Registration of Non-Stiff Stalk Maize Germplasms : Increasing Short-Season Genetic Diversity with the EarlyGEM Program

NDEarlyGEM4 (Reg. No. GP-600, PI 675375), NDEarlyGEM5 (Reg. No. GP-601, PI 675376), NDEarlyGEM22a (Reg. No. GP602, PI 675381), NDEarlyGEM22b (Reg. No. GP-603, PI 675382), and NDEarlyGEM26 (Reg. No. GP-604, PI 675383) are new nonStiff Stalk maize (Zea mays L.) populations from the EarlyGEM program. The designation “ND” represents North Dakota for national registration to recognize efforts of germplasm adaptation and improvement at North Dakota State University, Fargo. “EarlyGEM” was added to also recognize all Germplasm Enhancement Maize (GEM) network efforts that made original GEM germplasms available for adaptation to short-season environments. The North Dakota Agricultural Experiment Station released these populations on 17 Jan. 2014. The North Dakota maize breeding program is the sole genetic provider of EarlyGEM genetic materials. The objective of the EarlyGEM program is to increase the genetic diversity of short-season hybrids with unique sources of germplasm currently not available in the northern US industry. Seven EarlyGEM inbred lines have been developed from these released populations and have been shared with industry exclusively.

N DEarlyGEM4 (Reg.No. GP-600, PI 675375), NDEarlyGEM5 (Reg.No. GP-601, PI 675376), NDEarlyGEM22a (Reg.No. GP-602, PI 675381), NDEarlyGEM22b (Reg.No. GP-603, PI 675382), and NDEar-lyGEM26 (Reg.No. GP-604, PI 675383) are new non-Stiff Stalk (SS) maize (Zea mays L.) populations from the EarlyGEM program.They were coded as "ND" populations once the process of adaptation to North Dakota ended.The designation represents North Dakota for national registration to recognize efforts of germplasm adaptation and improvement at North Dakota State University (NDSU), Fargo."EarlyGEM" has been added to also recognize all Germplasm Enhancement Maize (GEM) network efforts that made original GEM germplasms available for adaptation to short-season environments.The North Dakota Agricultural Experiment Station released these populations on 17 Jan.2014.The North Dakota maize breeding program is the sole genetic provider of EarlyGEM genetic materials.Populations were released due to their uniqueness to expand the genetic diversity of temperate short-season breeding programs and their unique trait properties.The objective of the EarlyGEM program is to increase the genetic diversity of shortseason hybrids with unique sources of germplasm currently not available in the northern US industry.Seven EarlyGEM inbred lines have been developed from these released populations and have been shared with industry exclusively.Carena and Sharma (2016) explain in detail the introduction, adaptation, evaluation, and development of GEM and EarlyGEM materials.A brief summary is included below.

