Development of maize inbred lines with elevated grain methionine concentration from a high methionine population

Abstract

Methionine is a nutritionally limiting amino acid in poultry diets based on maize (Zea mays L.) grain. Synthetic dietary supplements are available but are costly and not preferred by organic poultry producers. The development of high methionine maize varieties would reduce the need for supplementation. Several approaches have been reported for achieving this goal. Here, we report a novel approach that can produce diverse inbred lines with higher content of methionine than other methods. Inbred lines were developed using doubled haploid technology from a broad-based synthetic population that has undergone mass selection for grain methionine concentration. Out of 18 randomly selected inbred lines, one was significantly higher in methionine concentration than the high methionine check and 11 were not significantly different from it. The inbred lines developed in this way also exhibited useful genetic diversity for several agronomic and kernel quality traits, including flowering date, and orangeness of the kernel. This approach is an excellent complement to other breeding methods for development of varieties for production of poultry feed. Because the approach does not rely on transgenic technology, the resulting lines are suitable for use by organic producers and are well suited to organic production systems.

Abbreviations

DH

doubled haploid

DO

distance from orange

GDD

growing degree days

LSD

least significant difference

OD

optical density

RGB

red green blue

1 INTRODUCTION

Methionine is an essential amino acid that is limiting in corn-based poultry diets. This limitation is normally corrected by the addition of supplements such as soy protein or synthetic methionine. These supplements add to the cost of the diet. In addition, the National Organic Regulations limit the use of chemical supplements to 2 pounds per ton of feed for laying and broiler chickens and to 3 pounds per ton of feed for turkeys and other poultry (USDA, 2017). Corn varieties with elevated levels of methionine would alleviate the need for supplementation, allowing poultry producers to design diets that are less expensive and do not rely on synthetic methionine.

Researchers have used several approaches to increase corn grain methionine concentration. A natural mutation called dzr1 was identified in an inbred line called BSSS53 (Phillips et al., 1981) that was later released as B101 (Hallauer & Wright, 1995). This line was used as a donor line in a backcrossing program to develop several high methionine inbred lines (Olsen et al., 2003; Phillips et al., 2008). The dzr1 gene has been shown to regulate levels of the high methionine 10 kDa delta zein transcript (Cruz-Alvarez et al., 1991). Breeding with this gene is complicated because the genetic locus has not been characterized at the molecular level. Most zein genes occur in clusters of tandemly repeated copies derived from a founder sequence. This leads to high expression levels of the family and divergence among family members. Dzr1 has been mapped to a region of the genome that contains a gene cluster of alpha-zein genes (Benner et al., 1989; Chaudhuri & Messing, 1995). The repetitive nature of this region complicates development of molecular markers linked to the high methionine trait. A further complication is that the inheritance of the dzr1 locus is subject to imprinting, making the outcome of crosses less predictable than Mendelian traits (Chaudhuri & Messing, 1994; Wu et al., 2009).

A transgenic approach based on an understanding of the molecular basis of dzr1 function allowed researchers to design a transgene that increases grain methionine levels by manipulation of the methionine sink strength (Lai & Messing, 2002). More recently, researchers used a second transgenic approach that involves manipulation of the sulfur assimilation pathway to increase grain methionine concentration to greater levels. This approach increases the strength of the methionine source (Planta et al., 2017).

