Printer Friendly

Genetic analysis of corn kernel chemical composition in the random mated 10 generation of the cross of generations 70 of IHO X ILO.

THE ILLINOIS long-term selection experiment is a classic illustration of efficacy of selection for chemical composition in the corn kernel. Initiated in 1896 with the objective of changing kernel composition in an open-pollinated corn cultivar, divergent recurrent selection has been applied for over 100 generations (Dudley and Lambert, 2004). Mean oil concentration in IHO after 104 generations is 206 g [kg.sup.-1], with no apparent reduction in selection response. Selection in ILO was discontinued after Generation 89 because oil was so low that differences among genotypes could not be reliably measured. Before selection, the initial population contained 47 g [kg.sup.-1] oil, which is similar to current commercial hybrids.

Grain quality is an important objective in cereal breeding (Sadimantara et al., 1997; Mazur et al., 1999; Tan et al., 1999; Wang and Larkins, 2001; Xu et al., 2003). In corn grain, a typical hybrid cultivar contains approximately 40 g [kg.sup.-1] oil, 90 g [kg.sup.-1] protein, 730 g [kg.sup.-1] starch, and 140 g [kg.sup.-1] other constituents (mostly fiber). The oil is stored mainly in the form of triacylglycerols in the germ, while starch and protein are found primarily in the endosperm, which makes up the majority of the kernel (Tan and Morrison, 1979). Grain quality can be improved by alteration of the relative amounts of the primary kernel constituents to create a product more adapted for a specific use. For example, livestock studies have shown a greater rate of weight gain per pound of feed in swine using high oil (>60 g [kg.sup.-1]) corn as opposed to standard feed corn (Lambert, 1994; Lambert et al., 2004). This is because the caloric content of oil is 2.25 times that of starch on a dry weight basis. Accordingly, high oil corn has added value for animal feed, the primary use of corn in developed countries (National Corn Growers Association, 2001). Another possible source to add value would be the use of high starch corn in the production of ethanol, since starch is the carbon source for ethanol production from corn. For these two examples, to the degree oil and starch are negatively correlated, high oil corn would be synonymous with low starch corn and vice versa.

In comparison with commercial cultivars, the yield and agronomic characteristics of IHO and ILO are poor (Dudley et al., 1974; Lambert, 1994) and are therefore not amenable to direct commercialization. Yet, understanding the genetic factors underlying kernel traits in IHO and ILO may be useful in two ways. First, it may be possible to introgress desirable alleles for kernel composition into elite germplasm by means of conventional plant breeding procedures. This approach would be practical if a small number of genes with relatively large effects control oil concentration. Second, identification and subsequent sequencing of the genes underlying kernel trait variation would provide the possibility of genetic engineering through plant transformation and might allow for identification of favorable alleles for kernel composition in elite germplasm.

Random mating breaks up linkages between markers and QTL and between linked QTL. Thus, it is an experimental tool for increasing resolution for detection of QTL (Darvasi and Soller, 1995). In a study of the effects of random-mating on QTL detection using the cross of Illinois High Protein X Illinois Low Protein, Dudley et al. (2004) found, as expected by theory, a significant reduction in the percentage of markers declared as significant between the Syn0 (the generation produced by random-mating the [F.sub.1]) and Syn4 (fourth generation of random-mating after the Syn0) generations. Likewise, Willmot et al. (2006) found a significant reduction in the percentage of markers declared as significant between the Syn0 and Syn4 of the cross of IHO X ILO.

Although IHO and ILO were developed for divergent kernel oil concentration, differences exist for other kernel traits. Specifically, at Generation 70, IHO kernels had a mean of 166 g [kg.sup.-1] oil, 150 g [kg.sup.-1] protein, and 460 g [kg.sup.-1] starch. In contrast, ILO kernels from Generation 70 contained 4 g [kg.sup.-1] oil, 120 g [kg.sup.-1] protein, and 660 g [kg.sup.-1] starch. The existence of differences between IHO and ILO in kernel traits other than oil allows for QTL analysis of those traits as well. For example, Berke and Rocheford (1995) studied a cross between Generation 90 of IHO and a strain of ILO selected for 19 generations for earliness after 70 generations of selection for low oil and found 31 marker loci clustered in 11 regions associated with oil concentration, 16 in eight regions associated with protein concentration, and 28 in 13 regions associated with starch. The fact that differences exist between IHO and ILO in kernel traits not directly selected on suggests that genetic correlations, resulting from linkage or pleiotropy, are present between these traits and oil concentration. The distinction between pleiotropy and linkage is important, since it could greatly affect possible commercial development. The number and relative effects of genetic factors underlying a trait also affect the potential for altering chemical composition by traditional breeding or by molecular techniques.

By using a large number of markers and progenies in the R[M.sub.10] from the cross of IHO x ILO, this study takes advantage of the effects of random mating on identifying marker-QTL associations. A previous study by Laurie et al. (2004) using the same progenies and markers identified approximately 50 markers significantly associated with QTL for oil using a regression approach to analyze the data. The current study is unique in extending current statistical methodology to allow development of a molecular marker map and use of simple interval mapping on progenies from a population cross. In addition, this study reports analysis of testcross progenies and of protein and starch data not reported in the study of Laurie et al. (2004). In sum, the objectives of this study were to (i) develop and implement statistical methodology appropriate to do simple interval mapping of data from this particular experiment, (ii) identify QTL affecting oil, protein, and starch concentration, (iii) determine associations among significant QTL for these traits, and (iv) compare results from per se and testcross progenies.

MATERIALS AND METHODS

Genetic Materials

In 1971, a reciprocal cross was made between Cycle 70 of IHO and ILO using about five to seven plants from each population. Progeny were subsequently intermated at random for 10 generations with an effective population size of approximately 200 plants (100 male and 100 female parents). About 1500 R[M.sub.10] individuals were then self mated two generations without selection to produce 500 lines per se, each represented by an [S.sub.2] ear. About 50 kernels from each line per se were planted adjacent to Monsanto 7051, a commercial inbred tester, and two types of crosses made. In one, the tester was fertilized with pollen from the [S.sub.2] plants to produce 500 hybrid progeny lines. In the other, plant to plant crosses were made within an [S.sub.2] row to produce seed required for line per se evaluation. Thus two types of progeny were developed: the lines per se (represented by a mating structure of [F.sub.1]R[M.sub.10][S.sub.2]) and their corresponding testcrosses with Monsanto 7051.

Molecular Marker Data

Details of molecular marker data generation and its quality have been reported (Laurie et al., 2004). Briefly, SNP genotypes were determined for each line with DNA extracted from seedling tissue germinated from a bulk of about 50 seeds. This provided an estimate of the per se genotypes of the preceding generation ([F.sub.1]R[M.sub.10][S.sub.1). Because of the nature of SNPs all markers were biallelic. Markers were chosen largely on the basis of allele frequency difference between IHO and ILO (henceforth IHO and ILO refer to Cycle 70 unless otherwise noted), as estimated by genotyping a random sample of 24 individuals each from IHO and ILO (see Laurie et al., 2004 for details). Loci were coded consistently by assigning the same designation (e.g., "A" as opposed to "a") to alleles with higher frequency in IHO. Marker order was assumed to be identical to that in a composite map constructed by Monsanto for three different populations (Laurie et al., 2004). Data from 479 segregating SNP markers genotyped on 499 lines per se (one of the 500 lines was dropped because of poor quality DNA) were used for the analysis reported here.

Phenotypic Data

Kernel phenotypes were measured on both lines per se and testcrosses over 2 yr at three locations with two replications each. In each year-location combination plots were grown in a generalized lattice [[alpha](0,1)] experimental design (Patterson and Williams, 1976). The lines per se and the testcrosses were grown in separate experiments. Plots consisted of 15 plants in a row 5.3 m long with 0.76 m between rows. The locations were Monmouth, IL, Urbana, IL, and Williamsburg, IA. A replicate consisted of 50 blocks of ten rows per block, with one row per line. Approximately ten pollinations were made within each row, with one pollen parent per mating and each pollen parent being used only once. This yielded 6 to 8 ears per row from which kernels were bulked for analysis.

Oil, protein, and starch concentration were measured in the lines per se with near infrared reflectance (NIR) on ground seed samples with a Dickey-john instrument (Hymowitz et al., 1974; Dudley and Lambert, 1992; Dyer and Feng, 1997). For testcrosses, Monsanto company analyzed whole kernel samples by near infrared transmittance (NIT) with an Infratec Grain Analyzer. The NIR method was used for the lines per se because of a lack of enough seed to analyze by NIT. Seed of 27 lines, spanning the range of values expected, were analyzed by both methods. Means from the two lines were comparable and correlations between methods ranged from 0.93 for starch and protein to 0.97 for oil. Thus the methods were considered to give comparable results.

The phenotypic data were analyzed by the PROC MIXED procedure of SAS software (version 9.00, SAS Institute, Inc., Cary, NC, www.sas.com) 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, we calculated point estimates of and exact confidence intervals for heritabilities (Nyquist, 1991; Knapp et al., 1985).

Estimates of phenotypic and genotypic correlations were used to measure relationships among traits. Per se and testcross BLUP estimates were used to calculate phenotypic correlations and their standard errors (Lynch and Walsh, 1998; Kendall and Stuart, 1977). A method described by Holland (personal communication, see http://www4.ncsu.edu/%7Ejholland/ correlation.html (verified 30 November 2005), see also Holland et al., 2002) was used to calculate genotypic correlations and standard errors. In this method, trait data are analyzed as a repeated measures experiment by SAS (PROC MIXED), with one measure being one trait and a second measure being a second trait. Other variables such as years, locations, reps, and blocks were included in the same manner as in the phenotypic data analysis and were considered random.