Methods
The GEM program was created with the support of US Congress to increase the genetic base of US hybrids (Pollak, 2003).The GEM program identified ~2% of the accessions evaluated by the Latin American Maize Program (Sevilla and Salhuana, 1997) as desirable germplasm.Breeding crosses were made between coded industry lines and elite accessions (Salhuana, 1997), and heterotic group information was preserved.From these breeding crosses, the GEM program initiated inbred line development with early-generation testing to provide elite S 3 lines to network cooperators.Original GEM efforts concentrated on central and eastern areas of the United States.In 1999, the North Dakota maize breeding program joined the GEM network with the long-term goal of increasing the genetic diversity of short-season hybrids.As a consequence, the GEM program was expanded to the northern United States through the EarlyGEM program.The goal was to move elite tropical and temperate maize germplasm northward and westward to shortseason cold-and drought-prone environments (Carena, 2003;Carena et al., 2009aCarena et al., , 2009b;;Sharma andCarena, 2012, 2016).The EarlyGEM program currently offers unique early-maturing competitive products not available in the northern US industry.
The first sets of elite GEM S 3 lines were grown in the Fargo breeding nursery for observation.Those with most adaptation potential to short seasons were crossed to released and experimental North Dakota lines.Only one backcross generation to the elite early-maturing NDSU parents was produced to reduce the time for inbred line development (Carena et al., 2009b).Over 5000 BC 1 :S 1 early-generation lines were produced for inbred line development and population creation.Long-term cooperators from foundation seed companies provided elite testers representing the most current early SS and non-SS heterotic groups for early-and late-generation testing.Testcross trials were used to evaluate the potential of early-generation lines in hybrid combinations to create the new SS and non-SS Early-GEM populations (Carena and Sharma, 2016).For the creation of non-SS populations, yield trials across multiple northern US locations per year were arranged in experiments ranging from 8 ´ 8 to 15 ´ 15 partially balanced single lattice designs.Given all data, the bulk-entry method was used to intermate top BC 1 :S 1 progenies and develop the new populations.Ranks of adjusted means for lattice effects combined over environments were used to select BC 1 -derived lines.New populations include non-SS donors SCR01:N1310-265-1-B-B (internally coded as GEM4), BR52051:N04-70-1 (coded as GEM5), CHO5015:N12-123-1-B-B (coded as GEM22), and FS8B(T):N1802-35-1-B-B (coded as GEM26).GEM22 was used in the development of two new populations carrying different groups of lines.Table 1 shows the number of progenies used as effective population size in the Ear-lyGEM released populations.
North Carolina II mating designs were utilized to evaluate the release potential of EarlyGEM populations per se and in hybrid combinations across 27 North Dakota environments.To represent each population, sample sizes were >100 individuals (e.g., >200 plants) for seed production.For population hybrid production, pair crosses were used between populations of the opposite heterotic group.Seed production of population per se and hybrids targeted equal representation of gametes.As a consequence, tassels were removed from plants used as males and females.Lattice experiments included top-performing hybrids as experiment control checks.Checks were recommended by industry long-term cooperators of the North Dakota maize breeding program, represented the 80 to 95 relative maturity (RM) range, and were annual benchmarks for all traits evaluated.Plots were planted and harvested with equipment adapted for small experimental plots.Husbandry practices were those recommended for maximum yields at the respective locations.The target population for all trials was 90,000 plants ha −1 .Different benchmarks and heritability and rank-summation indices were used not only for release purposes, but also for creating new and elite short-season exotic populations.
Data were collected for each plot at each location for grain yield (t ha −1 , adjusted on a 15.5% grain-moisture basis), grain moisture at harvest (%), test weight (kg L −1 ), stalk lodging (%), and grain quality traits.A subsample of 500 g of kernels was collected from every plot and used to measure the grain quality of all the genotypes.The near-infrared spectroscopy grain analyzer (OmegAnalyzer G) was used as a rapid and nondestructive measure of analyzing quality parameters in maize grains.Data were imported to SAS (SAS Institute, 1990) for analyses.Individual ANOVAs were computed using SAS (SAS institute, 1990) for traits within environments.The ANOVAs for each location were performed using the PROC LATTICE procedure.Entries were considered as a fixed factor, whereas replicates and blocks were treated as random.The relative efficiency of the lattice design with the randomized complete block design (RCBD) was calculated for each trait.If the relative efficiency was >105%, adjusted means by incomplete blocks were used.If the efficiency was <105%, means were not adjusted.For high-efficiency traits, the effective error was used as the denominator in the F-test instead of the RCBD error mean square.Traits with homogeneous variance were considered for combined analyses.In these cases, each location ´ year combination was considered as an environment.Adjusted and unadjusted means from each trait were used.Expected mean squares were based on a mixed linear model that considered environments and replications as random effects and entries as fixed effects.Combined error mean squares (pooled error) were calculated by pooling the corresponding individual error mean squares weighed by their corresponding degrees of freedom.Mean comparisons were assessed by Fisher's protected LSD, since it has been shown to be an adequate test for detection of differences (Carmer and Swanson, 1971).
The bulk-entry method allowed not only for the creation of new populations, but also for the identification of new earlygeneration lines that followed additional pedigree selection for the development of new lines and hybrids.The BC 1 :S 1 lines were subjected to the NDSU maize breeding pedigree selection process that has included drought and cold tolerance screening under managed-stress winter nursery environments since the 2000-2001 season (Carena et al., 2009a).Top BC 1 -derived lines from released populations were advanced and extensively tested across 51 to 64 short-season environments, in cooperation with the industry.Annually, new testers were included as recommended by industry cooperators.The purpose of both early-and late-generation testing was to discard thousands of genotypes using different experiments and testers over years to develop the next generation of diverse lines and hybrids.New EarlyGEM lines were compared, in hybrid combinations, with commercial checks.

