Source: AGRICULTURAL RESEARCH SERVICE submitted to
MAIZE HEIRLOOM VARIETIES OF THE UNITED STATES
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
ACTIVE
Funding Source
AFRI COMPETITIVE GRANT
Reporting Frequency
Annual
Accession No.
1030682
Grant No.
2023-67013-40042
Cumulative Award Amt.
$649,402.00
Proposal No.
2022-10249
Multistate No.
(N/A)
Project Start Date
Sep 15, 2023
Project End Date
Sep 14, 2026
Grant Year
2023
Program Code
[A1141]- Plant Health and Production and Plant Products: Plant Breeding for Agricultural Production
Project Director
Flint-Garcia, S.
Recipient Organization
AGRICULTURAL RESEARCH SERVICE
1815 N University
Peoria,IL 61604
Performing Department
(N/A)
Non Technical Summary
Prior to adoption of hybrids in the 1930's, US corn was comprised of farmer-selected and farmer-maintained populations of diverse "heirloom varieties." Many heirlooms had specific culinary value (e.g., milling, baking, fermentation), and niche markets for these varieties still exist and are expanding today. Today's 'field corn' is high yielding but lacks both the genetic and phenotypic diversity of heirloom varieties. Over 1000 US heirloom varieties are maintained in the USDA National Plant Germplasm System, but they are rarely used in modern breeding. This may be because no comprehensive analysis of the diversity and relationships of US heirloom varieties exists, and because heirlooms lack modern agronomic characteristics (e.g., uniformity and high yield potential under high input conditions). A better understanding of heirloom diversity, along with breeding for improved production would go a long way towards making heirloom corn varieties available and profitable for small scale American farmers.This project has three major goals: (1) describe 1000 US heirloom varieties at a genetic relationship level and for a large number of plant, ear, and kernel characteristics that will indicate their level of productivity and potential usage; (2) identify the genes that control how well these 1000 heirloom varieties may grow in field conditions across the US, and the genes that control various ear and kernel traits; and (3) evaluate the yield potential of a set of promising heirloom varieties. Ultimately, this research project will lead to an understanding of local/regional adaptation of US heirloom varieties and improved taste, aroma, and nutrition.
Animal Health Component
50%
Research Effort Categories
Basic
50%
Applied
50%
Developmental
0%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
2021510108050%
2021510108150%
Knowledge Area
202 - Plant Genetic Resources;

Subject Of Investigation
1510 - Corn;

Field Of Science
1081 - Breeding; 1080 - Genetics;
Goals / Objectives
To effectively utilize the available Zea germplasm, it is essential to understand the heirloom varieties of maize of the US. This proposal will set the stage for future work exploring the culinary value of US heirloom varieties.Aim 1. Characterize the Heirloom Varieties of Maize from the United States. We will acquire germplasm from NPGS and seed savers organizations, genetically characterize a core set of that germplasm using a pooled DNA sequencing (Pool-seq) method to determine phylogenetic relationships and measure diversity within and among varieties, and to identify relationships between the US heirloom varieties and representative Mexican accessions, and characterize the heirloom varieties for a wide range of whole plant, ear, and kernel phenotypes related to adaptation, agronomic performance, and food characteristics. These analyses will be the first ever systematic characterization of the Heirloom Varieties of Maize of the United States.Aim 2. Genome Wide Association Studies and Population Genetic Analyses to Identify Genes with Large Effects on Key Traits. We will conduct genome wide association studies (GWAS) for a number of traits related to adaptation, agronomic performance, and food characteristics. We will develop new methods and conduct GWAS using allelic frequency estimates from Pool-seq for genotypes and trait values on an accession-mean basis. In addition, we will perform specialized GWAS to detect allelic variants associated with long-term climate characteristics of collection sites to detect genes related to environmental adaptation. We will use population genetic analyses to identify SNPs most highly differentiated between different maize groups defined on the basis of kernel type/food use. These analyses will provide a catalog of heirloom haplotypes affecting adaptation, morphology, and food characteristics that can be targeted for selection in future breeding efforts.Aim 3. Evaluate Yield Potential and Test Effects of Selection in a Subset of Heirloom Varieties. We previously identified 12 heirloom varieties with an array of kernel and ear morphologies from the experiments described above in Preliminary Data. Classical breeding methods were used to improve these heirloom varieties for agronomic traits while retaining their primary characteristics and food quality traits. We will phenotypically compare the original heirloom varieties to the improved heirloom varieties to determine the response to selection for traits related to agricultural productivity and food uses. Based on initial field screenings of heirloom varieties, we will identify a set of an additional 50 heirloom varieties that appear to have agronomic merit. We will evaluate yield for these 62 (12 + 50) heirloom varieties compared to commercial white food corn hybrids in replicated yield trials in MO and NC. The breeding program will result in a set of improved heirloom varieties and an understanding of the potential for heirloom varieties in general to respond to selection for higher productivity.