Researchers have also used recurrent selection to develop broad-based breeding populations with elevated grain methionine levels (Scott et al., 2008). After eight cycles of recurrent selection, methionine levels in populations selected for high methionine averaged 0.26 g/100 g of tissue compared to a value of 0.20 g/100 g of tissue for populations that were selected for low methionine levels. These populations have elevated methionine levels resulting from changes in both source and sink strengths (Newell et al., 2014). Following eight cycles of selection for only grain methionine concentration, the agronomic performance of the populations was poor. To restore agronomic performance, the two cycle 8 high methionine populations described in Newell et al. 2014 were combined and two additional cycles of selection for agronomic traits as well as grain methionine concentration were carried out. This process included selection for ability to germinate following exposure to cold temperature, disease resistance, and other traits of interest to organic corn producers. Methionine concentrations measured during selection suggest that this population is near the level required for poultry diets, and that it contains individuals well in excess of this level. Because grain methionine concentration was selected as a quantitative trait rather than a single gene trait, it may be possible to increase methionine concentration to extreme levels, analogous to what has been done with grain protein and oil concentration in the Illinois long-term selection program (reviewed in Dudley & Lambert, 2004). In addition, because mutation breeding and recurrent selection likely rely on different genetic mechanisms for increasing grain methionine concentration, it may be possible to combine these methods to produce varieties that are superior to varieties produced by either method alone.

While recurrent selection population improvement programs have created populations containing individuals with extreme phenotypes in many traits including grain protein and oil concentration (Dudley & Lambert, 2004), ear length (de Jesus Lopez Reynoso & Hallauer, 1998), kernel mass (Odhiambo & Compton, 1987), and endosperm appearance in sweet corn (Tracy & Chang, 2007), the individuals in these populations cannot be reproduced. One limitation of working with intermated populations is the difficulty of capturing extreme phenotypes in reproducible lines. Traditionally this requires many generations of self-pollination to produce inbred lines. With the ready availability of doubled haploid (DH) technology in maize, creation of inbred lines from breeding populations requires much less time and expense than the multiple cycles of inbreeding that were required in the past. The objective of this study is to demonstrate the use of DH technology to capture extreme phenotypes in DH lines derived from a broad-based synthetic breeding population that has undergone recurrent selection for grain methionine concentration, thereby immortalizing the extreme phenotypes from the population in reproducible lines. The ability to rapidly develop reproducible lines carrying extreme phenotypes from recurrent selection population improvement programs adds value to these populations for plant breeders. In addition, the inbred lines developed in this study are an important step toward developing hybrids that meet the needs of organic poultry producers for high methionine grain.

Core Ideas

• Methionine is a limiting nutrient in maize-based poultry diets and few options exist for organic feed.
• The high-grain methionine trait from a selected population was transferred to doubled haploid lines.
• The lines produced in this study have higher methionine than checks and diversity in other traits.
• This approach addresses the need of organic poultry producers for diverse high methionine maize germplasm.
• Breeding populations are more valuable when phenotypic extremes can be readily captured in DH lines.

2 MATERIALS AND METHODS

2.1 Germplasm used in the study

The experimental entries in this study consisted of 18 maize inbred lines produced from a broad-based synthetic population of maize that was selected for high methionine concentration for 10 generations. This population was produced by combining two high methionine populations (BS11HMC8 and BS31HMC8; Newell et al., 2014), followed by random mating of the resulting population with selection for high methionine individuals for two more generations. This population was used to produce DH lines using a protocol developed by the Iowa State University Doubled Haploid Facility (Vanous et al., 2017). In addition, a normal methionine check inbred line (B73) and a high methionine check inbred line B101 (Hallauer & Wright, 1995) were included. B101 is also known as BSSS53 (Phillips et al., 1981). These lines were selected for their methionine levels and are not used in current commercial hybrids.

2.2 Experimental design for inbred line evaluation

The entries were evaluated in 2 years during the summers of 2017 and 2018 at the Iowa State University, Agronomy and Agricultural Engineering research farm near Boone, Iowa. The soils at the ISU Agronomy and Agricultural Engineering farm are classified as Clarion, Nicollet, and Webster with slopes from 0%–5%. Tile drainage is in place on much of the farm. The weather in the two growing seasons differed during both rainfall and temperature, with 2018 receiving about 20 cm more rainfall over the season, with most of the excess coming in June and July (Figure 1).

FIGURE 1

Hourly temperature (above) and cumulative rainfall (below) for the 2017 and 2018 growing seasons.