Characteristics
The properties of each germplasm registered are based not only on the combining ability information of top BC 1 -derived progenies recombined to form the new populations, but also on the performance of populations per se and in hybrid combinations.
The NDSU maize breeding program coordinated the generation of data across up to 64 northern US environments for testcross trials.Public and private institutions participated in the extensive testing efforts to identify the top progenies used in the development of the new populations proposed for registration.Mating design data of populations per se and in hybrid combinations across 27 environments validated the release potential of new populations (Table 2).

NDEarlyGEM4
NDEarlyGEM4 is a unique synthetic germplasm created at NDSU for breeding programs developing short-season, non-SS genetically inbred lines carrying tropical alleles for grain yield and grain quality properties.It was developed by intermating 22 elite early-generation lines giving top performance with industry SS testers across North Dakota environments (Table 1).Results showed that the tropical accession SCRO1-derived lines produced hybrids with early flowering and low grain moisture at harvest.There were 31 testcrosses statistically similar (P £ 0.05) to the driest industry hybrid (Sharma and Carena, 2012).In addition, testcrosses showed the highest average grain oil (4.7 vs. 4.5%) and grain protein (9.5 vs. 8.8%) compared with top industry checks.For extractable starch content, there were four testcrosses as good as the top check hybrid for the trait.
As a population per se, NDEarlyGEM4 showed above average earliness, grain moisture at harvest, test weight, and grain starch content compared with adapted North Dakota populations and checks across 27 environments (Table 2).NDEarlyGEM4 had above average grain yield performance in combination with an NDSS (PI 675366, North Dakota Stiff Stalk) population, a synthetic made of seven elite inbred lines (A632, A641, ND278, B37, ND2000, CG102, and B73) and BSSS.However, it expressed best heterosis in combination with NDBSK(HI-M)C3 (PI 650889), population derived from open-pollinated cultivar Krug Yellow Dent, developed at the Nebraska Agriculture Experiment Station.NDBSK(HI-M)C3 was improved by 11 cycles of half-sib recurrent selection for stalk strength in Iowa [BSK(HI)C11] and three cycles of stratified mass selection for days to silk emergence in ND (Carena et al., 2008).As a consequence, NDEarlyGEM4 ´ NDBSK(HI-M)C3 can be used as a low-cost, populationhybrid alternative (Carena, 2005).
The exotic portion of the NDEarlyGEM4 germplasm was brought from the Caribbean (Table 1) to the central US Corn Belt through the USDA-GEM program and then adapted to the northern Untied States through the EarlyGEM adaptation program.NDEarlyGEM4 brought exceptional new genes not only for grain yield and early maturity combination but also for high grain quality, especially for farmers in need of maize with better ethanol utilization.NDEarlyGEM4 is a good source of shortseason genotypes for the northern United States, with ~85 RM.