Single Marker Analysis

Single marker analysis (SMA) was performed for the kernel quality traits and the results used for comparison by simple interval mapping (SIM). SMA consisted of linear regression of marker genotypes on marker class phenotypic means on a per marker basis using SAS (Kearsey and Hyne, 1994). Results produced include a significance value, an estimate of the additive effect (slope of the regression line), and a goodness of fit value (adjusted [r.sup.2]). Markers with a p value [less than or equal to] 0.01 which had a minimum of 20 individuals in each homozygous marker class were considered significant.

Simple Interval Mapping (SIM)

In this study, there were two primary challenges to the use of SIM. First, traditional QTL interval mapping is based on crosses between homozygous parents, enabling unequivocal assignment of marker and QTL alleles to each parent with frequency 1 or 0. When one intermated population is crossed with another, as was the case here, this condition no longer holds. Second, the mating structure to be analyzed, rather than being a simple [F.sub.2] or backcross, comprised ten generations of random mating of the population cross followed by two generations of selfing.

The parents used to make the initial cross no longer exist. Hence, initial estimates of allele frequencies, recombination frequencies, and gametic disequilibrium (Table 1) for the parents used to make the original cross were not directly observable. Nonetheless, estimates of these parameters are necessary to enable interval mapping in the final generation conditional on the mating scheme. Thus, two steps were taken to enable interval mapping. First, a method of estimating genetic parameters (Table 1) for the parents used in the study was devised. Second, the interval mapping procedure described by Lander and Botstein (1989) was modified to accommodate heterogeneity of the parents.

Estimation of Parental Parameters

Since genotypes and phenotypes for the lines per se were assessed on samples of [F.sub.1]R[M.sub.10][S.sub.2] germplasm, the mating design was assumed to be equivalent to [F.sub.1]R[M.sub.10][S.sub.1]. Estimation of expected two-locus genotype frequencies in the [F.sub.1R[M.sub.10][S.sub.1], given initial values of the parental parameters listed in Table 1, requires the use of recursive formulae, such as those described by Moreno-Gonzalez (1993). Since direct estimates of the parental parameters are nonexistent, we seek derived estimates that provide a good fit to the observed genotypic frequencies in the [F.sub.1]R[M.sub.10][S.sub.1] generation. To provide these estimates, the quasi-Newton method in SAS PROC NLP (version 9.1 SAS/ OR) was employed. The equations used in the optimization procedure are given in Appendix A.

The quasi-Newton method was employed to obtain least-square estimates of the initial parental parameters at each marker interval. These estimates minimize the sum of squares of the deviations of observed genotypic frequencies in the final generation from those expected conditional on initial conditions and the mating scheme. As the last generation in the mating design to be analyzed was produced by selling, nine bilocus genotypes are possible providing eight degrees of freedom for parameter estimation. Details of options used in the optimization procedure, constraints imposed, and methods of obtaining starting estimates of parameters to be estimated are given in Appendix A.

Interval Mapping

Simple interval mapping in this study requires phenotypic data, estimates of parental parameters, and three-locus equations developed for a population cross followed by random-mating and selfing. BLUP estimates of line means for both testcross and per se data were used for the phenotypic data. Given estimates of the parental parameters, three locus equations for interval mapping in the the R[M.sub.10][S.sub.1] generation were derived as in Hospital et al. (1996). Three locus genotypes are shown in Table 2.

Simple interval mapping was then performed by extending previously established methods (Lander and Botstein, 1989) to accommodate non-homozygous parents and multiple generations of random mating followed by selfing. To address these additional needs, software was specially developed and implemented in Mathcad (version 2001i, MathSoft Engineering & Education, Inc., Cambridge, MA, www.mathcad.com). The key differences between standard interval mapping procedures (Lander and Botstein, 1989) and the method used in this study are described in Appendix B.

Permutation analysis (Churchill and Doerge, 1994) was used to ascertain the significance of QTL. Using parameters corresponding to the highest observed LOD score a random distribution of 1001 LOD scores was generated for each locus. An observed LOD score was considered significant if it exceeded the 1000th highest value in its corresponding random distribution. This level of stringency corresponds with an value of 0.001. Where a series of adjacent significant loci with effects in the same direction was observed a single QTL was assumed to be present at the point of maximum positive deviation of the observed LOD score from the permutation threshold LOD value.

Comparisons among Methods of Analysis

Results from the SIM procedure were compared with results from the maximum [R.sup.2] improvement (MAXR/BIC) procedure used by Laurie et al. (2004) and to the results from single marker analysis (SMA). For comparative purposes, regions were identified by SIM where there was a gap of at least 10 centimorgans between significant intervals. Markers were then identified which had been determined significant by MAXR/BIC or SMA and placed into the regions identified by SIM. Results are reported as number of regions significant. In some cases, two or more markers declared significant by MAXR/BIC or SMA fell into the same region. In such cases, those markers were counted as only one region. To be consistent with Laurie et al. (2004), a false discovery rate of 0.05 (equivalent to p = 0.014) was used for SMA. For SIM, a p-value of 0.001 based on permutation results was used because the number of regions identified as significant was similar to the numbers identified by MAXR/BIC and SMA.

Comparisons among Traits and Generations

Comparisons of oil, protein, and starch QTL were made by means of the results of the SIM analysis at the 0.001 probability level. As for comparisons among methods of analysis, regions were identified as having significant QTL for pairs of traits and the number of regions containing QTL affecting a pair of traits determined. In a similar manner, the number of regions having significant QTL in both the per se and testcross progenies was determined.

RESULTS AND DISCUSSION

Means, Genetic Variances, and Heritability

Line per se means for oil and protein were significantly higher than for testcrosses, whereas for starch, line per se means were significantly lower than those of the testcrosses (Table 3). Although significantly different, the per se and testcross means for protein were similar, with lines per se exhibiting a wider range of observations than testcross lines (Table 3). Genetic variance (the line component in Table 4) for lines per se is 10x that observed in the testcrosses for oil, 3x for protein, and 20x for starch. Although many of the line x environment variance components were significant (Table 4), they account for much less variation than the line effect. Heritability estimates on a family mean basis are high, indicating good potential for QTL detection (Table 5). The highest heritability values were observed for oil. Confidence intervals (CIs) for heritability for oil did not overlap with CIs for any other trait. There were no significant differences in heritability values between per se and testcross populations for oil and starch, but for protein, the upper limit for the testcross population was less than the lower limit for the per se population.

QTL Analysis Results

Statistical Methodology for SIM

Assumptions made in modeling the transmission of QTL in this experiment include diploid Mendelian inheritance and negligible natural selection or genetic drift during the process of random mating and selfing. Given prior experience with the germplasm the assumption of negligible natural selection appears valid (Dudley and Lambert, 2004). The large effective population size of ~200 plants limited the potential for genetic drift.

The multivariate optimization procedure was used to estimate parameters (see Table 1) required to perform interval mapping. Markers used in the study were chosen on the basis of a large frequency difference between a sample of IHO and ILO plants. The mean interpopulation marker frequency difference observed for these samples was 0.77 [+ or -] 0.01. This contrasts with the optimization result of 0.58 [+ or -] 0.01. The smaller difference estimated by the optimization procedure may be ascribed to the fact that the samples of IHO and ILO used to estimate marker frequency were not the same as the parents used to make the original cross. In addition, inadvertent selection for particular alleles (e.g., alleles related to nicking) during the ten generations of random mating could lead to differences in expected allele frequencies.

The optimization procedure provided recombination fraction estimates (on an [F.sub.2], rather than an expanded map basis) among markers, which correlate relatively well with the estimates obtained from the Monsanto composite map (r = 0.71, p < 0.01); however, the total map distances differ. Using Haldane's mapping function, the [F.sub.2] map length estimated in this study (2981 cM) is approximately twice the Monsanto composite map length (1410 cM). One explanation for this lies in the differing number of individuals and markers used to construct the composite map and this map. In developing the composite map, the majority of markers showed polymorphism in one or two of three mapping populations, two of which are of an [F.sub.2] mating structure and one consisting of [F.sub.2]-derived recombinant inbreds. This provided an average sample of genotypes equivalent to ~160 [F.sub.2] individuals. With 479 markers the possibility exists for few to no recombination events being detected among closely linked adjacent markers. When this occurs the lack of detection of recombination events would result in underestimation of the recombination fraction and an indefinite map order. In this study, 499 families and 10 generations of random mating were used, so recombination between closely linked markers was more likely to be detected. In addition, even a very small error rate in genotyping of samples for so many markers could lead to a large increase in apparent map distance (Lincoln and Lander, 1992). This is exacerbated by the fact that if the recombination fraction between two markers is on the order of 0.1, it would be very hard to detect any linkage between the two markers in an [F.sub.1]R[M.sub.10][S.sub.1] (Winkler et al., 2003). An error in marker order would also lead to inflation of erroneous intervals and therefore the overall map. These factors all lead to an increased estimate of map length.

In estimating [Q.sub.H] and [Q.sub.L], the SIM procedure extended current methodology by sampling the parameter space at fixed values. This was necessary because the parental germplasm contained unknown and varying QTL frequencies across loci. The theory here assumes that QTL frequencies which produce the most significant result do so because they are closest to the true frequencies and hence model the data best. An example of this concept can be found in the application of multiple regression in that, for a fixed number of factors, the better model is assumed to be the one which fits best (i.e., has the least residual variance).

Interval Mapping Results

Tables 6 through 8 contain a detailed summary of simple interval mapping results for oil, protein, and starch. Points listed are those at which the LOD + values (the amount an observed LOD score exceeds the permutation generated LOD value) are positive for either per se or testcross results and which are maximum for an adjacent set of points which all had positive LOD+ values.

Because in several cases the points listed were very closely linked, the significant points were grouped into regions by considering all significant points not separated by at least 10 cM (on the de novo map) as being in one region. Thus while for oil in the per se progenies, there are 70 significant points, (Table 6) these were grouped into 51 significant regions (Table 9).

