Enhancing Crop Domestication Through Genomic Selection, a Case Study of Intermediate Wheatgrass

Origin of current intermediate wheatgrass germplasm being developed for Kernza grain production

Intermediate wheatgrass (IWG, Thinopyrum intermedium [Host] Barkworth & D. R. Dewey) has been developed as a perennial grain crop to provide ecosystem services, environmental benefits, and human food. Grain and products derived from IWG varieties improved for food production have been marketed under the registered trademark, Kernza. In the 1980s, a joint breeding effort between the Rodale Institute (RI) and the Big Flats Plant Material Center used IWG plant introductions (PI) from the National Plant Germplasm System (NPGS) and recurrent phenotypic selection to improve populations of IWG with the goal of developing a perennial grain. Initial selections were provided to The Land Institute where they were subsequently improved for grain production, yet the identity of the founder material of improved, food-grade IWG has not been publicly documented. Recently recovered original documents have been used to reconstruct the early breeding program to identify the most likely 20 PIs that form the founders of modern food-grade IWG. Molecular data using genotyping-by-sequencing in current elite breeding material, remnant seed from the initial RI selections, and preserved sample material have provided supporting evidence for the historical records. The genetic origin for food-grade IWG is focused between the Black Sea and Caspian Sea in the Stavropol region of Russia, with smaller contributions likely from collections as distant as Kazakhstan in the east to Turkey in the west. This work connects the flow of germplasm and utility of NPGS PIs to present day IWG grain cultivars being developed in multiple breeding programs around the world.

A genome-wide association study and genomic prediction for Phakopsora pachyrhizi resistance in soybean

Soybean brown rust (SBR), caused by Phakopsora pachyrhizi, is a devastating fungal disease that threatens global soybean production. This study conducted a genome-wide association study (GWAS) with seven models on a panel of 3,082 soybean accessions to identify the markers associated with SBR resistance by 30,314 high quality single nucleotide polymorphism (SNPs). Then five genomic selection (GS) models, including Ridge regression best linear unbiased predictor (rrBLUP), Genomic best linear unbiased predictor (gBLUP), Bayesian least absolute shrinkage and selection operator (Bayesian LASSO), Random Forest (RF), and Support vector machines (SVM), were used to predict breeding values of SBR resistance using whole genome SNP sets and GWAS-based marker sets. Four SNPs, namely Gm18_57,223,391 (LOD = 2.69), Gm16_29,491,946 (LOD = 3.86), Gm06_45,035,185 (LOD = 4.74), and Gm18_51,994,200 (LOD = 3.60), were located near the reported P. pachyrhizi R genes, Rpp1, Rpp2, Rpp3, and Rpp4, respectively. Other significant SNPs, including Gm02_7,235,181 (LOD = 7.91), Gm02_7234594 (LOD = 7.61), Gm03_38,913,029 (LOD = 6.85), Gm04_46,003,059 (LOD = 6.03), Gm09_1,951,644 (LOD = 10.07), Gm10_39,142,024 (LOD = 7.12), Gm12_28,136,735 (LOD = 7.03), Gm13_16,350,701(LOD = 5.63), Gm14_6,185,611 (LOD = 5.51), and Gm19_44,734,953 (LOD = 6.02), were associated with abundant disease resistance genes, such as Glyma.02G084100, Glyma.03G175300, Glyma.04g189500, Glyma.09G023800, Glyma.12G160400, Glyma.13G064500, Glyma.14g073300, and Glyma.19G190200. The annotations of these genes included but not limited to: LRR class gene, cytochrome 450, cell wall structure, RCC1, NAC, ABC transporter, F-box domain, etc. The GWAS based markers showed more accuracies in genomic prediction than the whole genome SNPs, and Bayesian LASSO model was the ideal model in SBR resistance prediction with 44.5% ~ 60.4% accuracies. This study aids breeders in predicting selection accuracy of complex traits such as disease resistance and can shorten the soybean breeding cycle by the identified markers.

Crop genomic selection with deep learning and environmental data: A survey

Machine learning techniques for crop genomic selections, especially for single-environment plants, are well-developed. These machine learning models, which use dense genome-wide markers to predict phenotype, routinely perform well on single-environment datasets, especially for complex traits affected by multiple markers. On the other hand, machine learning models for predicting crop phenotype, especially deep learning models, using datasets that span different environmental conditions, have only recently emerged. Models that can accept heterogeneous data sources, such as temperature, soil conditions and precipitation, are natural choices for modeling GxE in multi-environment prediction. Here, we review emerging deep learning techniques that incorporate environmental data directly into genomic selection models.