NDEarlyGEM5
The NDEarlyGEM5 maize synthetic germplasm was created at NDSU for breeding programs developing short-season, non-SS, diverse inbred lines carrying tropical alleles for grain yield, starch content, and fast drydown.It was developed by intermating 26 EarlyGEM lines with high combining ability for grain yield, fast drydown, drought tolerance, and disease resistance.Tropical germplasm was brought from Brazil to the US Corn Belt through the USDA-GEM program and then adapted to the northern United States through the EarlyGEM program.The major disadvantages of exotic incorporation from tropical areas are high grain moisture and late flowering (Goodman, 1999).However, large sample sizes of tropical exotic germplasms allowed the combination of above average  grain yield, lower moisture at harvest, stalk lodging resistance, and high starch content (Table 1).
NDEarlyGEM5 is a unique genetic provider for earliness with yield.Testcrosses carrying this germplasm had the lowest grain moisture at harvest of all the entries of the experiment while not statistically (P £ 0.05) different from industry checks for grain yield (Sharma and Carena, 2012).Lines used to develop NDEarlyGEM5 were 22 d earlier maturing than original GEM S 3 versions.There were 16 additional testcrosses that were statistically similar to the top industry check hybrid for grain yield.Many of the testcrosses represented in this particular group had high yield and test weight but low grain moisture at harvest.These traits are most desirable for short-season environments like North Dakota (Carena et al., 2009a(Carena et al., , 2009b)).Days to silking showed 17 testcrosses as early as the earliest industry hybrid check.These findings reject the general hypothesis that highyielding genotypes are not found in earlier hybrids (Hawbaker et al., 1997).If more efforts were put into breeding genetically diverse genotypes for short-season environments, the relationship between moisture and yield would change even faster.Using genetically narrow germplasm provides fewer chances to obtain desirable trait combinations.
As a population per se, NDEarlyGEM5 showed top values for grain yield, earliness, grain moisture at harvest, and stalk lodging resistance compared with adapted North Dakota populations and checks across 27 environments (Table 2).NDEarlyGEM5 also had above average starch and test weight performance.In combination with NDSS, it shows best performance for grain yield, grain moisture at harvest, and grain oil content.As a consequence, NDEarlyGEM5 ´ NDSS can be used as a low-cost, population-hybrid alternative (Carena, 2005) or for reciprocal recurrent selection programs emphasizing population and inbred line development for short-season quality hybrids.
NDEarlyGEM5 is a source of lines and hybrids with early maturity for short seasons (80 RM), high yield, and the highest extractable starch content for the wet milling and ethanol processing industries.NDEarlyGEM5 is especially characterized by its fast drydown after the use of the nondestructive NDSU breeding methodology created for the accurate and highthroughput phenotyping of drydown rate (Yang et al., 2010).The incorporation of exotic alleles from this germplasm will be important for short-season hybrids targeted at ethanol production and starch-derived products.

NDEarlyGEM22a
NDEarlyGEM22a is a maize population created at NDSU for breeding programs developing short-season, non-SS, diverse inbred lines with an emphasis in test weight and grain protein.NDEarlyGEM22a was developed by recombining 21 BC 1 elite lines for top hybrid yield and quality performance with SS industry testers across North Dakota environments (Table 1).Late-maturing temperate germplasm was brought from Chile to the central US Corn Belt through the USDA-GEM program and then adapted to the northern Untied States through the EarlyGEM late ´ early backcross adaptation program.Earlygeneration lines used to develop NDEarlyGEM22a were not statistically (P £ 0.05) different from the top check for grain yield but were better for grain protein content.As a consequence, the incorporation of exotic germplasm provided superior alleles for grain quality, especially test weight and grain protein content.
As a population per se, NDEarlyGEM22a has confirmed top values for test weight, grain protein quality, and total contents compared with adapted North Dakota populations and checks across 27 environments (Table 2).In combination with NDBSK(HI-M)C3, it showed best performance for grain yield, test weight, and grain protein content.NDEarlyGEM22a is a source of lines and hybrids with intermediate maturity for short seasons (90 RM) and the highest protein content for the livestock industry.NDEarlyGEM22a is especially characterized by its highest quality and feed value due to its consistent values in test weight, grain protein, and amino acid content.The incorporation of exotic alleles from this germplasm will be important for short-season hybrids targeted at the next generation of healthier maize-derived products (Dong et al., 2012).