Plots were managed using practices optimized for commercial production of corn in this area. Each plot consisted of one row of 5.9 m in length, with 0.76 m between adjacent plots. The 20 entries were randomly assigned to plots in three complete blocks (replications) in each of the 2 years of the study. The 60 plots in each season were arranged in a 10 × 6 grid, with each complete block consisting of a 2 × 10 portion of the grid. This field design was selected because it can be replicated on a standard 96 (12 × 8) well micro titer plate without requiring the use of edge wells. At least three internal plants in each plot were self-pollinated by hand. Grain from one randomly selected, fully developed ear from each plot was ground for analysis.

2.3 Trait measurements on inbred lines

Grain methionine concentration was determined using a microbial method that has been shown to correlate well with the AOAC standard method of analysis (Scott et al., 2008). This method can be carried out in 96-well plates and costs about 1% of the AOAC standard method, allowing greater replication and higher precision than the standard method. Each of the 120 grain samples was assigned to a well of a 96-well plate in the same randomized complete block design that was used in the field, with the 60 samples from 2017 analyzed on one plate and the 60 samples from 2018 analyzed on another. Preserving the experimental design across the field and laboratory analysis confounds the sources of error introduced by laboratory and field operations, resulting in fewer sources of variation and a better estimation of error than would be possible if field samples were re-randomized prior to laboratory analysis. We submitted one replicate of each genotype to the University of Missouri Experiment Station Chemical Laboratory for analysis by the AOAC standard method (AOAC, 2006) as a control.

In addition to grain methionine concentration, several traits related to agronomic performance and plant morphology were measured on a plot basis by averaging the values of five randomly selected plants in each plot. These traits included plant height (measured from the ground to the highest part of the plant) and ear height (measured from the ground to the node the ear is attached to), tassel branch number, and the number of growing degree days (GDD) from planting to the day the plants were pollinated. Average kernel mass was determined from the ears that were harvested for methionine analysis by weighing 10 randomly selected kernels from each ear.

In order to quantify the visual degree of orangeness of the kernels, a subsample of about 20 kernels from each of the 120 grain samples produced in the multienvironment trial was photographed in natural daylight with a visible light digital camera. Each image contained a color standard. Color measurement was based on the red green blue (RGB) color system. In this system, three color parameters represent the relative contributions of red, green, and blue, respectively. RGB values with a scale of 0–255 were extracted from each image using ImageJ software (version 5.2a, National Institutes of Health). In order to create a quantitative measure of orangeness of each sample we defined a distance from orange (DO) value by comparing the observed RGB value to the standard RGB value of orange color (RGB = 255, 140, 0). The DO value was calculated with the following equation:

$$$$\begin{equation*}{\text{DO value}} = \sqrt {{{({R}_{Obs} - {R}_{Std})}}^2 + {{({G}_{Obs} - {G}_{Std})}}^2 + {{({B}_{Obs} - {B}_{Std})}}^2} \end{equation*}$$$$

where R_Obs, G_Obs, and B_Obs are RGB values of the observed value and R_Std, G_Std, and B_Std are the standard orange RGB values (R = 255, G = 140, and B = 0). The resulting DO values were used as the response variables in the statistical analysis described below.

2.4 Evaluation of agronomic performance of hybrids made with high methionine inbred lines as parents

In order to evaluate the agronomic performance of the high methionine inbred lines in hybrid combinations, lines with high grain methionine concentrations were crossed to produce seed of six hybrids. This seed was planted in a multilocation yield trial in nine different environments in 2021 and 2022. A current commercial hybrid with normal methionine concentration and recommended for production of organic grain in the growing area was included as a check entry in the experiment. One of the experimental hybrids exhibited poor stands and was removed from the experiment. The experiment was a randomized complete block design with two replications at each location. Each plot consisted to two rows that were 17.5 m long, with 0.76 m between rows. Plots were harvested after physiological maturity using a plot harvesting combine equipped with a weighing bucket and a moisture meter. Plot weight and moisture were converted to grain yield in Mg/Ha at 15.5% moisture.