For oil and starch, the number of significant regions detected by SIM in the per se and testcross lines was similar (Table 9) even though average absolute values of all additive effects in testcrosses were approximately 2/3 those in the lines per se for oil and 0.45 those for starch. This finding agrees with Laurie et al. (2004) and is the result of reduced error and interaction variances in the testcross progenies (Table 4). For protein, average additive effects for per se and testcrosses were similar. Protein and starch have a similar number of significant regions in both per se and testcrosses (Table 9). The number of significant regions (>40 for oil, protein, or starch in both per se and testcross populations) suggests a large number of QTL with small effects separate the parental strains for these traits.

Approximately 2/3 of the regions significant for oil were the same in the per se and testcross progenies (Table 9). For starch and protein, only approximately half of the regions were in common. This finding is in agreement with the correlations between testcross and per se results where the correlation for oil was 0.72 x x , whereas the correlations were 0.49 x x and 0.43 x x for starch and protein, respectively. Regions of most value for breeding may be those found in both populations. These would not only have a higher probability of being real but have LOD+ values larger on average than other QTL. For example the mean LOD+ value across all traits for QTL significant in both populations is 1.55 in comparison with a mean of 0.99 for QTL significant in only one population.

SMA Results vs. SIM Results

Single Marker Analysis

Results from SMA contain some ambiguity concerning number, location and effect of QTL within a region as these factors are confounded (Doerge et al., 1994; Whittaker et al., 1996; Bernardo, 2002) and colinearity among linked markers is not accounted for. SIM analysis is expected to provide more significance than SMA (Liu, 1997). Fifty, 40, and 66 makers were identified by SMA in per se lines for oil, protein, and starch, respectively (Fig. 1). In testcross lines respective numbers of markers identified by SMA were 63, 37, and 51. Because a number of markers in the SMA analysis were closely linked, the number of regions is less than the number of markers (Table 10). SMA identified only about 2/3 the number of regions identified by SIM.

[FIGURE 1 OMITTED]

For oil, SMA analysis did not detect significant marker QTL associations on chromosome 10 in the lines per se or testcrosses or on chromosome 8 in the testcrosses (Fig. 1). SIM identified QTL on all 10 chromosomes in both the per se and testcross populations (Table 6). SIM identified QTL on all 10 chromosomes for protein in both per se and testcrosses while SMA failed to identify significant associations on chromosomes 4, 8, and 9 in the testcross population. Similar results were obtained for starch. Given that the probability used for detection by SMA was 0.014 and that for SIM was 0.001, the SIM method appears to be quite powerful.

QTL Effects on Multiple Traits

Relationships among kernel composition traits can be inferred from correlations among them (Table 11), the proportion of regions having significant QTL for pairs of traits, and the signs of effects for the different traits. Oil and starch were highly negatively correlated, with significantly higher correlations for testcrosses than for lines per se (Table 11). This agrees with the SMA results in that the signs of the effects for QTL significant for oil in both per se and testcrosses were opposite those for starch. For SIM results, 56 of the 70 oil QTL listed in Table 6 have opposite signs of additive effect from starch in per se lines, whereas 55 of 69 oil QTL have additive effects of opposite signs from starch in testcross lines. Protein and starch are also highly negatively correlated and correlations are similar for lines per se and testcrosses. The negative correlations of oil with starch may result from changes in the germ to endosperm ratio as oil is primarily contained in the germ and starch in the endosperm.

Correlations of oil with protein were small but positive (Table 11). This was substantiated by both the SMA and SIM results. For SMA, additive effects for 10 of 21 QTL significant for oil in both per se and testcrosses had the same sign in both per se and testcrosses as QTL for protein. For SIM, signs of additive effects for 42 of the 70 oil QTL identified in per se lines (Table 6) are the same for protein, while signs of 41 of the 69 QTL identified in testcrosses are same for oil and protein.

Comparison with Other Studies

Berke and Rocheford (1995) evaluated 200 [S.sub.1] lines from a cross between an IHO cycle 90 plant and an ILO-derived plant. Single marker analysis for several traits was then performed on [S.sub.1] samples using 80 RFLPs. At the [alpha] = 0.01 significance level 36, 21, and 38 QTL markers were detected for oil, protein, and starch, respectively. For the 479 markers used here (averaged for per se and testcrosses) 56, 39, and 59 respective QTL markers were detected at the 0.01 level using SMA. The proportion of markers per trait found significant (0.45, 0.26, and 0.48 for oil, protein, and starch, respectively) by Berke and Rocheford (1995) differs from this study (0.12, 0.08, and 0.12). The reduced proportion of significant markers found in this study may be expected due to our use of random mated generations. A notable similarity is seen in the ratio of markers significant for oil and starch versus protein. In both studies, approximately 1.5 x as many markers were significant for oil and starch as for protein.

Using a multiple regression approach to association analysis, Laurie et al. (2004) analyzed the oil data from this study. Because of differences in the methods of analysis, some differences in results were expected. In those results, 49 per se and 39 testcross markers were identified as being associated with QTL, 15 of which were common to both populations. For comparison with SIM analysis, markers identified as significant in Laurie's analysis (labeled MAXR/BIC) were placed in regions determined by SIM analysis. Regions which contained markers significant in both the MAXR/BIC and SIM analyses were considered to have identified the same QTL. On this basis, for oil in the per se progenies, the 49 markers significant from Laurie's analysis grouped into 36 regions of which 22 were also significant in the SIM analysis (Table 12). For oil in the testcross progenies, the 39 markers significant by the MAXR/BIC analysis were grouped into 35 regions of which 26 were also significant in the SIM analysis. On the basis of individual markers, 61.2% of the 49 markers significant for oil in the per se lines were identified as being in a significant region by SIM and 71.8% of the 39 markers significant in the testcross progenies were include in significant SIM regions. These results represent reasonable agreement between the two methods given that the probability levels for the two analyses are likely different and that the SIM analysis does not take into account the multiple regression aspects of MAXR/BIC.

FINAL COMMENTS

In this study, we present for the first time a method of developing a molecular marker linkage map for a random-mated population resulting from a cross between selected, intermated populations along with a method for performing SIM using that map. Applying this method to [F.sub.1][RM.sub.10][S.sub.1] progeny and their testcrosses from a cross of IHO x ILO demonstrated that kernel oil, protein, and starch concentrations in this cross are controlled by a relatively large number of QTL, with each exerting a relatively small effect. A high negative correlation of oil with starch was observed and in most cases QTL significant in both traits had additive effects opposite in sign, suggesting development of high oil-high starch lines would be difficult. In contrast, the correlation between oil and protein was positive with 59.7% of all significant oil QTL having additive effects of the same sign as for protein. Thus selection for high oil and high protein should be possible.

The SIM results identifying 51 significant regions for oil in the per se progenies and 54 in the testcross progenies agree remarkably well with an estimate of 54 effective factors differentiating generations 76 of IHO and ILO obtained by Dudley (1977) using a quantitative genetics approach to estimating numbers of effective factors. The control of kernel quality traits by a large number of QTL with small effects has implications for breeding and for the potential efficacy of molecular genetic approaches to improvement of grain quality. Specifically, quality traits will be need to be treated as quantitative traits in a breeding program. Thus, backcrossing to increase oil, protein, or starch in a line will likely be difficult as will molecular approaches aimed at modifying a single gene to make major shifts in oil, protein, or starch concentration.

APPENDIX A

Equations used to estimate parental parameters were as follows where: AH = A allele frequency in the high population, BH = B allele frequency in the high population, DH = Linkage disequilibrium in the high population, AL = A allele frequency in the low population, BL = B allele frequency in the low population, DL = Linkage disequilibrium in the low population, r = recombination fraction between A and B, f1 - f9 = observed frequencies of genotypic classes (AABB, AABb, AAbb, AaBB, AaBb, Aabb, aaBB, aaBb, aabb) in [RM.sub.10][S.sub.1], t = number of random mated generations, and diff1 - diff9 = deviation of observed from expected genotypic frequencies.

A = AH + AL;

B = BH + BL;

AB = A + B;

W = (0.5 - r) x (0.5 - r) + 0.25;

X = (AH - AL) x (BH - BL) + DH + DL;

Y = AH x BH + AL x BL + DH + DL;

Z = (AH - AL) x (BH - BL);

E = [(1 - Z/(2 x X)) x [(1 - (1 - r).sup. x ] x (t - 1)) + r x [(1 - r).sup. x ] x (t - 1);

diff1 = - f1 + X x X x W/4 x E x E - X/4 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB + 2) x r + Y + 1) x E + 0.125 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r - (2 x Y x Y - 2 x (AB - 2) x Y) x r + (Y + 2) x Y);

diff2 = - f2 + X x X x W/-2 x E x E + X/2 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB + 1) x r + Y - A/2) x E - 0.25 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r (2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r + (Y - A) x Y);

diff3 = - f3 + X x X x W/4 x E x E - X/4 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB) x r + Y - A - 1) x E + 0.125 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r - 2 x (Y x Y - AB x Y + A x B) x r + (Y - A - 2) x (Y - A));

diff4 = - f4 + X x X x W/-2 x E x E + X/2 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB + 1) x r + Y - B/2) x E - 0.25 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r(2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r + (Y - B) x Y);

diff5 = - f5 + X x X x W x E x E - X x (2 x Y - AB + 1) x W x E + 0.5 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r - (2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r + Y x Y (AB - 1) x Y + A x B/2);

diff6 = - f6 + X x X x W/- 2 x E x E + X/2 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB + 1) x r + Y - A - B/2 + 1) x E 0.25 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r (2 x Y x Y- 2 x (AB - 1) x Y + A x B) x r + (Y - AB + 2) x (Y - A));

diff7 = - f7 + X x X x W/4 x E x E - X/4 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB) x r + Y - B - 1) x E + 0.125 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r - 2 x (Y x Y - AB x Y + A x B) x r + (Y - B - 2) x (Y - B));

diff8 = - f8 + X x X x W/-2 x E x E + X/2 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB + 1) x r + Y - A/2 - B + 1) x E - 0.25 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r - (2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r + (Y - AB + 2) x (Y-B));

diff9 = - f9 + X x X x W/4 x E x E - X/4 x ((2 x Y - AB + 1) x r x r - (2 x Y - AB + 2) x r + Y - AB + 3) x E + 0.125 x ((2 x Y x Y - 2 x (AB - 1) x Y + A x B) x r x r - 2 x (Y x Y - (AB - 2) x Y) x r + (Y - AB + 4) x (Y - AB + 2));

PROC NLP for the quasi-Newton algorithm options used were: the Broyden, Fletcher, Goldfarb, and Shanno (DBFGS) update method, INHESSIAN (sets the initial estimate of the approximate Hessian to the cross product Jacobian) and LSPRECISION = 0.1 [specifies the degree of accuracy that should be obtained with updates (Fletcher 1987)]. Defaults for all other options were accepted. Available options are explained in the version 9.1 SAS/OR documentation.