NDEarlyGEM22b
NDEarlyGEM22b is a maize population created at NDSU for breeding programs developing short-season, non-SS, diverse inbred lines with an emphasis in test weight and grain oil content.NDEarlyGEM22b was developed by recombining 20 Ear-lyGEM elite lines for top hybrid yield and quality performance with SS industry testers across North Dakota environments (Table 1).The same germplasm source used for NDEarly-GEM22a was used to develop NDEarlyGEM22b.Early generation lines used to develop NDEarlyGEM22b were not statistically (P £ 0.05) different from the top industry check for grain moisture at harvest but were better for grain oil content.
As a population per se, NDEarlyGEM22b has confirmed top values for test weight and grain oil contents compared with adapted North Dakota populations and checks across 27 environments (Table 2).In combination with NDSS, it showed the best performance for grain yield and test weight.NDEarly-GEM22b is a source of short-season lines and hybrids with high oil content for the livestock industry and ~90 RM.

NDEarlyGEM26
NDEarlyGEM26 maize population was created at NDSU for breeding programs developing short-season, non-SS, diverse inbred lines with unique southern US alleles.This new germplasm was developed by recombining 20 BC 1 elite lines for top hybrid performance with SS industry testers across North Dakota environments.Late-maturing temperate germplasm was brought from the southern United States to the central US Corn Belt through the USDA-GEM program and then adapted to the northern United States through the EarlyGEM program.Selected early-generation lines had, in hybrid combinations with SS testers, values that were not statistically (P £ 0.05) different from industry checks for days to silking, grain yield, grain moisture at harvest, test weight, and starch content.
The accession FS8B(T) from the southern United States was selected for special value-added traits (Pollak, 2003).As a population per se, NDEarlyGEM26 showed top values for lodging resistance and grain protein content compared with adapted North Dakota populations and checks across 27 environments (Table 2).In combination with NDSS, it showed above average performance for grain yield, grain moisture at harvest, and test weight.NDEarlyGEM26 is a source of lines and hybrids with high protein content for the livestock industry and a good source of short-season genotypes for the northern United States, with ~95 RM.
NDEarlyGEM4, NDEarlyGEM5, NDEarlyGEM22a, NDEarlyGEM22b, and NDEarlyGEM26 are unique and genetically diverse germplasm sources of short-season inbred lines and hybrids not found in current genome sequences or mapping populations, often using very small and low-diversity samples.This supports the vision of Wellhausen (1965), Goodman and Brown (1988), Salhuana (1997), Hallauer (1999), Pollak (2003), Hallauer and Carena (2009, 2013, 2015), and Hallauer et al. (2010), among others, to continue strong coordinated efforts to evaluate and incorporate exotic germplasm.They are alternative options if maize breeders have an interest in expanding the base of their current breeding programs.The Ear-lyGEM program encourages the development of the next generation of sustainable US hybrids carrying diverse genetics for future climate challenges.Adaptation and improvement programs carrying unique and diverse germplasm will be essential in the development of new cultivars tolerant to climate changes.

Population Maintenance and Distribution
The development of new NDEarlyGEM populations included a two-step process.First, top BC 1 :S 1 lines were included in the recombination nursery section, and the bulkentry method was used to create Syn1 populations.Second, the newly created populations were included in the population maintenance nursery section for additional recombination and seed increases.As a consequence, Syn2 populations were created.Crosses within each pedigree were produced among plants avoiding full-sib family production.Tassels were removed from plants used as males and females to reduce the number of overall pollinations needed and to maintain as much equal representation as possible from gametes.NDEarlyGEM maize germplasms were multiplied in Fargo and New Zealand breeding nurseries within their population maintenance nursery sections.Balanced bulks of 100 to 500 kernels were used for this purpose.Two hundred kernels will be available for distribution, and NDSU maize breeders at Fargo will coordinate multiplication of populations for the first 5 yr.