2.5 Statistical analysis

For each of the response variables in the study, linear models were constructed to partition the variance in the 120 measurements (Y) among different effects as follows:

$$$$\begin{equation*}{Y}_{ijkl} = m + {G}_i + {E}_j + G{E}_{ij} + E{[{\rm{Row}}]}_{jk} + E{[{\rm{column}}]}_{jl} + {e}_{ijkl}\end{equation*}$$$$

where Y is a measurement; m is the mean value of all measurements in the experiment; G_i is the effect of the ith genotype (the range of i is 1–20); E_j is the effect of the jth environment (the range of j is 1–2); E[Row]_jk is the effect of the kth row in the experimental design, with k being the same in the field and laboratory (k ranges from 1 to 6), nested within the jth environment; E[Column]_jl is the effect of the lth row in the experimental design, with l being the same in the field and laboratory (l ranges from 1 to 10), nested within the jth environment; and the E[Row]_jk and E[Column]_jl effects were not significant for all traits except GDD to pollination date so all models were re-run without these terms.

The statistical significance of each effect in the model is reported in Table 1. For traits that were not analyzed in the laboratory, the above model was used, but row and column effects were only due to field effects. Models were fit to the data using the standard least squares method, with all effects treated as fixed. Studentized residuals were plotted in normal quantile plots (Figure S1) to examine the uniformity of the error and to detect outliers. No outliers were identified. Genotypic least squares means for methionine concentration and resulting significance groupings are based on the data from the microbial method and are reported in Table 2. For all other traits genotypic mean values were reported with a least significant difference (LSD) value in Table 3.

TABLE 1. Degrees of freedom (DF), mean squares, significance probabilities for linear models of 20 lines for methionine concentration, plant height (PH) and ear height (EH), tassel branch number (TB), kernel mass (KB), orangeness (distance from orange, a measure of color), and growing degree days (GDD) to pollination date (PD).

	DF	Met (O.D.)	PH (cm)	EH (cm)	TB	KM (g)	Color	PD (GDD)
Genotype	19	0.0052^**	2078^**	954.7^**	123.6^**	0.6368^**	2269^**	1.274^**
Environment	1	0.0414^**	8546^**	194^**	100^**	1.091^**	1971^**	509.1^**
GxE	19	0.0016^**	101.6^n.s.	74.10^*	6.106^n.s.	0.1617^**	372.2^**	0.0663^n.s.
Error	80
Total	119
Model R^2		0.76	0.85	0.84	0.85	0.80	0.81	0.93

• Abbreviation: O.D., optical density.
• *, **Significant at respectively. n.s., not significant.

TABLE 2. Methionine concentration of 18 doubled haploid lines and two checks with significance groupings.

Genotype	Significance grouping^a	Methionine conentration (microbial) (g/100 g tissue)	Methionine concentration (AOAC) (g/100 g tissue)
PS1-9	A	0.321	0.35
PS1-14	AB	0.309	0.35
PS1-24	ABC	0.310	0.35
PS1-6	ABCD	0.296	0.37
B101	BCDE	0.287	0.29
PS1-15	CDEF	0.285	0.29
PS1-17	CDEF	0.286	0.28
PS1-22	DEFG	0.283	0.29
PS1-10	DEFGH	0.267	0.32
PS1-23	EFGH	0.259	0.28
PS1-8	EFGHI	0.263	0.27
PS1-25	EFGHIJ	0.264	0.21
PS1-3	FGHIJ	0.246	0.26
PS1-4	GHIJ	0.230	0.25
B73	HIJK	0.246	0.25
PS1-2	HIJK	0.246	0.26
PS1-7	IJK	0.253	0.26
PS1-13	IJK	0.247	0.27
PS1-26	JK	0.231	0.27
PS1-11	K	0.209	0.20

• ^a Two samples are significantly different from each other (α = 0.05) if no letters in their significance grouping overlap. Significance groupings are based on the microbial method. The AOAC method was carried out on one replicate of each entry as a control.

TABLE 3. Genotypic mean values of agronomic and grain quality traits of 18 inbred lines and two check varieties.