The use of wheatgrass (<i>Thinopyrum intermedium</i>) in breeding

Wheatgrass (Th. intermedium) has been traditionally used in wheat breeding for obtaining wheat-wheatgrass hybrids and varieties with introgressions of new genes for economically valuable traits. However, in the 1980s in the United States wheatgrass was selected from among perennial plant species as having promise for domestication and the development of dual-purpose varieties for grain (as an alternative to perennial wheat) and hay. The result of this work was the creation of the wheatgrass varieties Kernza (The Land Institute, Kansas) and MN-Clearwater (University of Minnesota, Minnesota). In Omsk State Agrarian University, the variety Sova was developed by mass selection of the most winter-hardy biotypes with their subsequent combination from the population of wheatgrass obtained from The Land Institute. The average grain yield of the variety Sova is 9.2 dt/ha, green mass is 210.0 dt/ ha, and hay is 71.0 dt/ha. Wheatgrass is a crop with a large production potential, benef icial environmental properties, and valuable grain for functional food. Many publications show the advantages of growing the Kernza variety compared to annual crops in reducing groundwater nitrate contamination, increasing soil carbon sequestration, and reducing energy and economic costs. However, breeding programs for domestication of perennial crops are very limited in Russia. This paper presents an overview of main tasks faced by breeders, aimed at enhancing the yield and cultivating wheatgrass eff iciency as a perennial grain and fodder crop. To address them, both traditional and modern biotechnological and molecular cytogenetic approaches are used. The most important task is to transfer target genes of Th. intermedium to modern wheat varieties and decrease the level of chromatin carrying undesirable genes of the wild relative. The f irst consensus map of wheatgrass containing 10,029 markers was obtained, which is important for searching for genes and their introgressions to the wheat genome. The results of research on the nutritional and technological properties of wheatgrass grain for the development of food products as well as the differences in the quality of wheatgrass grain and wheat grain are presented.

Genetic architecture and QTL selection response for Kernza perennial grain domestication traits

Analysis of multi-year breeding program data revealed that the genetic architecture of an intermediate wheatgrass population was highly polygenic for both domestication and agronomic traits, supporting the use of genomic selection for new crop domestication. Perennial grains have the potential to provide food for humans and decrease the negative impacts of annual agriculture. Intermediate wheatgrass (IWG, Thinopyrum intermedium, Kernza®) is a promising perennial grain candidate that The Land Institute has been breeding since 2003. We evaluated four consecutive breeding cycles of IWG from 2016 to 2020 with each cycle containing approximately 1100 unique genets. Using genotyping-by-sequencing markers, quantitative trait loci (QTL) were mapped for 34 different traits using genome-wide association analysis. Combining data across cycles and years, we found 93 marker-trait associations for 16 different traits, with each association explaining 0.8-5.2% of the observed phenotypic variance. Across the four cycles, only three QTL showed an F_ST differentiation > 0.15 with two corresponding to a decrease in floret shattering. Additionally, one marker associated with brittle rachis was 216 bp from an ortholog of the btr2 gene. Power analysis and quantitative genetic theory were used to estimate the effective number of QTL, which ranged from a minimum of 33 up to 558 QTL for individual traits. This study suggests that key agronomic and domestication traits are under polygenic control and that molecular methods like genomic selection are needed to accelerate domestication and improvement of this new crop.

Perennials as Future Grain Crops: Opportunities and Challenges

Perennial grain crops could make a valuable addition to sustainable agriculture, potentially even as an alternative to their annual counterparts. The ability of perennials to grow year after year significantly reduces the number of agricultural inputs required, in terms of both planting and weed control, while reduced tillage improves soil health and on-farm biodiversity. Presently, perennial grain crops are not grown at large scale, mainly due to their early stages of domestication and current low yields. Narrowing the yield gap between perennial and annual grain crops will depend on characterizing differences in their life cycles, resource allocation, and reproductive strategies and understanding the trade-offs between annualism, perennialism, and yield. The genetic and biochemical pathways controlling plant growth, physiology, and senescence should be analyzed in perennial crop plants. This information could then be used to facilitate tailored genetic improvement of selected perennial grain crops to improve agronomic traits and enhance yield, while maintaining the benefits associated with perennialism.