Prior information was used to set initial values of the parental parameters. For allele frequencies (AH, AL, BH, BL) initial values were taken as the estimate obtained from genotyping 24 random individuals each from generations 70 of IHO and ILO. For the recombination fraction, results from the Monsanto composite map were used. Initial linkage disequilibrium values within IHO and ILO were set at zero.

Further efforts to obtain accurate solutions were made by placing appropriate constraints on all parameters. The recombination fraction (r) was constrained linearly to be between 0 and 0.9999. Although recombination fractions with linkage are only valid in the interval [0, 0.5], the wider constraint interval allowed the algorithm to vary outside the valid range in search of an optimal solution. This allowance however was later observed to produce the same results had the variable been constrained to the valid interval. Linkage disequilibrium parameters were constrained according to their nonlinear dependence on allele frequencies. This dependence is established as follows.

Consider the array of gametes possible in the bilocus case. Each of the four gametic frequencies is naturally constrained to the interval [0, 1], as are the allele frequencies. This produces the following inequality, where A and B are frequencies of particular alleles at each of the loci A and B, and D is the gametic linkage disequilibrium (Lynch and Walsh, 1998) associated with loci A and B.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

Solving for D in each case results in

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

.D must be [greater than or equal to] the maximum of the four values on the left, and [less than or equal to] the minimum of the four values on the right. For the left side, in all cases:

max[A(1 - B) - 1, (1 - A)B - 1] [less than or equal to] min[-A x B, - (1 - A)(1 - B)].

Likewise, for the right side:

max[A(1 - B), (1 - A)B] [less than or equal to] min[1 - A x B, 1 - (1 - A)(1 - B)].

Therefore, D is effectively constrained as:

max[- A x B, - (1 - A)(1 - B)] [less than or equal to] D [less than or equal to] min[A(1 - B), (1 - A)B] (Weir, 1996).

A 99.9% confidence interval was used as a linear constraint on allele frequencies. To construct a confidence interval about the allele frequency parameters the specifics of this particular experiment were considered. An estimated five to seven plants each from IHO and ILO were chosen at random to form the mapping population. That is to say, from each parental population a sample of [y.sub.par] copies of one particular allele was taken at random from a total of 10, 12 or 14 alleles ([n.sub.par]) resulting in an actual allele frequency of [y.sub.par]/[n.sub.par]. The probability of observing a certain allele frequency in a sample of parent individuals depends on two separate probabilities: one associated with the true allele frequency in IHO and ILO and one associated with the sampling of parent individuals from IHO or ILO. Both are calculated using the binomial probability distribution function and the overall joint probability becomes the product of the two separate probabilities.

For the first probability, information about IHO and ILO was available from marker genotypes of 24 random individuals from each population. This genotyping produced unbiased frequency estimates with a sample size equal to 48 alleles. Using this information the first of the two probabilities is

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

where P(v) is the probability of obtaining an allele sample of [y.sub.pop] copies of one particular allele out of [n.sub.pop] alleles in IHO or ILO given a true population allele frequency of v. In this case, we considered the value of v obtained from the sample of 24 plants in IHO and ILO as the true value of v and calculated P(v) based on [n.sub.pop] = 48. The second probability is

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

where [n.sub.par], takes values of 10, 12, and 14 (representing 5, 6, and 7 parent plants) while [y.sub.par] varies across the integers [0, [n.sub.par]. Finally, a 99.9% confidence interval was calculated by minimizing and maximizing [y.sub.par]/[n.sub.par] subject to the joint probability. With the exception of markers located on the end of a chromosome, the optimization procedure estimates each allele frequency twice. Consequently, a single allele frequency estimate was taken to be the mean of the two estimates, weighted by the inverse of the associated errors. Obtaining parameter estimates for the actual parents used allowed for subsequent interval mapping.

APPENDIX B

Interval mapping requires specification of QTL genotypic frequencies conditional on flanking marker genotypic frequencies and recombination frequencies between the QTL locus and the marker loci. This is the point at which the SIM procedure used in this study differs from the Lander and Botstein (1989) approach. As pointed out before, current theory assumes unequivocal assignment of a QTL allele with a frequency of 1.0 to either parent. Since this was not the case, the entire QTL allele frequency space was sampled. This consisted of calculating LOD scores conditional of observed phenotypic and initial marker parental parameter estimates where the QTL allele (Q) frequency in IHO and in ILO was allowed to vary from 0 to 1 in increments of 0.1. Erratic results were obtained when the QTL frequencies in IHO and ILO were set equal to one another. When frequencies were equal, the system of equations became ill conditioned (i.e., approached singularity) providing no unique solution when estimating the [[mu.sub.j] s. Accordingly, such combinations were omitted.

For each locus tested, the parameter set giving the highest LOD score was used to estimate additive and dominance effects in the lines per se and additive effects in the testcross lines. For the per se population, a QTL additive effect was taken to be half the phenotypic difference between QTL homozygote values and a dominance effect was taken to be the deviation of the QTL heterozygote value from the mean of the QTL homozygote values, while for the testcross population, a QTL additive effect was equal to the phenotypic difference between the QTL homozygote values, because the two homozygous classes AA and aa would generate Ax and ax genotypes, where x is the tester allele.

ACKNOWLEDGMENTS

We gratefully acknowledge the assistance of Cathy Laurie, Scott Chasalow, Lyle Crossland, Donald Roberts, and Randy Rich. Special thanks to G.R. Johnson for a critical review and valuable suggestions.

REFERENCES

Berke, T., and T.R. Rocheford. 1995. Quantitative trait loci for flowering, plant and ear height, and kernel traits in maize. Crop Sci. 35: 1542-1549.

Bernardo, R. 2002. Quantitative traits in plants, p. 291. Stemma Press, Woodbury, MN.

Churchill, G.A., and R.W. Doerge. 1994. Empirical threshold values for quantitative trait mapping. Genetics 138:963-971.

Darvasi, A., and M. Soller. 1995. Advanced intercross lines, and experimental population for fine genetic mapping. Genetics 141: 1199-1207.

Doerge, R.W., Z.B. Zeng, and B.S. Weir. 1994. Statistical issues in the search for genes affecting quantitative traits in populations, p. 15-26. In Analysis of molecular marker data (Supplement). Joint Plant Breed. Symp. Ser., Am. Soc. Hort. Sci., CSSA, Madison, WI.

Dudley, J.W. 1977. Seventy-six generation of selection for oil and protein percentage in maize, p. 459-473. In E. Pollak et al (ed.) Proc. Intl. Conf. on Quantitative Genetics. Iowa State Univ. Press, Ames.

Dudley, J.W., A. Dijkhuizen, C. Paul, S.T. Coates, and T.R. Rocheford. 2004. Effects of random-mating on marker-QTL associations in the cross of the Illinois High Protein X Illinois Low Protein maize strains. Crop Sci. 44:1419-1428.

Dudley, J.W., and R.J. Lambert. 1992. Ninety generations of selection for oil and protein in maize. Maydica 37:1-7.

Dudley, J.W., and R.J. Lambert. 2004. 100 Generations of selection for oil and protein in corn. Plant Breed. Rev. 24(Part 1):79-110.

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 in maize. CSSA, Madison, WI.

Dyer, D., and P. Feng. 1997. NIR destined to be a major analytical influence. Feedstuffs 69:16-25.

Fletcher, R. 1987. Practical methods of optimization, Second ed. John Wiley & Sons, Chichester:, UK.

Holland, J.B., V.A. Portyanko, D.L. Hoffman, and M. Lee. 2002. Genomic regions controlling vernalization and photoperiod responses in oat. Theor. Appl. Genet. 105:113-126.

Hospital, F., C. Dillmann, and A.E. Melchinger. 1996. A general algorithm to compute multilocus genotype frequencies under various mating systems. Comput. Appl. Biosci. 12:455-462.

Hymowitz, T., J.W. Dudley, F.I. Collins, and C.M. Brown. 1974. Estimations of protein and oil concentration in corn, soybean, and oat seed by near-infrared light reflectance. Crop Sci. 14:167-170.

Kearsey, M.J., and V. Hyne. 1994. QTL analysis: A simple 'marker-regression' approach. Theor. Appl. Genet. 89:698-702.

Kendall, M., and A. Stuart. 1977. The advanced theory of statistics. p. 250. Vol. 1. 4th ed. Macmillan, New York.

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.

Lambert, R.J. 1994. High-oil corn hybrids, p. 123-145. In A. R. Hallauer (ed.) Specialty corns. CRC Press, London.

Lambert, R.J., D.E. Alexander, and I.J. Mejaya. 2004. Single kernel selection for increased grain oil in maize synthetics and high-oil hybrid development. Plant Breed. Rev. 24:153-175.

Lander, E.S., and D. Botstein. 1989. Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185-200.

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.

Lincoln, S., and E. Lander. 1992. Systematic detection of errors in genetic linkage data. Genomics 14:604-610.

Liu, B.H. 1997. Statistical genomics: Linkage, mapping, and QTL analysis CRC press. Boca Raton, FL.

Lynch, M., and B. Walsh. 1998. Genetics and analysis of quantitative traits, p. 819. Sinauer Associates, Inc., Sunderland, MA.