Project Methods
We will characterize the heirloom varieties of the US in a similar manner to the "Landraces of Maize" monographs, but also modernize the analyses by inclusion of genotypic and unoccupied aerial vehicle (UAV) data. We will also include heirloom varieties from Canada when available. This characterization of the maize heirloom varieties will be an invaluable source of information to curators at NPGS, breeders, small holder farmers, and heirloom and indigenous seed organizations.Aim 1. Characterize the Heirloom Varieties of Maize from the United StatesWe will acquire existing germplasm to assemble a comprehensive set of available US and Canadian accessions. We will also include a small number of key Mexican heirloom varieties for rooting phylogenies and as controls to verify previously proposed relationships between US and Mexican heirloom varieties. In total, we anticipate assembling a set of 1000 heirloom varieties.The 1000 accessions will be grown in partially replicated field trials in both North Carolina and Missouri in years 1 and 2. We will use a combination of UAV and manual methods to collect data plant maturity and architecture. Ears and shelled kernels will be imaged using a high-throughput ear imaging protocol. The grain from the self-pollinated ears will be used for physical quality analysis and NIR-based estimation of kernel starch, protein, and oil content.We will use Pool-seq for genotyping the heirloom accessions. We propose to select a core sample of 300 accessions and sequence pooled DNA samples of 12 plants within each accession. The core set for sequencing will be selected to represent the known diversity of geographic and ecological origins, and ear and kernel phenotypes. In addition, we will sample 20 typical accessions of key Mexican maize groups and 6 accessions of the closest wild relative of maize (Zea mays ssp. parviglumis). This method is expected to provide average 24X coverage of each accession, and allele frequencies rather than SNP calls for up to 83 million SNPs.The raw plot-level data will be analyzed using linear mixed models to estimate accession mean values adjusted for field block and macro-environment effects, and to provide estimates of the heritability of accession mean differences, the importance of accession-by-environment interaction variance, and adjusted accession mean values across all environments for each trait. The adjusted accession means will then be standardized within traits, used for trait correlation analysis, and subjected to numerical taxonomy analysis to measure Euclidean distances between accessions in multivariate trait space. Formal cluster analysis will be applied to identify appropriate distances for delineating groups and obtaining natural groupings of the accessions/heirloom varieties. Groupings will be compared to geographic origins, kernel types, and heirloom names to understand how these factors impact overall phenotypic relatedness.Marker data will be used to estimate hierarchical partitioning of genetic variation among groups defined by geographic regions or kernel types and among accessions within groups. We will test if accessions with common names are genetically distinct. We will also perform phylogenetic and cluster analysis on the marker data [89], permitting comparison of similarity measures and natural groupings on the basis of morphological data versus genotypic data. This information will be valuable for optimizing sampling and maintenance of accessions within NPGS, giving breeders some insight into how much sampling should prioritize adding more collections at the cost of fewer plants per collection.Aim 2. Genome Wide Association Studies (GWAS) and Population Genetic Analyses to Identify Genes with Large Effects on Adaptation and Ear/Kernel MorphologyWe will use GWAS to identify genetic variants with large effects on adaptation and ear or kernel morphology traits. In order to connect trait data to marker