10 131.0 108.1 Starch IHPxILP PS – 2 10 158.0 66.0 Starch IHPxILP PS – 3 3 8.1 142.0 Protein IHPxILP PS TC 3 3 8.1 144.5 Protein IHPxILP PS – 3 3 19.7 59.7 Starch IHPxILP PS – 3 3 19.7 59.7 Protein IHOxILO PS – 3 3 52.3 98.2 Starch IHPxILP – TC 3 3 54.5 59.3 Protein IHPxILP PS – 3 3 55.0 142.0 Starch IHPxILP PS – 3 3 57.2 142.0 Starch IHPxILP – TC 3 3 59.3 67.0 Starch IHOxILO PS – 3 3 59.3 67.0 Protein IHOxILO PS – 3 3 59.3 68.5 Starch IHPxILP – TC 3 3 59.3 112.3 Protein IHOxILO PS TC 3 3 59.5 84.0 Starch IHPxILP – TC 3 3 61.0 61.7 Protein IHPxILP PS – 3 3 61.0 142.0 Protein IHPxILP – TC 3 3 61.7 64.0 Starch IHPxILP – TC 3 3 61.7 81.9 Protein IHOxILO PS – 3 3 61.7 123.8 Protein IHPxILP – TC 3 3 64.0 112.3 Protein IHPxILP – TC 3 3 80.0 88.6 Starch IHPxILP – TC 3 3 80.0 89.5 Protein IHPxILP PS – 3 3 111.9 142.0 Starch IHPxILP PS – 3 3 111.9 142.0 Protein IHPxILP PS – 3 3 123.7 142.0 Starch IHPxILP PS – 3 3 123.7 142.0 Protein IHOxILO PS – 3 4 9.1 92.1 Protein IHOxILO PS – 3 4 19.7 69.5 Protein IHOxILO PS – 3 4 19.7 92.1 Protein IHPxILP PS TC 3 4 60.0 60.0 Protein IHPxILP – TC 3 4 61.7 121.6 Protein IHOxILO PS – 3 4 68.5 1.0 Protein IHOxILO PS TC 3 4 69.5 109.5 Protein IHPxILP PS TC 3 4 88.7 113.5 Protein IHOxILO PS – 3 4 96.1 69.5 Protein IHOxILO PS – 3 4 98.6 135.1 Protein IHPxILP PS – 3 4 106.9 49.6 Protein IHPxILP – TC 3 4 106.9 127.5 Protein IHOxILO PS – 3 4 109.4 92.1 Protein IHPxILP PS – 3 4 142.0 52.8 Protein IHPxILP PS – 3 4 142.0 66.5 Protein IHPxILP PS – 3 4 142.0 125.5 Protein IHPxILP PS – 3 4 142.0 146.6 Protein IHPxILP – TC 3 4 145.0 129.2 Protein IHOxILO PS TC 4 6 38.7 38.0 Protein IHPxILP PS – 4 6 58.6 60.8 Protein IHPxILP PS – 4 6 63.6 106.1 Protein IHPxILP PS – 4 6 80.0 59.9 Protein IHOxILO PS – 4 6 109.2 92.8 Protein IHPxILP PS – 4 6 117.6 7.0 Protein IHPxILP PS – 4 6 119.8 86.9 Protein IHPxILP PS – 4 6 124.8 43.1 Protein IHPxILP PS – 4 6 129.2 58.1 Protein IHOxILO PS – 4 6 142.1 22.9 Oil IHOxILO PS – 4 7 27.6 86.0 Oil IHOxILO PS – 4 7 38.7 189.4 Oil IHOxILO PS – 4 7 