Mazur, B., E. Krebbers, and S. Tingey. 1999. Gene discovery and product development for grain quality traits. Science (Washington, DC) 285:372-375.

Moreno-Gonzalez, J. 1993. Relationships for linked loci in advanced random and nonrandom mating populations. J. Genet. Breed. 47:9-14.

National Corn Growers Association. 2001. The World of Corn. National Corn Growers Association, St. Louis, MO.

Nyquist, W.E. 1991. Estimation of heritability and prediction of selection response in plant populations. Crit. Rev. Plant Sci. 10:235-322.

Patterson, H.D., and E.R. Williams. 1976. A new class of resolvable incomplete block designs. Biometrika 63:83-92.

Sadimantara, G.R., T. Abe, and T. Sasahara. 1997. Genetic analysis of high molecular weight proteins in rice (Oryza sativa L.) endosperm. Crop Sci. 37:1177-1180.

Tan, S.L., and W.R. Morrison. 1979. Lipids in the germ, endosperm and pericarp of the developing maize kernel. J. Am. Oil Chem. Soc. 56:759-764.

Tan, Y.F., J.X. Li, S.B. Yu, Y.Z. Xing, C.G. Xu, and Q. Zhang. 1999. The three important traits for cooking and eating quality of rice grains are controlled by a single locus in an elite rice hybrid, Shanyou 63. Theor. Appl. Genet. 99:642-648.

Wang, X.L., and B.A. Larkins. 2001. Genetic analysis of amino acid accumulation in opaque-2 maize endosperm. Plant Physiol. 125: 1766-1777.

Whittaker, J.C., R. Thompson, and P.M. Visscher. 1996. On the mapping of QTL by regression of phenotypes on marker-type. Heredity 77:23-32.

Weir, B.S. 1996. Genetic data analysis II. Sinauer Associates, Sunderland, MA.

Willmot, D.B., J.W. Dudley, T.R. Rocheford, and Al Bari. 2006. Effect of random mating on marker-QTL associations for grain quality traits in the cross of Illinois High Oil X Illinois Low Oil. Maydica (In press).

Winkler, C.R., N.M. Jensen, M. Cooper, D.W. Podlich, and O.S. Smith. 2003. On the determination of recombination rates in intermated recombinant inbred populations. Genetics 164:741-745.

Xu, C., X. He, and S. Xu. 2003. Mapping quantitative trait loci underlying triploid endosperm traits. Heredity 90:228-235.

doi:10.2135/cropsci2005.06-0153

Darryl Clark, John W. Dudley, * Torbert R. Rocheford, and John R. LeDeaux

Darryl Clark, John W. Dudley, and Torbert R. Rocheford, Dep. of Crop Science, Univ. of Illinois, Urbana, IL 61801; John R. LeDeaux, Monsanto Co., St. Louis, MO 63167. This research was supported by the Illinois AES and a grant from Renessen, LLC. Received 22 June 2005. * Corresponding author (jdudley@uiuc.edu).
Table 1. Population parameters characterizing the bilocus model.

Parameter    Definition

[A.sub.H]    Frequency of one of the two alleles at locus A in the high
               population (IHO)
[B.sub.H]    Frequency of one of the two alleles at locus B in the high
               population (HO)
[D.sub.H]    Gametic linkage disequilibrium in the high population
               (IHO)
[A.sub.L]    Frequency of one of the two alleles at locus A in the low
               population (ILO)
[B.sub.L]    Frequency of one of the two alleles at locus B in the low
               population (ILO)
[D.sub.L]    Gametic linkage disequilibrium in the low population (ILO)
[theta]      Recombination fraction between loci A and B (presumed
               equal in 1110 and ILO)

Table 2. A matrix showing trilocus genotypes as products of
trilocus gametes fisted in the first row and column. Each cell
represents the expected frequency of that genotype arising
from a particular combination of corresponding gametes.

        AQB       AQb       AqB       Aqb

AQB    AAQQBB    AAQQBb    AAQqBB    AAQqBb
AQb    AAQQBb    AAQQbb    AAQqBb    AAQqbb
AqB    AAQqBB    AAQqBb    AAqqBB    AAqqBb
Aqb    AAQqBb    AAQqbb    AAqqBb    AAqqbb
aQB    AaQQBB    AaQQBb    AaQqBB    AaQqBb
aQb    AaQQBb    AaQQbb    AaQqBb    AaQqbb
aqB    AaQqBB    AaQqBb    AaqqBB    AaqqBb
aqb    AaQqBb    AaQqbb    AaqqBb    Aaqqbb

        aQB       aQb       aqB       aqb

AQB    AaQQBB    AaQQBb    AaQqBB    AaQqBb
AQb    AaQQBb    AaQQbb    AaQqBb    AaQqbb
AqB    AaQqBB    AaQqBb    AaqqBB    AagqBb
Aqb    AaQqBb    AaQqbb    AaqqBb    Aaqqbb
aQB    AaQqBb    aaQQBb    aaQqBB    aaQqBb
aQb    aaQQBb    aaQQbb    aaQqBb    aaQqbb
aqB    aaQqBB    aaQqBb    aaqqBB    aagqBb
aqb    aaQqBb    aaQqbb    aaqqBb    aaqqbb

Table 3. Means ([+ or -] standard deviation), ranges, and coefficients
of variation (CV) for oil, protein, and starch in per se (PS) and
testcross (TC) populations.

Trait      Population           Mean             Range      CV

                                    g [kg.sup.-1]            %

oil            PS        70.6 [+ or -] 0.13    43.4-103     9.7
               TC        49.2 [+ or -] 0.04    41.5-60.1    4.5
Protein        PS         129 [+ or -] 0.14     106-154     6.1
               TC         126 [+ or -] 0.11     115-136     5.7
Starch         PS         585 [+ or -] 0.52     490-644     4.4
               TC         677 [+ or -] 0.12     657-690     1.0

Table 4. Estimated components of variance for oil, protein, and
starch in per se (PS) and testcross (TC) populations.

Variance
Component                  Population      Oil      Protein    Starch

                                            [(g [kg.sup.-1]).sup.2]

Line                           PS         119 **    71.7 **     716 **
                               TC        10.2 **    21.4 **    33.2 **
Line x year                    PS        6.67 **    5.83 **     116 **
                               TC        0.44 **    1.66 *     4.22 **
Line x location                PS        1.11       3.53 **    28.3 *
                               TC        0.26 **    0.8        1.65 *
Line x year x location         PS        4.99 **    3.07       85.4 **
                               TC        0.00       6.58 **    2.09
Residual                       PS        47.3 **    62.3 **     661 **
                               TC        4.88 **    50.9 **    42.9 **

* Significant at p < 0.05.

** Significant at p < 0.01.

Table 5. Heritability estimates and 95% confidence intervals
for oil, protein, and starch in per se (PS) and testcross (TC)
populations.

Trait      Population    [H.sup.2]    Lower limit    Upper limit

Oil            PS          0.93          0.92           0.94
               TC          0.94          0.92           0.94
Protein        PS          0.88          0.86           0.90
               TC          0.77          0.73           0.80
Starch         PS          0.84          0.81           0.86
               TC          0.83          0.81           0.86

Table 6. Oil QTL found to be significant in per se and/or testcross
lines by simple interval mapping. "Chrom" is the chromosome
number, "Marker Number" indicates the marker nearest the
QTL, and "Marker Position" gives a marker's genetic map
position in centimorgans. "QTL Distance" is the QTL distance
relative to the nearest marker in cM. A negative value indicates
the QTL is toward the beginning of the chromosome. "LOD +"
refers to the amount an observed LOD score exceeds the
permutation generated LOD threshold value. "a" and "d" are
the estimated additive and dominance effects, respectively.

                                                 Per se
        Marker    Marker      QTL
Chrom   number   position   distance   LOD +        a          d

                                               g [kg.sup.-1]