data properly, we will summarize allele frequencies within each accession at each locus using custom scripts. An additive GWAS will be performed by regression of accession mean values for each trait on minor allele frequencies at each locus with appropriate adjustment for population structure (principal components) and kinship.We will conduct environmental GWAS to identify genes associated with environmental adaptation using historical climate data from collection site. We will extract historical (1984 to 2013) climate data (minimum and maximum temperatures, precipitation, solar radiation, and atmospheric humidity) using the nasapower package in R. We will use historical information on cropping in each region and photoperiod data for the first 60 days of the estimated growing seasons from the U.S. Naval Observatory database. Then, for each accession, we will conduct an environmental GWAS by connecting allele frequencies to the historical climate variables, again adjusting for population structure and kinship. We will use these results to investigate how the major flowering time genes and perhaps currently unknown genes influence the adaptation of these accessions to different geographical regions of the US and Canada.We will test if the allelic distribution of these major effect variants depends on geographic origins and historically recognized groupings. We will overlay the allelic variation at haplotypes surrounding the major candidate flowering time genes on the collection sites of the accessions to understand how much selection on these specific genes has driven their adaptation to different latitudes and ecologies within geographic regions. We will also test individual SNPs for allele frequency differentiation between the geographic and kernel type groupings defined in Aim 1, so see if we can identify specific genomic regions and candidate genes associated with group differentiation.Aim 3. Evaluate Yield Potential and Test Effects of Selection in a Subset of Heirloom Varieties.Maize heirloom varieties are, by definition, open-pollinated populations that have been subjected to natural selection and to farmer selection, but not to formal 'scientific' selection. Previously, we chose 12 heirloom populations of mostly Southeastern US origin for improvement in our breeding program. We subjected these 12 heirlooms to two cycles of recurrent S1 selection based on visual evaluations of productivity, lodging resistance, and healthy ears at harvest in Clayton, NC and Columbia, MO. Importantly, we performed selection and intermating only within each heirloom, so that the resulting improved generations will still represent the named variety and maintain their unique food use characteristics.We will test whether agronomic and productivity traits were improved and if kernel composition characteristics within each heirloom were maintained. In addition, we will test whether the 12 heirlooms responded to selection differently, and if selection in different environments impacted the resulting C2 generation. The original C0 and two C2 (MO and NC) generations of each heirloom will be planted in split-plot designs with two-row plots and three replications at two locations in each of MO and NC. Heirloom will be the whole-plot factor and generation of selection will be the sub-plot factor, maximizing precision on the comparison of selected vs. original generations within each heirloom. Key agronomic performance traits including ear height, lodging resistance, grain yield, test weight, and grain moisture will be measured with a mechanical combine.Based on initial field screenings of heirloom varieties for Aim 1, we will identify a set of 50 additional heirloom varieties that appear to have agronomic merit and represent the diversity of grain types used for special culinary uses. We will evaluate yield for these 50 heirloom varieties in a replicated yield trial consisting of 3 replicates in each of MO and NC in year 3.

Progress 09/15/23 to 09/14/24

Outputs
Target Audience:To our knowledge there is no private investment in the use of diverse heirloom varieties, and the current market demand is insufficient to attract significant private industry investment. Nevertheless, specialty food corn markets are developing, and small-scale farmers are interested in supplying these markets, but they need information on the performance of the many available varieties, and they need varieties with better yield, disease, lodging resistance, and superior culinary value. Our research aims to address these needs in an open-source public-serving manner. Changes/Problems:We originally planned to perform pooled sequencing on each heirloom accession, but have modified our sequencing strategy following feedback from the community that individual plant level whole genome sequencing will be of greater use to the community and likely involve fewer complications with sample processing and analysis. We've modified our genotyping plan to instead perform whole genome sequencing at approximately 2x depth on 10 individual plants per accession for 846 accessions. This modified plan exceeds the depth and breadth of genotyping outlined in the project proposal. What opportunities for training and professional development has the project provided?Training activities Danielle Davis, Senior Capstone Project in Plant Sciences, University of Missouri - "Exploring Maize Heirloom Varieties for Whiskey and Tortilla Flavor" Less than one percent of corn grown in the United States is used for human consumption. Despite this low percentage, there is a growing demand for corn food in the United States as distilleries and restaurants recognize the effect of ingredients on flavor. Heirloom corn varieties are especially suited to culinary uses because they possess high genetic diversity for traits which can be used in a culinary setting. My project evaluated the protein and starch content of corn from 360 crosses among 180 heirloom varieties, the so-called "Heirloom Marriages." Protein and starch are the primary factors in determining the quality of corn used in cooking. We used Near Infrared Spectroscopy to measure grain protein and starch content in intact kernels, which allows for downstream taste testing. Because the metabolites underlying flavor are unknown, the taste testing can be used to establish a relationship between protein and starch content and flavor. Danielle researched the requirements for and successfully obtained Institutional Review Board (IRB) approval for conducting a sensory panel to taste the tortillas we produced over two separate dates in February 2023. We chose ten heirloom varieties from the US and Latin America and five Heirloom marriages for tortilla production based on their grain composition profiles and grain color. Corn was nixtamalized following standard protocols using 1% pickling lime (NaOH), the nixtamal was ground on our stone mill to produce masa (dough), and tortillas were produced manually using tortilla presses. We hosted two sensory panels of 30 volunteers each to determine whether participants prefer tortillas made from Heirloom corn varieties or from Heirloom Marriage. For each panel, participants were asked to evaluate three sets of tortillas, where each set contained three tortillas: one made from the Heirloom Marriage grain and a tortilla made from each of the two parent heirlooms. One set was repeated to account for differences between the two panels, and allowed data from both panels to be combined. The participants were asked to evaluate the tortillas based on seven characteristics: color appeal, chewiness, surface uniformity, taste, aroma, hardness, and rollability. While not a publishable study, the processes for obtaining IRB approval and working through the logistics of conducting an untrained sensory panel. This information will be valuable as we move forward on food projects in the lab. Professional development activities Graduate student, Melissa Draves. Corn Breeding Meeting. Raleigh, NC. March 2024. Graduate student, Jordan Cummings. Corn Breeding Meeting. Raleigh, NC. March 2024. Graduate student, Melissa Draves. Maize Genetics Meeting. Raleigh, NC. March 2024. Graduate student, Jordan Cummings. Maize Genetics Meeting. Raleigh, NC. March 2024. Graduate student, Jordan Cummings. NC State Plant Breeding Consortium Spring Symposium. Raleigh, NC. April 2024. Graduate student, Melissa Draves. "Entering Mentoring" workshop at University of Missouri through the Office of Undergraduate Research. January-March 2024. Graduate student, Melissa Draves. Maximizing Access to Research Careers (MARC) mentor for two senior undergraduate students. Meeting topics focus on graduate school applications, practice interviews, and poster presentations. September 2024-present. Graduate student, Melissa Draves. Steering Committee for the 2025 Maize Genetics Meeting. March 2024-present. Graduate student, Jordan Cummings. NC State Genetics and Genomics Scholars Showcase. Raleigh, NC. March 2024. How have the results been disseminated to communities of interest?"Tastier tortillas and whiskey? The secret is in the corn." University of Missouri student researcher Danny Davis is crossbreeding heirloom varieties of corn to create healthier, tastier corn tortillas and whiskey. SHOW ME MIZZOU. April 17, 2024. https://showme.missouri.edu/2024/tastier-tortillas-and-whiskey-the-secret-is-in-the-corn/. "What type of corn makes the best tortilla?" Danny Davis is working to answer that question. She's one of the many students presenting at Show Me Research Week. Mizzou Instagram. April 8, 2024. https://www.instagram.com/reel/C5gUgdLOuaB/ "A corn homecoming" by Lindsey Liles describes corn rematriation efforts of Nancy Strickland Chavis and the Lumbee Tribe of North Carolina, in collaboration with Dr. Holland. Photographs show plants and ears from our experimental fields. Garden and Gun magazine, 2024 Oct/Nov Issue, page 62 - 64. https://gardenandgun.com/articles/a-homecoming-for-corn/ What do you plan to do during the next reporting period to accomplish the goals?We will phenotype the ears, cobs, and kernels from the 2024 season using the ear scanning pipeline described above. The images will be uploaded to Cyverse and analyzed for yield component traits using the Maize Kernel-Ear-Cob Analysis pipeline developed by the University of Wisconsin. The grain will be scanned with NIR to estimate protein, starch and oil content, and will be analyzed for physical quality analysis such as test weight and kernel hardness. We plan to proceed with DNA extraction, library preparation, and whole genome sequencing for 70% of samples during Fall 2024 through Spring 2025. The additional 30% will be completed in Fall 2025. Simultaneously, we are testing protocols for high throughput RNA extraction for the seedling leaf samples collected in NC and MO in summer 2024 and will proceed with 3' RNAseq library prep and sequencing of some or all samples. The 2025 field trial will be planted in NC and MO to complete the partially replicated design. The same traits will be collected in 2025 as described above for the 2024 trial.