Discussion
The North Dakota maize breeding program leads the Ear-lyGEM project targeted at increasing the genetic diversity of short-season maize hybrids.This research is part of a long-term continuous effort to incorporate GEM germplasm into the northern Untied States and southern Canada.As a consequence, the industry can access unique genes for various traits and trait combinations.Released populations were demonstrated to be outstanding sources for economically important traits essential for short-season environments (Sharma and Carena, 2012).Significant (P £ 0.05) differences were detected among genotypes (Table 2), and data have shown that new NDEarlyGEM populations were very competitive per se and in hybrid combinations.In addition, new populations have served as excellent sources of new elite short-season EarlyGEM inbred lines and hybrids.Seven new EarlyGEM inbred lines (ND2018, ND2019, ND2024, ND2026, ND2027, ND2034, and ND2041), derived from released NDEarlyGEM non-SS populations, have been exclusively released to the industry and recoded.
Early-generation hybrids including non-SS EarlyGEM BC 1 :S 1 lines showed very good adaptation to short seasons.Stalk lodging was <2%, except for one testcross with an industry sister line.The nonsignificant (P £ 0.05) differences between industry checks and EarlyGEM hybrids demonstrated successful adaptation, although exotic tropical and late temperate germplasm were incorporated.Nursery observation and screening across inbreeding generations during incorporation was efficient in selecting lodging-resistant lines and hybrids.There was a significant reduction in plant height in lines derived from exotic germplasm, as compared with their initial stage of adaptation (Fig. 1).Many early-flowering lines were recovered from late exotic incorporations.There were 99 early-generation testcrosses that were not statistically (P £ 0.05) different from the earliest-maturing industry check for days of silking.Most EarlyGEM lines flowered earlier than most commercial hybrid checks.One backcross generation provided 15 to 20 d earlier maturity than the original S 3 GEM lines.On average, results showed that non-SS lines represented early to intermediate maturity for North Dakota compared with industry checks (80-95 RM).As a consequence, only one backcross was sufficient to adapt non-SS EarlyGEM lines.
Exotic incorporation of GEM germplasm was not only successful for adaptation to short-season environments, but it was also positive for economically important traits.Exotic lines, derived from released NDEarlyGEM populations, were adapted to short North Dakota seasons due to the fact that they could produce adapted and competitive hybrids with public and industry testers.There were 45 testcrosses not statistically different (P £ 0.05) from the top industry check for grain yield, 54 testcrosses for test weight, and 63 for grain moisture at harvest.Many non-SS EarlyGEM lines were identified with outstanding grain quality characteristics.There were 90 Early-GEM lines not statistically (P £ 0.05) different from the Monsanto processor-preferred hybrid for high fermentable content in grain.In addition, most EarlyGEM lines have shown above average grain protein and oil content in hybrid combinations, as well as average grain starch content.For grain starch, 23 testcrosses were not statistically different from the top 'Pioneer' check for this trait.For grain protein and oil content, 42 testcrosses were not statistically (P £ 0.05) different from the top entry, not including any of the industry checks used.The high grain protein content found in many of the EarlyGEM-derived hybrids has shown that exotic incorporation can be a unique source for ranchers in short-season environments.Overall, the EarlyGEM materials carried unique grain quality genes to add significant value to short-season hybrids.
Non-SS EarlyGEM testcrosses have shown, on average, higher yields than SS EarlyGEM testcrosses (Carena and Sharma, 2016).In addition, testcrosses including NDEarlyGEM4 and 26 lines flowered earlier than industry checks, whereas testcrosses including NDEarlyGEM5 and 22 have shown faster drying properties according to the method proposed by Yang et al. (2010).Large samples of NDEarlyGEM5 lines have shown desirable associations of yield with grain moisture at harvest, indicating that there was enough genetic diversity to achieve this trait combination, which is not often found in genetically narrow germplasm.Sharma and Carena (2012) showed testcrosses with NDEarlyGEM4, NDEarlyGEM5, and NDEarly-GEM22 lines that had high grain oil and protein contents as compared with top industry check benchmarks.