1          3        18       -0.3       0.51       -4.8        3.9
1          4        34        0.7       0.88        3.2       18.2
1         15       119        0.6       0.05        3.8        9.3
1         20       132       -1.0      -1.13        4.7        5.4
1         22       132        4.2       1.47        0.2       12.4
1         28       175        0.1       0.36        5.4        0.7
1         35       204       -0.2       0.01       -3.5       16.6
1         41       213        2.3       0.84        6.5        0.9
1         52       260       -2.6      -1.33        2.4        8.9
1         59       303        1.6      -0.06       -1         -2.3
1         61       326       -5.2      -0.03       11.1      -11.5
1         63       328        0.9       0.60        6         -0.6
1         67       338        1.3      -0.20        0.9        2.2
1         68       344       -1.6       0.40        5.3        2.6
1         70       344        5.7      -0.18        6         -2.4
1         73       402       -1.0      -1.87        7.3       -1.8
1         76       431       -2.0      -2.79        6.1        2.8
1         79       452        5.8      -1.59       -5.1       -0.1
2         84         3       -0.7      -2.09       -3.1       16.2
2         88        25        0.3       0.53        7.9       -3.8
2         92        46       -1.1       0.09       -5.1        1.2
2         95        54        0.4      -0.54       -5.2       -1.1
2         95        54        1.4       0.56       -5.7       -1.8
2         96        63        0.7      -2.54        2.2       -5.5
2         97        69       -1.9       0.23        5.8       -1.2
2        100        77        0.7       0.31        5         -2.3
2        101        84       -0.1      -0.04       -5.1        2.2
2        103        92       -1.7       1.68        4         -6.7
2        105       100       -1.2      -1.52        6          1.8
2        108       116        0.0      -0.41       -5.7        1.2
2        112       153        8.4       0.51        7.3       -4.3
2        115       180       -0.7      -2.27        8         -0.1
2        120       198       -0.4       0.42       -0.7       10.7
2        120       198        3.7      -2.46        0.6       14.2
2        126       232        2.6      -1.34       -7         -7.6
2        126       232        3.6       0.31       -5.9       -5.8
2        126       232        4.7      -1.29       -8.5      -12.7
3        131         3        0.3       0.17       -3.9        7.5
3        131         3        2.4      -0.06       -4.8        9.2
3        133        20       -0.2      -0.64        4.8        4.8
3        138        74        1.9      -1.76        6.3        3.2
3        138        74        8.6       0.40        5.4        0.5
3        139        95       -1.2      -0.32        9.8       -4.6
3        139        95        1.9       0.41        5.5       -9.7
3        142       132       -0.1      -0.09        6.2       -0.6
3        143       142       -1.6       1.66        6         -3.7
3        154       203       -0.1      -1.73        3.9        6.2
3        157       212       -0.1       2.29        6.3        1.8
3        170       227        2.9       0.16       -8.9      -13.6
3        182       299       -0.6      -2.14        5.4        0.7
3        187       306       -0.9       0.43        6.1        4.1
4        189         0        2.0      -0.98        5.7        2.5
4        189         0        6.4       0.70        5.5       -0.8
4        191        47       -4.5       0.87        5.1       -1.1
4        191        47       -0.3       1.56        5.2        0.0
4        193        58       -1.2       1.72        5.6        5.8
4        194        63       -1.1       1.16       -0.6       12.8
4        214       168        0.6       0.52       -6.8       -0.9
4        215       178       -1.8       0.70        8.1       -8.3
4        218       182       -1.0       0.92        8.2       -8.4
4        221       187        1.5      -0.07       15.4      -44.8
4        232       233      -11.3       0.09        7.4       -9.8
4        232       233       -2.2       0.42        5.5       -9.8
4        234       260        0.3      -1.06        5.4        2.7
4        237       275       -0.8       1.91        1.6       17.9
4        240       286        0.0       2.70        5.2        0.7
5        246         0        2.0       0.43        5.7       -6.4
5        246         0        4.2      -1.18        6.2       -8.9
5        254        52        0.4       4.57        5         -0.8
5        255        63       -0.2       0.42        4.7        3.4
5        256        97        0.2       3.54        5.4       -1.3
5        258       101        0.8       2.59        3.8        6.6
5        259       113       -1.6       0.52        4.6        5.0
5        260       119       -2.3      -3.96       -8.7       -7.7
5        265       140       -2.0       4.21        5.2        1.1
5        269       157        0.9      -0.08       28.2        6.0
5        273       223       -0.3       0.91        6.1       -2.0
5        273       223        1.8       1.34        6.4       -5.7
5        279       324        1.3      -1.80        5.5        1.8
5        284       368       -0.1      -2.39       -7.9       -4.1
5        286       426       -1.1       0.81       -7.6       -0.7
5        286       426        5.2      -1.76        3.1        7.9
5        287       441        1.0       0.08        4.1        5.5
5        288       465      -10.3      -0.04       16.7        4.6
5        288       465        3.8       0.95        8.1       -4.6
5        289       478        4.6      -1.71        5.2       -5.6
5        289       478        5.7       0.42        5.9       -6.0
5        290       528       -4.2       0.62        5.3        1.5
6        309        58       -2.1       0.03        3.8       17.5
6        311        69        0.2       5.64        5.2        3.7
6        318        86       -1.5       2.67        3.9        6.7
6        319        87        2.5       1.67        5.7       -3.8
6        320        93       -1.1      -0.92        5.2       -5.6
6        331       131       -5.0      -1.61       -7.3        3.3
6        331       131       -3.9       0.34       -6.7       -0.9
6        332       136       -2.0       0.14        6.6       11.6
6        334       163       -3.2       1.73        5.5       -0.7
7        348        46       -2.9       0.25       -7.3        3.9
7        355        87        0.8       1.40        5.5        0.2
7        364       132       -0.9      -1.71        6.6       -0.2
7        365       132        0.1       0.58        8.1       -5.0
7        369       170        1.1      -0.84       -4.4        3.9
7        375       252       -1.8      -0.06       -4.2      -22.0
7        377       266        0.2       0.37       -8.2       -4.0
8        386        28       -1.3      -0.53       -6.1       -3.1
8        387        55       -0.7       0.21        5.2        1.2
8        391       122       -4.3      -0.11        3.4       10.9
8        391       122       -2.1       0.23        6.4        5.8
8        402       231       -4.0       2.10        4.5        4.0
8        412       274       -2.0       2.21       -0.3       10.6
9        419        27       -5.2       0.55        5.8       -6.5
9        422        67        6.2       0.40        0.5       13.4
9        433       127       -8.2      -1.86        5.2        1.9
9        443       149        6.7      -1.51        4.8        3.6
9        446       236       -5.8      -2.64      -10.7       -5.5
10       453        37       -1.2       0.58        3         10.3
10       453        37       -0.2       0.55        2.7       12.9
10       468        62        2.4      -1.73       -6.2        0.4
10       471        84       -0.8      -3.80       -5.2      -10.6
10       479       102       -0.7      -2.49       -7.6        2.1

            Testcross

Chrom   LOD +         a

                 g [kg.up.-1

1        -2.62        3
1        -2.05        3.4
1        -2.74        2.4
1         2.69        3.2
1        -0.02        2.8
1        -1.79        2.6
1         0.01       -3
1        -1.62        3.4
1         0.02       -3.8
1         0.87        3.8
1         0.60        3.2
1        -3.50        3.8
1         0.03       -3
1        -1.74        4
1         0.40        4.2
1         1.09        3.4
1         0.16        2.8
1         0.27       -3.2
2         0.95        2.8
2         1.56        3.2
2         0.17       -2.8
2         0.58       -2.6
2        -1.06       -2
2         1.44        2.2
2         1.37        2.4
2        -1.39        3.6
2         0.32       -5.4
2        -1.45        2.6
2         0.76        4.8
2         1.45        3.4
2        -1.60        4
2         0.74        3.4
2        -1.69        2.2
2         0.21        5.4
2         0.20        3.4
2        -0.16        3.6
2         1.06        3
3        -0.48       -1.2
3         0.27        0
3         0.84        0.2
3         3.04        3
3        -0.03       15
3         0.85        4.4
3        -0.01       13.2
3         7.29        3.2
3         4.54        2.8
3         3.44        4.2
3         3.05        3.2
3        -1.62        3.6
3         0.89        7.4
3        -2.44        3
4         0.16       -3.2
4        -0.84       -3
4         1.07        3
4        -0.14        2.8
4        -1.28        2.8
4         1.58        0
4        -3.32       -1.6
4        -0.77        5.2
4        -2.66        7.6
4         0.30        3
4        -0.04        6
4        -0.04        8.8
4         0.34        2
4        -2.34       -0.8
4        -0.32        4.2
5        -0.13        2.8
5         0.40        3.4
5         3.57        2.8
5        -1.73        3.2
5         0.15        2.6
5        -1.14        1.2
5        -1.22        2.8
5         0.43        3.4
5         5.05        3
5         1.67        2.8
5         1.00        3.2
5        -0.44        3
5         0.53        3.8
5         0.22       -3.2
5        -1.04        3.2
5         0.57        2.2
5        -1.77        3.8
5         1.25        3.2
5         0.04        4.6
5         1.57        3.6
5        -0.65        3
5        -3.91       15.2
6        -3.55        2.2
6         8.62        3
6        -0.14        3.2
6         3.08        3.2
6         5.17        3.2
6         1.96        3.2
6         0.16        2.8
6        -2.73        2.4
6        -1.30        3.4
7        -3.89       -0.2
7        -0.88        3
7         0.52        2.8
7        -0.01       14.6
7         1.43       -2.6
7         0.04        3.2
7         0.11       -4.6
8         0.32        3.4
8        -0.83        3.2
8         1.59        5.4
8        -0.15        4.2
8        -0.46        3.8
8         1.50       -0.6
9        -0.91        5.8
9        -5.80        4
9         0.30        4.2
9         0.89        3
9         0.49       -3.4
10       -0.53        1.8
10        0.77        1.6
10        0.49       -4.2
10        0.97       -4
10        0.74       -3.8

Table 7. Protein QTL found to be significant in per se and/or
testcross lines by simple interval mapping. "Chrom" is the
chromosome number, "Marker Number" indicates the marker
nearest the QTL, and "Marker Position" gives a marker's
genetic map position in centimorgans. "QTL Distance" is the
QTL distance relative to the nearest marker in cM. A negative
value indicates the QTL is toward the beginning of the chro-
mosome. "LOD +" refers to the amount an observed LOD
score exceeds the permutation generated LOD threshold
value. "a" and "d" are the estimated additive and dominance
effects, respectively.