	EH (cm)	PH (cm)	TB	PD	KM (g)	Color
B73	93.6	182	8.7	1607	2.65	69.8
B101	93.5	156	28.0	1790	2.59	72.9
PS1-2	71.6	148	15.2	1560	2.55	96.0
PS1-3	101	190	12.8	1733	2.61	77.4
PS1-4	74.9	161	13.2	1557	2.12	88.4
PS1-6	58.5	155	14.8	1559	2.67	91.2
PS1-7	69.6	141	11.7	1601	2.81	83.0
PS1-8	66.2	140	18.6	1615	2.63	78.7
PS1-9	61.5	162	12.9	1562	2.36	102
PS1-10	68.8	155	9.44	1635	1.80	78.0
PS1-11	82.4	193	20.8	1602	2.25	93.1
PS1-13	95.1	200	13.3	1488	2.94	74.0
PS1-14	90.0	168	12.7	1622	2.84	67.9
PS1-15	87.1	186	13.6	1610	2.26	66.0
PS1-17	77.9	165	13.3	1539	2.15	78.2
PS1-22	69.9	150	15.1	1547	2.01	62.7
PS1-23	75.2	130	15.5	1692	2.03	84.6
PS1-24	99.3	169	13.6	1663	2.89	73.2
PS1-25	80.3	163	22.9	1494	2.23	94.4
PS1-26	75.7	163	13.1	1524	2.58	83.4
Mean	79.6	164	15.0	1600	2.45	80.8
LSD	8.36	15.9	2.99	108	0.335	9.99

• Abbreviations: Color, grain color measured as distance from orange; EH, ear height; KM, kernel mass; LSD, least significant difference; PD, growing degree units to pollination date; PH, plant height, TB, tassel branch number.

Hybrid yield data were evaluated using an approach similar to the inbred line experiment, except that row and column effects were not included in the model.

Pearson product-moment correlations between trait values were calculated using the pairwise method.

3 RESULTS AND DISCUSSION

3.1 Analysis of variance

Genotype was a significant source of variation for all the traits measured (Table 1). Since all DH lines were derived from the same population, this is likely a reflection of the diversity contained in the population. Some diversity is contributed by the two check samples as well. The effect of environment was significant for all traits. The difference in mean grain methionine levels between the 2 years of the study was 0.037, which is a relatively large difference. For comparison, the mean difference between the high methionine check and the normal methionine check was 0.042. The GxE effect was only significant for grain methionine concentration and kernel mass. Due to these significant effects, it will be important to use varieties with the genetic potential to exceed the desired grain methionine level, since the genetic potential will likely be modulated by the production environment.

3.2 Grain methionine levels

Seven of the lines in this study had grain methionine levels that were significantly higher than that of the normal methionine check, including B101 which was included as the high methionine check (Table 2). Ten of the experimental lines were not significantly different from B101 and one line was significantly higher than B101.

Several approaches have been used to increase the concentration of grain methionine. We have demonstrated that it is possible to produce inbred lines with higher methionine concentration by this method than by other methods. However, the different methods of increasing methionine concentration have different advantages and disadvantages and having multiple approaches to address the problem is desirable. For example, mutation breeding and recurrent selection are both acceptable in organic production systems, while transgenic technology is not. Population improvement by recurrent selection is time consuming, but once a population with desirable traits is established, deriving inbred lines with DH technology can be done more rapidly than can be done by either transgenic approaches or by backcross breeding with the dzr1 mutation. It is striking that more than half of the randomly selected DH lines were not significantly different from the high methionine check inbred. This observation demonstrates the effectiveness of using recurrent selection to create a population with the desired traits for use as a donor population in DH line development.