58.6 144.6 Oil IHPxILP PS – 4 7 68.3 90.0 Oil IHOxILO PS – 4 7 69.5 69.1 Oil IHPxILP PS – 4 7 118.4 67.9 Oil IHPxILP PS – 4 7 127.5 90.0 Oil IHPxILP PS – 4 7 127.5 173.7 Oil IHOxILO PS – 4 7 135.3 122.5 Oil IHOxILO PS – 4 7 136.4 62.0 Starch IHPxILP PS – 5 5 -3.4 100.9 Starch IHPxILP PS TC 5 5 1.6 70.2 Starch IHOxILO – TC 5 5 16.7 40.0 Starch IHPxILP PS TC 5 5 28.0 70.5 Starch IHOxILO – TC 5 5 31.3 117.1 Starch IHPxILP PS TC 5 5 41.0 70.2 Starch IHOxILO PS – 5 5 57.7 117.1 Starch IHPxILP PS – 5 5 70.8 99.6 Oil IHPxILP PS – 5 6 1.6 7.5 Oil IHPxILP PS – 5 6 13.8 7.5 Oil IHPxILP PS – 5 6 13.8 30.0 Oil IHPxILP PS – 5 6 13.8 103.4 Oil IHPxILP PS – 5 6 22.3 7.5 Oil IHPxILP PS – 5 6 22.3 100.0 Oil IHPxILP PS – 5 6 34.4 32.8 Oil IHOxILO PS – 5 6 62.5 40.0 Oil IHOxILO PS – 5 6 62.5 52.8 Oil IHPxILP PS – 5 6 69.3 71.7 Starch IHOxILO PS – 5 6 70.0 31.8 Starch IHOxILO PS – 5 6 71.2 112.4 Starch IHOxILO – TC 5 6 117.1 35.3 Starch IHOxILO – TC 5 6 117.1 53.1 Starch IHOxILO PS – 6 6 27.8 37.5 Starch IHOxILO PS – 6 6 37.8 53.3 Starch IHOxILO – TC 6 6 37.8 53.3 Starch IHOxILO PS – 6 6 54.0 58.0 Starch IHOxILO – TC 6 6 54.0 58.0 Starch IHPxILP – TC 6 6 55.1 106.1 Oil IHPxILP PS – 6 8 38.4 53.9 Oil IHPxILP – TC 6 8 55.1 77.5 Oil IHPxILP PS – 6 8 59.0 55.0 Oil IHPxILP PS – 6 8 71.7 71.9 Oil IHPxILP – TC 6 8 78.8 14.7 Oil IHPxILP PS – 6 8 113.1 102.1 Oil IHPxILP PS TC 6 8 115.0 77.5 Oil IHPxILP – TC 6 8 117.8 117.3 Protein IHPxILP PS TC 6 10 18.0 56.0 Starch IHPxILP PS TC 6 10 18.0 60.0 Starch IHOxILO PS – 6 10 33.3 68.0 Starch IHOxILO PS – 6 10 36.2 55.8 Protein IHPxILP – TC 6 10 58.7 83.5 Protein IHPxILP – TC 6 10 67.5 65.8 Protein IHPxILP – TC 6 10 103.5 56.0 Starch IHPxILP – TC 6 10 103.5 71.6 Starch IHPxILP – TC 6 10 103.5 108.1 Starch IHPxILP PS TC 6 10 104.4 57.0 Protein IHPxILP – TC 6 10 117.8 55.6 Starch IHPxILP – TC 6 10 117.8 55.6 Oil IHPxILP – TC 9 10 78.2 55.6 Oil IHPxILP PS – 9 10 96.0 62.0 Oil IHPxILP PS – 9 10 100.6 108.1 Oil IHPxILP – TC 9 10 110.3 66.0 Oil IHPxILP – TC 9 10 166.6 108.1