                                                      Per se
         Marker     Marker       QTL
Chrom    number    position    distance    LOD +         a           d

                                                   g [kg.sup.-1]

1           3          18        -1.3      -1.84       -0.8        -13
1           4          34         2.7      -0.04        2.8        -5.1
1           6          41         2.7       0.91        4.1         1.5
1           9          85         1.0       0.10       -4          -0.9
1          16         130        -1.3       0.28        4          -0.9
1          25         150         1.3       1.42       -3.9        -2.4
1          29         177        -0.6       1.27       -4          -0.4
1          32         193        -0.8       1.86       -4          -1.2
1          47         233         0.0      -1.95       -4.6        -4.5
1          50         243         3.6      -2.58       -1.1        -7.6
1          53         279        -3.6      -1.62        3.7        -2.9
1          59         303        -4.7      -0.05       -3.1         5.5
1          60         307        -0.6       0.30        3.9         0.2
1          61         326         0.0       0.05        1.1         8.1
1          67         338         0.3       1.28        3          -4.2
1          70         344        13.0      -0.07        8.1        -6.7
1          72         393         4.0      -1.10       -4           0.6
1          75         427        -0.7      -1.41        3.6        -4
1          77         435         2.7      -0.57       -4.8        -3.2
1          77         435         6.0       0.02       -3.8        -2.5
2          83           0         0.0      -0.43        3.9        -1.7
2          86          20         0.6       2.02        3.9        -1.3
2          93          47         1.1       0.41        2.8        -5.2
2          97          69        -0.8       0.92       -4.1        -4.4
2          98          73        -0.9      -0.04        3.7        -2.5
2         101          84        -0.1       1.00       -3.6         3
2         104          92         2.4       0.18       -1.1        -7.6
2         110         146       -10.4      -0.05        6.2         3.9
2         114         177         0.0       1.24        3.5         4.1
2         114         177         1.0       0.44        3.4         3.7
2         119         188         0.6      -0.01        0          -1.8
3         133          20         2.9      -0.15        0.2         0.7
3         136          69         1.8      -0.04        4.1         0.7
3         138          74         6.3      -0.56        3.8        -1.3
3         139          95        -0.1      -0.58        1.9        -8.1
3         142         132        -2.1      -0.65        3.9         1.8
3         143         142        -1.6       0.82        4.1        -0.4
3         143         142        -0.6       0.99        4          -1.1
3         149         182       -11.5      -0.04       10.3         1.2
3         154         203        -0.1       3.06        3.2        -4.7
3         155         203         3.0       1.70        2.9        -5.3
3         172         238        -0.2       0.44       -6.4         7.4
3         178         249         6.2      -1.69       -4           0.8
3         179         289        -1.1      -1.13        1.7         6.9
3         185         299         0.4       0.66        3.7         2.9
3         188         324        -7.4      -2.58       -2.9        -1.5
4         190          22         6.3       0.11        3.2        -8.3
4         193          58        -1.2      -0.16       -4.5        -0.7
4         201         103        -2.6      -1.32        3.2        -2.4
4         210         144        -3.5      -3.12       -0.2         0.8
4         212         161        -0.4       0.30        5.2         0.5
4         212         161         1.7      -1.39       -5.9        -3.2
4         223         195         4.3       0.49        4.3         2.9
4         231         207         1.2       1.04       -4.6        -4.8
4         232         233         0.9       0.27       -5          -4.6
4         233         245        -3.6      -0.96       -2.6        -6.2
5         252          21         3.6       0.01       -3.8         2.5
5         253          32        -3.1      -0.13       -3.8         3.9
5         255          63         7.3       0.09       -3.5         3.9
5         256          97        -5.1       0.06       -2.6         5.8
5         257         101        -0.2      -1.31        3.4         3.3
5         263         126         0.2       0.95        4.9         4.4
5         264         131        -1.9      -1.21        2.4         7.6
5         273         223        -2.3      -0.60       -8.3        17.7
5         280         338        -1.3      -2.70        3.3        -2.3
5         280         338         3.9       0.13        4.2        -1.4
5         283         362        -2.0      -0.03       11.1        19.3
5         286         426        -2.1       1.05       -4.7        -0.4
5         287         441        -0.1      -0.05        6.4       -17
5         288         465         0.7       3.46        2.6        -5.9
5         289         478        -4.9       2.39        3.1        -4.6
6         309          58        -0.1      -0.69       -3.6         2.6
6         318          86        -2.6      -0.81        1.2         8.3
6         322          94         0.3      -2.78        3.5         3.5
6         325         108        -1.2      -1.22       -2.8         5.1
6         332         136         4.2       0.12       -9.2         -17
6         333         146        -2.7      -1.13        3.6        -2.9
6         337         185        -1.6       0.47        3.6         1.8
7         342          20        -3.1       0.27        5.6        -0.2
7         348          46         0.2       0.14        4.5        -0.8
7         364         132        -0.9      -3.22        3.1        -0.2
7         368         137         6.8       0.71        4.5        -4.1
7         369         170         1.1       1.26       -3.2         4
7         370         205       -10.0      -0.01      -15.8         7.1
7         373         220        -2.1       0.51       -3.6        -4.2
7         375         252       -10.9       0.46        2.6        -6.2
7         383         290        -1.4      -0.56       -4.2        -2.3
8         385          23        -9.5       0.22       -4.3         0.4
8         389         105       -15.8      -2.50       17.2        -5.7
8         389         105         0.5      -1.21        3.9        -3.9
8         393         131        -0.6       4.27        4           2.6
8         394         136         0.4       4.51        3.5         3.9
8         406         247        -2.2       0.64        2.8        -5.3
8         409         247         7.4      -0.86        3.9         0.7
8         414         291        -1.5      -2.81       -2.7        -1.9
9         418          11        -3.8      -3.68        0.1         0.3
9         419          27         4.3      -2.55        1.8        10.1
9         420          50         3.0       1.04       -4           1.6
9         422          67         0.8       0.38       -8.5         0.5
9         422          67         6.2      -1.04       -5.4        -2.3
9         426          82         1.2       1.05       -3.6        -5.7
9         427          86         1.4      -1.23       -3.5       -11
9         432         102         6.9      -1.58       -3.7        -2.7
9         433         127        -4.8      -0.86       -4          -1.4
10        452           9         0.6       0.05        4.3         4
10        453          37        -7.7      -2.95       18.8         5.5
10        453          37        -0.2       0.71        3.5         3.6
10        472          95        -0.6       0.23       -4.5        -2.3
10        478          99         1.2       0.95        7.5        18.9

               Testcross

Chrom    LOD +         a

                  g [kg.sup.-1]

1         0.08         4.6
1         1.52         3.8
1        -0.07         1.2
1        -2.63         2.8
1        -3.83         2.2
1        -2.49        -2.2
1        -3.25        -2.4
1        -3.63        -3.2
1         0.50         3.4
1         0.46         3.4
1         0.61         4.2
1         2.99         4.4
1         1.68         3.8
1        -1.24        -2.4
1        -1.94         3.2
1         0.21         4.8
1         0.07        -3.6
1         0.25         3.4
1         0.07         4.2
1        -1.31         4
2         0.78         4
2         0.64         4
2        -1.03         3.8
2        -2.63        -3.2
2         2.05        -1.8
2        -0.18        -3.2
2        -1.98         2.4
2         0.61         2
2        -0.88         3.6
2         0.55         4
2         0.02         3.6
3         0.92         3.8
3         1.53         4.6
3         0.31         4
3         1.05         3.8
3         0.91         3.4
3         0.77         3.8
3        -0.62         3.8
3         0.57         4
3         0.36         4.8
3         4.37         4.2
3        -2.82        -2.8
3         0.31         5.4
3         1.72         4
3        -0.11         2.8
3         0.33        -5.4
4        -3.51        10.8
4         0.03        -4.2
4         0.57         1.8
4         0.16         4.4
4        -3.08         3.6
4         0.20         4.8
4        -3.20         3.8
4        -1.86         3.6
4        -1.09        -4.6
4         0.69        -3.4
5         0.98         4.2
5         1.69         3.2
5        -2.90       -14.6
5        -0.04        -2.2
5         0.09        -4.6
5        -0.96         4
5         1.26         6
5         0.05         4
5         0.22        -5.2
5        -1.27         3.8
5         0.13         5
5        -0.59        -4.8
5         1.24        -3.8
5        -3.36         0.6
5        -1.78         3.4
6         0.48        -4.2
6         0.85        -4.2
6         0.10        -2.2
6         2.83         3.8
6        -0.78        -3.8
6         0.14        -3.6
6        -2.22        -3.4
7        -2.69         3.2
7        -2.08         4.6
7         0.08        -6.2
7        -1.32         1.8
7        -1.89        -3.6
7         0.88        -4
7        -0.04       -21.2
7        -0.36        -9.8
7         0.85        -3.6
8        -0.58        -4
8         1.06        -1
8         0.49        -3.4
8        -0.48         4.2
8         0.83         3
8        -3.50         1.6
8         0.11         4
8         0.47         3.8
9         0.07         4.4
9         0.21         5.2
9        -2.71         2.6
9         0.22        -3.8
9         0.61         3.4
9        -1.90         3
9         0.26        -4.4
9         0.01         3.6
9         0.44         3.8
10       -2.07         4.2
10        0.27         3.8
10       -1.11         4.6
10       -0.65        -4
10       -1.29        -2.8

Table 8. Starch QTL found to be significant in per se and/or
testcross lines by simple interval mapping. "Chrom" is the
chromosome number, "Marker Number" indicates the marker
nearest the QTL, and "Marker Position" gives a marker's
genetic map position in centimorgans. "QTL Distance" is the
QTL distance relative to the nearest marker in cM. A negative
value indicates the QTL is toward the beginning of the chro-
mosome. "LOD +" refers to the amount an observed LOD
score exceeds the permutation generated LOD threshold value.
"a" and "d" are the estimated additive and dominance ef-
fects, respectively.