3.3 Agronomic and grain traits

Examination of morphological traits of the plant revealed significant genetic diversity among the inbred lines in the study for all traits examined. Environmental variation was small. Kernel mass was the only trait with significant GxE. The significant genotypic effects observed for all agronomic traits measured in this study demonstrates the degree of diversity possible in inbred lines derived from the broad-based high methionine population used in this study. Genotypic mean values of agronomic and grain traits are presented in Table 3. This diversity will be useful in the development of varieties that are suited to specific locations and management practices. Lines with different GDD to pollination are valuable to breeders wishing to target products to different maturity zones, for example, while plant height may play a role in weed suppression, which is a trait of great importance to organic producers.

As an initial attempt to understand the combining ability of the inbreds included in this study, five hybrids were made and evaluated in a multienvironment yield trial (Table 4). The inbreds used for this study were chosen for their grain methionine concentration and other factors. This multienvironment trial showed significant genotype and environment effects for yield and moisture and significant GxE effects for moisture (Table 5). The experimental hybrids yielded between 8 and 10 Mg/Ha while the commercial check was 12.7 Mg/Ha. The commercial check had significantly better moisture at harvest than the experimental hybrids as well (Table 6). It would be possible to make 99 hybrids among the 11 inbreds that were not significantly different or higher than the high methionine check in grain methionine concentration, so the variation in yield and moisture among the five hybrids evaluated in this initial study is encouraging.

TABLE 4. Environments used in the hybrid yield trial.

Location	Year	Planting date	Longitude	Latitude
BRKAb	2022	May 24	−93.78603103	42.01451761
CRL06	2022	May 20	−94.72547115	42.06693165
CRL08	2021	April 27	−94.72547115	42.06693165
CRW14	2022	May 10	−91.48564539	41.19773777
KYS01	2021	May 6	−92.24518576	41.97497079
KYS02	2022	May 11	−92.24480778	41.98026185
MRSD	2021	May 14	−93.7823675	42.01464184
MRSE	2022	June 6	−93.77941463	42.01517119
WRL02	2021	April 29	−93.68885953	41.99746651

TABLE 5. Degrees of freedom (DF), mean squares, significance probabilities for effects in linear models of five hybrids and checks evaluated for grain yield and moisture concentration at harvest.

	DF	Yield (Mg/ha)	Moisture (%)
Genotype	5	66**	31**
Environment	8	12**	41**
GxE	40	2.0^n.s.	0.7**
Error	105	2.1	0.5
Total	158
Model R^2		0.80	0.93

• Abbreviation: n.s., not significant.
• **, significant at α = 0.05.

TABLE 6. Comparison of mean yields and moisture levels of five experimental hybrids and commercial checks at nine environments.

Genotype	Yield (Mg/ha)	Grouping^a	Moisture (%)	Grouping
Check	12.7	A	16.4	D
PS1-24/PS1-10	10.0	B	18.9	B
PS1-14/PS1-10	10.0	B	18.9	B
PS1-10/PS1-24	9.2	BC	18.9	B
S1-9/PS1-24	8.9	CD	18.2	C
PS1-9/PS1-10	8.0	D	19.4	A

• Note: Genotypes with the same grouping letter are not significantly different.
• ^a Two samples are significantly different from each other (α = 0.05) if no letters in their significance grouping overlap.

The inbreds examined in this study varied significantly in the degree of orange of their grain, even though it was not a target of selection (Table 1; Figure 2). Visually orange grain is desirable in poultry feed because it can impart a desirable color to meat and egg products of the animals that consume it. Orange grain color can be produced by elevated levels of certain compounds in the xanthophyll biosynthetic pathway and these compounds confer orange color to the yolks of eggs of hens consuming an orange corn diet (Heying et al., 2014; Ortiz et al., 2021). Yolk color was determined to be the most important attribute considered by consumers when selecting eggs from different production systems (Berkhoff et al., 2020). The xanthophyll pathway produces several carotenoid compounds with provitamin A activity, creating a correlation between orange grain color and provitamin A levels in the grain, and it has been established that chickens that were fed high carotenoid corn have increased levels of carotenoid compounds (Burt et al., 2013) The genetic control of degree of orange color has been examined (Chandler et al., 2013) and half of the quantitative trait loci identified were located at chromosome positions known to contain a gene in the xanthophyll pathway. It has been shown that visual selection for orange kernel color can be an inexpensive and effective method in improving the carotenoid concentration in grain (Burt et al., 2011). The variation in color of the lines examined here may make them useful for addressing questions about correlation of grain color with provitamin A levels or the ability to confer desirable colors through feed to meat or eggs.