PS = per se progenies; TC = testcross progenies.

Pair of chromosomes on which interacting QTL are located.

Position of QTL on chromosome based on Monsanto composite map. Left-hand column corresponds to chromosome in left-hand column under the heading chromosomes.

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication July 23, 2007.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

 

• Benjamini, Y., and Y. Hochberg. 1995. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B 57:289–300.
• Carlborg, Ö., and C.S. Haley. 2004. Epistasis: Too often neglected in complex trait studies? Nat. Rev. Genet. 5:618–625.[CrossRef][Web of Science][Medline]
• Carlborg, Ö., P.M. Hocking, D.W. Burt, and C.S. Haley. 2004. Simultaneous mapping of epistatic QTL in chickens reveals clusters of QTL pairs with similar genetic effects on growth. Genet. Res. 83:197–209.[CrossRef][Web of Science][Medline]
• Carlborg, Ö., L. Jacobsson, P. Åhgren, P. Siegel, and L. Anderson. 2006. Epistasis and the release of genetic variation during long-term selection. Nat. Genet. 38:418–420.[CrossRef][Web of Science][Medline]
• Carlborg, Ö., S. Kerje, I. Schutz, L. Jacobsson, P. Jensen, and L. Andersson. 2003. A global search reveals epistatic interaction between QTL for early growth in the chicken. Genome Res. 13:413–421.[Abstract/Free Full Text]
• Clark, D., J.W. Dudley, T.R. Rocheford, and J.R. LeDeaux. 2006. Genetic analysis of corn kernel composition in the random mated 10 generation of the cross of generations 70 of IHO x ILO. Crop Sci. 46:807–819.[Abstract/Free Full Text]
• Dudley, J.W. 1977. Seventy-six generations of selection for oil and protein percentage in maize. p. 459–473. In E. Pollak, O. Kempthorne, and T.B. Bailey, Jr. (ed.) Proc Int. Conf. on Quantitative Genetics. Iowa State Univ. Press. Ames.
• Dudley, J.W. 2007. From means to QTL: The Illinois long-term selection experiment as a case study in quantitative genetics. In Proc. 2nd Int. Plant Breeding Symp., Mexico City. 20–25 Aug. 2006. Crop Sci. (suppl.) [[(In press)]].
• Dudley, J.W., D. Clark, T.R. Rocheford, and J.R. LeDeaux. 2007. Genetic analysis of corn kernel chemical composition in the random mated 7 generation of the cross of generations 70 of IHP x ILP. Crop Sci. 47:45–57.[Abstract/Free Full Text]
• Dudley, J.W., R.J. Lambert, and D.E. Alexander. 1974. Seventy generations of selection for oil and protein concentration in the maize kernel. p. 181–212. In J.W. Dudley (ed.) Seventy generations of selection for oil and protein I maize. CSSA, Madison, WI.
• Eitan, Y., and M. Soller. 2004. Selection induced genetic variation. p. 154–176. In S. Wasser (ed.) Evolutionary theory and processes: Modern horizons, papers in honour of Eviatar Nevo. Kluwer Academic, the Netherlands
• Goodnight, C.J. 2004. Gene interaction and selection. Plant Breed. Rev. 24:269–291.
• Hua, J., Y. Xing, W. Wu, C. Xu, X. Sun, S. Yu, and Q. Zhang. 2003. Single-locus heterotic effects and dominance by dominance interactions can adequately explain the genetic basis of heterosis in an elite rice hybrid. Proc. Natl. Acad. Sci. USA 100:2574–2579.[Abstract/Free Full Text]
• Keightley, P.D. 2004. Mutational variation and long-term selection response. Plant Breed. Rev. 24:227–247.
• Knapp, S.J., W.W. Stroup, and W.M. Ross. 1985. Exact confidence intervals for heritability on a progeny mean basis. Crop Sci. 25:192–194.[Abstract/Free Full Text]
• Laurie, C.C., S.D. Chasalow, J.R. LeDeaux, R. McCarroll, D. Bush, B. Hauge, C. Lai, D. Clark, T.R. Rocheford, and J.W. Dudley. 2004. The genetic architecture of oil concentration in the maize kernel after 70 generations of divergent selection. Genetics 168:2141–2155.[Abstract/Free Full Text]
• Li, Z.-K., L.J. Luo, H.W. Mei, D.L. Wang, Q.Y. Shu, R. Tabien, D.B. Zhong, C.S. Ying, J.W. Stansel, G.S. Khush, and A.H. Paterson. 2001. Overdominant epistatic loci are the primary genetic basis of inbreeding depression and heterosis in rice. I. Biomass and grain yield. Genetics 158:1737–1753.
• Li, Z., S.R.M. Pinson, W.D. Park, A.H. Paterson, and J.W. Stansel. 1997. Epistasis for three grain yield components in rice (Oryza sativa L.). Genetics 145:453–465.[Abstract]
• Snedecor, G.W., and W.G. Cochran. 1989. Statistical methods. 8th ed. Iowa State Univ. Press, Ames.
• Walsh, B. 2004. Population and quantitative genetic models of selection limits. Plant Breed. Rev. 24:177–225.

This article has been cited by other articles:

				J. W. Dudley and G. R. Johnson Epistatic Models Improve Prediction of Performance in Corn Crop Sci., May 11, 2009; 49(3): 763 - 770. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dudley, J. W. Search for Related Content PubMed Articles by Dudley, J. W. Agricola Articles by Dudley, J. W. Related Collections Maize Crop Genetics