         Marker     Marker       QTL                 Per se
Chrom    number    position    distance
                                           LOD +       a            d

                                                    g [kg.sup.-1]

1           1           0         4.2      -1.70        19.2       15.7
1           9          85         1.0      -2.22         9.4      -14.2
1          15         119        -0.4       0.65       -13.9        9
1          20         132        -1.0      -0.70       -22.6        9.2
1          32         193        -1.8       1.22        13.4       -0.9
1          34         203        -2.6      -1.69       -13.6        5.8
1          34         203        -0.5       2.19        12.3       -5.1
1          40         208         1.6      -4.00        13.6       -3.2
1          52         260        -0.5       0.10       -15.9       -1.9
1          53         279        -2.5      -1.83        27.8       -0.4
1          59         303        -2.5      -0.03        19.4        7.3
1          63         328         0.9       3.18       -11.9       -7.8
1          68         344        -2.7       2.67       -12.1        6.4
1          71         382       -15.5      -4.82       -61.3       26.1
1          71         382        -9.1       0.11       -16.3       16.6
1          72         393        -1.2       0.20        12.7        8.6
1          74         410         3.7      -1.45       -13.6        2.4
1          75         427        -6.0       0.06       -18.6        4.5
1          76         431        -2.0      -1.04       -15.4       -7.4
2          83           0         1.0      -3.12       -11.8       -9.1
2          85          20        -6.9      -0.01       -76.2       14.2
2          86          20         0.6       3.99       -11.5       12.7
2          92          46        -2.1      -0.06       -29.1      -51.8
2          97          69         0.2       2.90        13.9        1.6
2          99          77        -0.3      -0.95        19.1        3.1
2         106         106        -2.6      -2.92         6.9       19.5
2         110         146         0.6       1.51        16.4       28.6
2         115         180        -0.7       1.10       -12.7       -2.2
2         119         188         0.6       0.83       -12.3       -6.1
2         120         198        -0.4       0.47       -10.2      -13.2
2         124         230        -1.9       0.58        11.5      -10.6
3         133          20        -0.2      -2.77       -16.6      -11.6
3         134          56         1.0      -3.37       -13.1        6.9
3         136          69         1.8       3.08       -12         -7.5
3         138          74         3.0       3.16       -11.7       -7.6
3         139          95        -1.2      -1.57       -15.7       13.7
3         143         142        -0.6       4.91       -12.3        0.1
3         148         143         7.2       0.72       -10.2      -38.9
3         151         188         0.7      -4.02        13.3      -53.3
3         154         203        -0.1       1.41       -11.1       -8.5
3         157         212        -0.1       1.21       -12.5        2.7
3         171         235        -2.9      -1.48        12.0        9.8
3         178         249        12.2       0.90        -2        -25.1
3         185         299         0.4       0.43        -7.7      -16.5
4         198          71         0.1      -2.59        36.4       17.4
4         199          74         3.5       1.59        -6.6      -20.8
4         200          95        -3.5       1.14        -9.4        -14
4         208         123        -0.9      -3.80       -12.4      -29.3
4         208         123         2.2       0.29       -18.1      -26.6
4         218         182        -2.0      -1.40       -15.5       -0.8
4         223         195         1.1       0.58       -14.2      -17.4
4         232         233         0.9      -1.32        -5.7       26.1
4         239         275         4.4       0.00       -15.4      -16.3
4         242         292        -0.3       0.16        -3.3      -48
4         242         292         0.7      -1.53        -4.6      -22.8
5         251          18         0.9       0.01       -13.1        7.8
5         254          52         1.4       1.03       -12.3       -5.8
5         255          63        -1.2       0.31       -11.4       -9.5
5         264         131         3.2      -0.68       -11.5       -2.6
5         265         140        -4.1       0.79       -13.7       -0.1
5         267         147         3.0       0.60       -17.5       -0.7
5         269         157         0.9      -0.11         0.3       -8.2
5         271         209         2.4       1.90        17.0       11.4
5         273         223        -1.3      -2.85        11.8      -46.6
5         282         356         0.4       1.19       -14.7       27.5
5         287         441        -3.2      -0.03       -52.4       48.6
5         288         465         0.7       4.39       -10.4       14.0
5         289         478        -4.9       1.84        -9.9       13.6
6         302          44        -1.3       0.42       -15.3       13.2
6         302          44        -0.3      -2.31       -12.5       16.4
6         311          69        -1.8      -3.17       -15.2        2.0
6         318          86        -0.5       0.38       -17.9       -9.1
6         318          86         0.5      -1.00       -12.1       -3.6
6         320          93        -0.1       1.29       -11.7       13.8
6         325         108        -0.2      -2.62        -7.4      -18.9
6         326         108         0.8       1.51       -23.6       13.6
6         332         136         1.0       0.10        13         18.1
6         334         163        -3.2       1.58       -12.6        5.1
7         342          20         0.0       0.60       -13.8        4.0
7         348          46        -2.9       1.66        13.2       -1.4
7         354          82        -3.1       0.39        -7.4       21.7
7         369         170         3.2       0.58        19.1        9.9
7         370         205        -1.1       0.47       -19         22.2
7         373         220        -4.2      -0.03       -61.1       35
8         387          55        -7.2       0.09       -15.2       -7.8
8         391         122         0.9       3.12       -11.4       -7
8         392         128        -2.0      -2.04       -13.7       -9.9
8         394         136         2.4       2.07       -11.4      -10.3
8         395         148        -0.6       0.42       -11.3       10.3
8         401         219         4.1       2.05       -13.4       -2.4
9         417           0         4.2      -1.13        11.7        6.9
9         419          27        -4.1      -0.59       -17.2      -11.5
9         420          50         6.3       0.71        14.3       -2
9         423          81        -5.0      -4.34        -8.1      -28.3
9         432         102         4.7      -3.39       -15.7       -5.7
9         441         139         0.0      -0.61       -14.8        2.4
10        452           9        -0.4      -1.62       -11.6      -13.6
10        453          37        -1.2      -2.67       -15.2       -6.1
10        468          62         0.3      -2.41        14.3        5.0
10        471          84         1.3       0.29       -19.5       16.0
10        472          95        -2.7       0.37        14.7       14.6
10        478          99         0.1      -1.02         8.3      -33.8

               Testcross
Chrom
         LOD +          a

                   g [kg.sup.-1]

1         0.91          5.4
1         0.06         -0.8
1        -0.29         -5.2
1         1.91         -5.6
1        -1.24          5.8
1         1.47         -4.8
1        -4.20         20.6
1         0.16         -6
1        -1.69         -4.2
1         1.14         -6
1         2.77         -9.2
1        -0.19         -7
1        -3.66         -6.6
1         0.34         -6.2
1        -0.44          6.8
1        -2.00          4.4
1         0.20         -5.2
1        -5.51        -10.2
1         0.91         -5.2
2         1.36         -5.4
2         1.27         -5.8
2         0.39         -5.2
2         1.55          5.6
2        -3.55          5.0
2         0.03          5.6
2         0.92         -5.2
2        -1.37         -3.4
2         1.60         -5.4
2        -1.08         -5.2
2        -1.89         -5
2        -0.44         -4.6
3         0.43         -2
3         1.42         -5
3         4.29         -5
3         4.79         -5.2
3         0.49         -4.6
3         6.57         -5
3        -0.05        -25.4
3         1.71         -5.8
3         2.51         -6
3         4.27         -5.2
3         0.30         -6.4
3        -0.01        -20
3        -0.21         -6.2
4         0.12         -0.6
4        -1.13          0.0
4        -1.90         -5.4
4         0.20         -0.8
4        -1.29          2.0
4         0.06         -8.2
4        -1.80         -5.6
4         1.25          9.8
4        -3.06         -4.4
4        -1.34         -5.6
4         0.28         -8
5        -2.55         -5.6
5        -1.39         -6.6
5        -2.50         -4.8
5         4.05         -5.4
5         0.31         -5.4
5        -3.87        -24.2
5         1.06         -9
5        -4.71          7.6
5         1.47         -6.8
5        -0.02          1.2
5         0.67          9.0
5        -1.88         -5.8
5        -0.03          0.0
6         0.91        -14.2
6         1.06        -13
6         5.13         -5.4
6        -1.49          6.6
6         0.40         -7.4
6         1.29         -5.4
6         0.87         -4.2
6        -0.11         17.6
6        -3.19          5.6
6        -5.26          0.6
7        -2.03          8.0
7        -1.78         -5
7        -4.99        -15
7        -0.84          4.8
7        -1.59        -10.8
7         0.76         13.6
8        -0.08        -19
8        -0.49         -6
8         1.37         -5.4
8        -2.95         -4.4
8        -0.07          2.8
8        -3.92         -3.4
9         0.60        -12
9         0.61         -9.4
9        -1.94         -5.4
9         0.43         -6.2
9         0.06         -7.2
9         0.60         -6.2
10        0.45         -7.4
10        0.47         -2.6
10        1.82          5.0
10        0.46          9.4
10       -0.19          5.6
10        0.34          5.0

Table 9. Number of regions identified by simple interval mapping
(SIM) for per se and testcross progenies.

Progeny      Oil    Protein    Starch

                  g [kg.sup.-1]

Per se       51       47         42
Testcross    54       51         42
Both         35       22         19

Table 10. Number of regions identified by simple interval map-
ping (SIM) and single marker (SMA) analyses for per se and
testcross progenies.

Population and method    Oil    Protein    Starch

                              g [kg.sup.-1]

Per se
  SIM                    51       47         42
  SMA                    33       24         29
  Both                   21       21         23
Testcross
  SIM                    54       51         42
  SMA                    29       26         29
  Both                   23       19         23

Table 11. Phenotypic (above diagonal) and genotypic (below
diagonal) correlations ([+ or -] one standard error) among kernel
quality traits in ner se (PS) and testcross (TC) populations.

Population    Trait            Oil                 Protein

                                    g [kg.sup.-1]

PS           Oil                              0.23 [+ or -] 0.05
TC                                            0.18 [+ or -] 0.05
PS           Protein    0.26 [+ or -] 0.04
TC                      0.22 [+ or -] 0.04
PS           Starch    -0.58 [+ or -] 0.03   -0.71 [+ or -] 0.05
TV                     -0.82 [+ or -] 0.02   -0.72 [+ or -] 0.04

Population          Starch

                g [kg.sup.-1]

PS           -0.55 [+ or -] 0.03
TC           -0.75 [+ or -] 0.02
PS           -0.65 [+ or -] 0.03
TC           -0.73 [+ or -] 0.02
PS
TV

Table 12. Number of regions identified by SIM, MAXR/BIC, and
SMA for oil in testcross and per se analyses.

Type of analysis    Per se results    Testcross results

SIM ([dagger])            51                 54
MAXR/BIC                  36                 35
SMA                       33                 29
SIM and MAXR/BIC          22                 26
SIM and SMA               21                 23

([dagger]) SIM is simple interval analysis, MAXR/BIC is maximum
regression analysis, and SMA is single marker analysis.
COPYRIGHT 2006 Crop Science Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Clark, Darryl; Dudley, John W.; Rocheford, Torbert R.; LeDeaux, John R.
Publication:Crop Science
Date:Mar 1, 2006
Words:14895
Previous Article:Aegilops species from semiarid areas of Lebanon: variation in quantitative attributes under water stress.
Next Article:Bayesian modeling of heterogeneous error and genotype x environment interaction variances.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters