Recipient Organization
UNIV OF MINNESOTA
(N/A)
ST PAUL,MN 55108
Performing Department
Agronomy & Plant Genetics
Non Technical Summary
Maize is a major crop in Minnesota and in the U.S., and it has long served as a model species for developing new breeding concepts and procedures applicable to other crops. This project continues, refines, and expands a long tradition of maize breeding research at the University of Minnesota. The proposed research, which will involve empirical and simulation studies, will focus on four emergent areas. First, we will further investigate the usefulness of targeted recombination, which refers to the ability to induce or select for chromosomal recombinations that occur in places where they are predicted to maximize performance for a given trait such as yield. Second, we will investigate if and how whole-genome sequence information on maize lines can be used in breeding for complex traits. Third, we will study how genomewide markers, doubled haploids, and year-round maize nurseries can be used to maximize genetic gains while keeping the budget and the timeframe constant. Fourth, we will try to update classical statistical genetics theory to make it more relevant to current breeding practices. These first four areas have a limited social or cultural component, and they will be complemented by a sweet corn and African vegetable breeding effort for local immigrant communities in Minneapolis-Saint Paul. Overall, the main outputs from this project will be new knowledge, increased information, better breeding methods, and new germplasm.
Animal Health Component
70%
Research Effort Categories
Basic
20%
Applied
70%
Developmental
10%
Goals / Objectives
The plant species Zea mays L.--which is called 'corn' by U.S. farmers and agronomists, and 'maize' by plant breeders and geneticists--is a major crop in Minnesota and in the United States. Corn helps feed the world, provides fuel for our vehicles, is an ingredient in a host of everyday products, and is a mainstay in rural communities in the state and nation. This proposed MAES project focuses on the breeding and genetics of this important American crop. The main outputs of this research will be new knowledge, increased information, better methods for breeding non-genetically modified corn, and new germplasm.This MAES project continues, refines, and expands a long tradition of maize breeding research at the University of Minnesota. Maize breeding and genetics research at Minnesota began in earnest 106 years ago when Herbert K. Hayes arrived on the Saint Paul campus to begin, among other assignments, a maize cultivar development program. The past century has seen many changes in maize breeding--open-pollinated cultivars gave way to hybrid cultivars, per-acre yields in the U.S. increased from <30 bushels per acre to >170 bushels per acre, and the tools and resources available to maize breeders to do their craft have changed. Yet two realities remain the same from 1915: (1) corn remains vital to the U.S. economy, and (2) maize has served as a model species for the development of new breeding concepts and procedures applicable to other crop species. While Mendel's garden peas exhibited traits that can be classified in a qualitative manner (e.g., yellow versus green seeds), the most important traits in maize (including yield) exhibit a continuous type of variation. Such continuous variation has necessitated statistical genetic approaches in maize breeding.During the last quarter century, maize breeders have gained an expanded set of tools for crop improvement. These tools include cheap and abundant single nucleotide polymorphism (SNP) markers; doubled haploid technology to create homozygous lines in only two generations instead of the usual 6-7 generations of self-pollination; year-round nurseries that allow up to seven growing seasons in two years; and computers, statistical methods, and software for analyzing, interpreting, and making breeding decisions on the basis of performance data and SNP data. Whereas previous iterations of this MAES project have focused on how to use these aforementioned tools, this current MAES project will focus on four emergent areas:Targeted recombination in plant breeding. Genes located on a given chromosome are shuffled as a package along with genes on other chromosomes, unless crossing over or recombination occurs between pairs of genes on that chromosome. For the most part, breeders so far have not attempted to control or to select where recombination occurs. Targeted recombination refers to the ability to induce or select for recombinations that occur in places where they are predicted to maximize performance for a given trait (Bernardo, 2017). This MAES project will further explore the usefulness of natural (i.e., non-induced) targeted recombination for improving different traits.Sequence information in breeding for quantitative traits. Information on the whole-genome sequence of lines may become increasingly available in maize breeding programs, particularly as the cost of sequencing continues to decrease. Can routinely available sequence information lead to more productive and more adapted maize hybrids? Would sequence information provide much more value over current SNP markers that provide information not at each location in the genome but at sparser "mileposts" across the genome?Genetic gains per unit time and cost. Initial investigations on the use of SNP markers, doubled haploids, and continuous nurseries have focused on proof-of-concept experiments to examine how such tools and resources can increase genetic gains (Mayor and Bernardo, 2009; Massman et al., 2013; Brandariz and Bernardo, 2018), without much regard for time and cost. Research is now needed to how such tools and resources can be used to increase genetic gains while keeping the total costs and time required constant.Updated quantitative genetics theory. The foundations of statistical (or quantitative) genetics theory were laid in the 1910s when tools to track the inheritance of genes or chromosome segments were unavailable. Because specific genes passed from parents to offspring could not be tracked, quantitative genetics theory had to rely on probabilistic approaches that led to outcomes that are achieved on average, but for which there is variation on a case-by-case basis. In particular, random mating was assumed so that the frequencies of genotypes can be deduced from the frequencies of genes, and vice-versa. Most germplasm in breeding programs, however, do not constitute random mating populations. The availability of SNP markers allows geneticists and breeders to track the inheritance of genes and chromosome segments, thus rendering much of the classical theory irrelevant (Bernardo. 2020). Updated theory coupled with empirical approaches are needed to provide more-relevant statistical genetics theory applicable to maize breeding.The above foci represent advanced developments in maize breeding and genetics, and they have a limited social or cultural component. Since 2015, the PD has conducted a small breeding program on sweet corn, and this vegetable breeding work will be incorporated in 2021-2026 into a larger vegetable breeding effort that will be accomplished through the new (since August 2020) Plant Breeding Center (of which the PD serves as director) at the University of Minnesota:Vegetable breeding for local immigrant communities. Minneapolis and Saint Paul have a wonderfully rich concentration of immigrant communities. Many recent Twin Cities immigrants are farmers who wish to continue their vocation in Minnesota; immigrant communities also continue to enjoy their native cuisines. Yet one challenge has been the lack of adapted and improved cultivars of native vegetables from their homelands. This project will focus on improving germplasm and developing cultivars of vegetable species important to Somali and Kenyan immigrant communities.The objecrtives of this proposed research are as follows:Assess the usefulness of targeted recombination in plant breedingDetermine if and how sequence information on homozygous lines can be meaningfully used in maize improvementRefine maize breeding schemes to increase genetic gains while keeping the cost and timeframe constantDevelop and evaluate quantitative genetics theory that relaxes key unrealistic assumptionsBreed germplasm and cultivars of vegetables important to local East African immigrant communities
Project Methods
Objective 1 Usefulness of targeted recombination in plant breeding A current collaborative experiment with Bayer Crop Scienceaims to validate predicted gains from targeted recombination in maize. In several early-maturing maize populations, targeted recombination positions were identified for yield while keeping grain moisture constant. Single copies of chromosomes that carry targeted recombinations have been introgressed into a common genetic background to create introgression libraries of lines that each carry one targeted recombination. Copies of multiple chromosomes, each carrying a targeted recombination, have also been stacked. Testcrosses of these lines with targeted recombinations will be evaluated in yield trials in 2021 as the first proof-of-concept experiment to evaluate the usefulness of targeted recombination in plants.Two follow-up studies to the above proof-of-concept experiment will be conducted as a continuation of the ongoing collaborative research with Bayer Crop Science. First, we will investigate whether chromosome-specific shrinkage factors will lead to better estimates of targeted recombination positions. Targeted recombination relies on having good estimates of where recombination should occur, and the current proof-of-concept experiment uses shrinkage factors that apply across all chromosomes. But would there be value in shrinkage factors that are tailored to each chromosome? If so, would the chromosomes targeted for recombination change, and would the marker intervals to be targeted on the chromosome change? Cross-validation studies using empirical data will be conducted to help answer these questions.Second, the current proof-of-concept study focuses on stacking within a biparental cross. How much value is there in stacking chromosome copies (which carry targeted recombinations) from different breeding populations? This approach is akin to applying targeted recombination in a multiparent population and could add value in terms of increased genetic diversity. While the concept is straightforward, having comparable estimates of per-chromosome effects in different populations would be more challenging.As opposed to targeted recombination, the across-genome recombination rate can be modulated, as has been demonstrated in riceand in rapeseed. Simulation studies will be conducted to compare short-term gains from natural targeted recombination versus gains when across-genome recombination rates are increased (to accumulate favorable gene combinations) or decreased (to prevent the loss of favorable gene combinations that have already been accumulated).Objective 2 Use of sequence information on homozygous lines in maize improvementBayer Crop Science will be providing access to phenotypic and sequence information for diverse maize lines, and this objective will involve two areas of investigation. First, can sequence information, coupled with phenotypic data, be used to have better estimates of where targeted recombinations should occur within a given chromosome? A limitation of SNP data is that the estimated positions of targeted recombination are coarse. Suppose recombination is estimated to ideally occur between marker 1, which is at the 10 cM position, and marker 2, which is at the 16 cM position. Under the current approach, recombinations at the 11, 12, 13, 14, and 15 cM positions are all deemed equivalent. The sequence and phenotypic data will be used to assess the sensitivity of the predicted genetic gain to the recombination point within a marker interval.Second, will the use of sequence information in breeding simply be akin to having information on an extremely large number of SNPs, or are there ways to utilize sequence information above and beyond having an extremely large number of generic markers? It is entirely possible that because large blocks of a chromosome or even entire chromosomes are transmitted from parents to offspring--especially with doubled haploids for which only one meiotic event is involved--that sequence information will not add any value of current SNP markers. This remains to be seen.Objective 3 Genetic gains per unit time and costSimulation research on maize breeding schemes with budget and time constraintswill be expanded. Possible extensions of this research may involve phenomic selection, which involves the use of hyperspectral imaging data from unmanned aerial vehicles as a substitute for SNP data. Prediction accuracies may be lower with hyperspectral data than with SNP data, but such lower accuracies could be offset by lower costs. The simulation software, which unlike other simulation packages include budget constraints, will be released as a stand-alone package to allow the maize breeding community to investigate breeding schemes tailored to their own breeding contexts.In maize and in other species, genetic improvement has been achieved not only through steady genetic gains but also through the identification of cultivars with outlier-type performance. Such outliers have generally been considered fortuitous. Simulation studies that incorporate empirical data will be conducted in an attempt to identify general factors or parental characteristics that lead to a higher frequency of outliers among the progeny. A challenge in this research is that outliers, by definition, defy expectations from a model. Machine learning approaches might be of value in this particular study.Objective 4 Updated quantitative genetics theoryThe covariance between relatives is a key concept in classical quantitative genetics. Such resemblance between relatives arises from alleles being identical by descent (i.e., copies of the same ancestral allele) between individuals that share a common ancestry. But could there also be a covariance between nonrelatives, particularly if population structure is present?In particular, genomewide prediction has been found effective in a sorghum germplasm collection in which relatives were assumed absent but in which population structure was present. The results from this sorghum study suggest the possibility of a covariance between nonrelatives. A covariance between nonrelatives, if it indeed exists, will be due to alleles that are alike in state but are not identical by descent. In turn, the probability alikeness in state between nonrelatives would putatively be driven by allele frequencies in the respective subpopulations from which two nonrelatives are derived.Theory for the covariance between relative and nonrelatives with be formalized, and such theory will be tested in collections of lines that have been phenotyped and genotyped and other published studies. Such covariance between nonrelatives is expected to contribute to a higher predictive ability in genomewide prediction. Input from quantitative geneticists will be sought.Objective 5 Vegetable breeding for local immigrant communitiesThis work will be done through a vegetable breeding working group, composed primarily of graduate students and postdoctoral research associates, at the Plant Breeding Center. With the help of Natalie Hoidal (extension education in the Department of Horticultural Science), the working group will identify a priority set of African vegetables, most likely including the following: jute mallow (Corchorus olitorius), spider wisp (Cleome gynandra), Ethiopian cabbage (Brassica carinata), and different Hibiscus species. Current work on sweet corn breeding will complement this proposed research. Similar work with other immigrant communities will be considered once this initial work becomes established.The PD will provide input to the working group on different aspects of germplasm collection, trait prioritization, progeny development, and evaluation of experimental lines. A participatory breeding approach will be implemented in partnership with local immigrant growers and farms, including Big River Farms, Dawn2Dusk Farm, and the Somali American Farmers Association.