FIGURE 2

Images of one rep of the grain samples used in this study, with the orange standard. The number in the upper left corner is the distance from orange value for the genotype shown, averaged across all reps of the study.

3.4 Correlations among traits

Table 7 shows correlations between all measured traits. While several of the correlations were statistically significant, only one was greater than 0.5 (plant height and ear height), which indicates that the variation among the traits examined is relatively independent. There are significant correlations between ear height and plant height, methionine concentration and GDD, ear height and GDD, plant height and GDD, ear height and kernel mass, ear height and DO, plant height and DO, and GDD and DO. The correlations of ear height and plant height, methionine concentration and GDD, ear height and GDD, ear height and DO, plant height and DO, and GDD and DO, were all significant at a level of α = 0.001. The correlations between DO and other traits were negative, and all other significant correlations were positive. The relatively low correlations between methionine concentration and the other agronomic traits are important because it suggests that there is not a strong biological connection between methionine concentration and the agronomic traits examined in this study. It may, therefore, be possible to improve agronomic and grain quality traits without sacrificing methionine concentration and vice versa.

TABLE 7. Correlation coefficients between measured traits. pairwise correlations (r-value) between the traits of methionine concentration, ear height (EH), plant height (PH), tassel branch number (TB), growing degree days to pollination date (PD), kernel mass (KM), and distance from orange are reported.

	MC	EH	PH	TB	PD	KM
EH	0.011
PH	0.030	0.600***
TB	−0.086	0.041	−0.157
PD	0.435***	0.313***	0.197*	0.014
KM	−0.151	0.217*	0.079	0.033	−0.178
Color	−0.113	−0.389***	−0.260***	0.191*	−0.277***	−0.105

• Abbreviations: Color, grain color measured as distance from orange; MC, methionine concentration (microbial method).
• *, ***Statistical significance at α = 0.05 and 0.001, respectively.

4 CONCLUSIONS

This study demonstrates the effectiveness of making DHs from a selected random mated population. Over half of the inbred lines produced in this way were not significantly different in grain methionine concentration from the high methionine check inbred, and one was significantly higher. In addition to having elevated levels of grain methionine, these inbred lines contain useful variation for agronomic and grain quality traits that will allow development of varieties that are well suited to use in poultry feed rations. This will benefit poultry producers by lowering the cost of their diets and provide organic poultry producers with a much-needed natural source of methionine. Using DH technology removes a significant barrier to the capture of extreme phenotypes in breeding populations developed by recurrent selection and may increase the value of these populations in breeding programs.

AUTHOR CONTRIBUTIONS

Taylor D. Hintch: Formal analysis; investigation; writing–review and editing. Adrienne Moran Lauter: Formal analysis; investigation; methodology; writing–review and editing. Shelly M. Kinney: Formal analysis; writing–review and editing. Thomas Lubberstedt: Methodology; project administration; writing–review and editing. Ursula Frei and Jode W. Edwards: Methodology; writing–review and editing. Prakasit Duangpapeng: Formal analysis; investigation; methodology; writing–original draft; writing–review and editing. Marvin Paul Scott: Conceptualization; data curation; formal analysis; funding acquisition; project administration; writing–original draft; writing–review and editing.

ACKNOWLEDGMENTS

This research was supported in part by the U.S. Department of Agriculture, Agricultural Research Service (project number 5030-21000-066-00D) and by grants from the U.S. Department of Agriculture, National Institute of Food and Agriculture, OREI program (award numbers 2020-51300-32180 and 2021-51300-34896). The findings and conclusions in this publication are those of the authors and should not be construed to represent any official USDA or United States Government determination or policy. Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.