Impacts
What was accomplished under these goals? Aim 1 The 2024 field trials in MO and NC were successful. An experimental design was developed to evaluate the 990 entries in a randomized, partially replicated design with 1.25X replication of each entry across the two locations and two years. Three plants per row were sibling-pollinated, and the remainder of the plants were allowed to open-pollinate. Field evaluations included 14-17 traits mostly measured on three plants per plot, totalling about 92,000 data points with less than 2% missing data. Field traits included: stand counts, anthesis and silking dates, plant and ear heights, leaf length and width, tassel and main spike length, tassel branch count, tillering, root and stalk lodging, ear length, ear rot (MO only), ear drop (MO only), and husk coverage (MO only). The three sibling-pollinated ears as well as three open pollinated ears were harvested. The sib-pollinated NC ears will be sent to MO, and both locations of sib-pollinated ears will be phenotyped using the ear scanning system described below. At each location, UAVs were flown weekly with RGB, and hyperspectral (MO) and/or multispectral (MO and NC) cameras. Flights began shortly after planting and continued through the growing season to maturity and into the harvest season. The UAV data will be used to extract Normalized Difference Vegetation Index (NDVI) and other plant health and productivity related indices, plant height, leaf area index, and other phenotypes. We developed a modular and scalable ear scanning pipeline in Columbia, MO for phenotyping intact ears, cobs, and kernels. A series of flatbed scanners and Raspberry Pi devices, each with a touch screen display, have been arranged, similarly to an assembly line, to maximize productivity and workflow. The number of scanner stations can be scaled to the number of operators available and the desired processing speed, but our current setup uses eight scanners. Three intact ears (our sib-pollinated ears) are scanned simultaneously on a single scanner, the kernels are shelled off the ear, and the empty cob and loose kernels are scanned separately, all at 1200 dpi. Software has been developed to automatically capture the plot information from the barcoded row tags which are scanned along with the ears, cobs, and kernels to create file names. After the scans are complete, kernel weights are collected using an external scale and images are uploaded to Cyverse and analyzed for yield component traits using the Maize Kernel-Ear-Cob Analysis software developed by the University of Wisconsin. Aim 2 We performed tissue sample collections during the Summer 2024 field experiment for RNA extraction and 3' RNAseq: seedling leaf tissue from five plants per row from both the NC and MO locations. Pooled tissue from the five plants at each location will be used to estimate transcript abundances at the plot level, where each plot is a representative sample of a larger heirloom accession. These RNAseq data will be used for transcriptome-wide association studies to understand how differential gene regulation contributes to phenotypic variation. We also collected leaf tissue for DNA extraction from up to 15 plants per row in MO. This was followed by leaf tissue collection from up to 12 plants per row in NC for any accessions that were not collected in MO. Leaf tissue will be used for genome sequencing. Aim 3 Nothing to report for this reporting period.

Publications