Using Genomic Selection to Develop Performance-Based Restoration Plant Materials

Effective native plant materials are critical to restoring the structure and function of extensively modified ecosystems, such as the sagebrush steppe of North America’s Intermountain West. The reestablishment of native bunchgrasses, e.g., bluebunch wheatgrass (Pseudoroegneria spicata [Pursh] À. Löve), is the first step for recovery from invasive species and frequent wildfire and towards greater ecosystem resiliency. Effective native plant material exhibits functional traits that confer ecological fitness, phenotypic plasticity that enables adaptation to the local environment, and genetic variation that facilitates rapid evolution to local conditions, i.e., local adaptation. Here we illustrate a multi-disciplinary approach based on genomic selection to develop plant materials that address environmental issues that constrain local populations in altered ecosystems. Based on DNA sequence, genomic selection allows rapid screening of large numbers of seedlings, even for traits expressed only in more mature plants. Plants are genotyped and phenotyped in a training population to develop a genome model for the desired phenotype. Populations with modified phenotypes can be used to identify plant syndromes and test basic hypotheses regarding relationships of traits to adaptation and to one another. The effectiveness of genomic selection in crop and livestock breeding suggests this approach has tremendous potential for improving restoration outcomes for species such as bluebunch wheatgrass.

Genome-wide association mapping and genomic prediction for kernel color traits in intermediate wheatgrass (Thinopyrum intermedium)

BACKGROUND

Intermediate wheatgrass (IWG) is a novel perennial grain crop currently undergoing domestication. It offers important ecosystem benefits while producing grain suitable for human consumption. Several aspects of plant biology and genetic control are yet to be studied in this new crop. To understand trait behavior and genetic characterization of kernel color in IWG breeding germplasm from the University of Minnesota was evaluated for the CIELAB components (L*, a*, b*) and visual differences. Trait values were used in a genome-wide association scan to reveal genomic regions controlling IWG's kernel color. The usability of genomic prediction in predicting kernel color traits was also evaluated using a four-fold cross validation method.

RESULTS

A wide phenotypic variation was observed for all four kernel color traits with pairwise trait correlations ranging from - 0.85 to 0.27. Medium to high estimates of broad sense trait heritabilities were observed and ranged from 0.41 to 0.78. A genome-wide association scan with single SNP markers detected 20 significant marker-trait associations in 9 chromosomes and 23 associations in 10 chromosomes using multi-allelic haplotype blocks. Four of the 20 significant SNP markers and six of the 23 significant haplotype blocks were common between two or more traits. Evaluation of genomic prediction of kernel color traits revealed the visual score to have highest mean predictive ability (r² = 0.53); r² for the CIELAB traits ranged from 0.29-0.33. A search for candidate genes led to detection of seven IWG genes in strong alignment with MYB36 transcription factors from other cereal crops of the Triticeae tribe. Three of these seven IWG genes had moderate similarities with R-A1, R-B1, and R-D1, the three genes that control grain color in wheat.

CONCLUSIONS

We characterized the distribution of kernel color in IWG for the first time, which revealed a broad phenotypic diversity in an elite breeding germplasm. Identification of genetic loci controlling the trait and a proof-of-concept that genomic selection might be useful in selecting genotypes of interest could help accelerate the breeding of this novel crop towards specific end-use.

Genomic Prediction and the Practical Breeding of 12 Quantitative-Inherited Traits in Cucumber (Cucumis sativus L.)

Genomic prediction is an effective way for predicting complex traits, and it is becoming more essential in horticultural crop breeding. In this study, we applied genomic prediction in the breeding of cucumber plants. Eighty-one cucumber inbred lines were genotyped and 16,662 markers were identified to represent the genetic background of cucumber. Two populations, namely, diallel cross population and North Carolina II population, having 268 combinations in total were constructed from 81 inbred lines. Twelve cucumber commercial traits of these two populations in autumn 2018, spring 2019, and spring 2020 were collected for model training. General combining ability (GCA) models under five-fold cross-validation and cross-population validation were applied to model validation. Finally, the GCA performance of 81 inbred lines was estimated. Our results showed that the predictive ability for 12 traits ranged from 0.38 to 0.95 under the cross-validation strategy and ranged from -0.38 to 0.88 under the cross-population strategy. Besides, GCA models containing non-additive effects had significantly better performance than the pure additive GCA model for most of the investigated traits. Furthermore, there were a relatively higher proportion of additive-by-additive genetic variance components estimated by the full GCA model, especially for lower heritability traits, but the proportion of dominant genetic variance components was relatively small and stable. Our findings concluded that a genomic prediction protocol based on the GCA model theoretical framework can be applied to cucumber breeding, and it can also provide a reference for the single-cross breeding system of other crops.

Nested association mapping reveals the genetic architecture of spike emergence and anthesis timing in intermediate wheatgrass

Intermediate wheatgrass (Thinopyrum intermedium) is an outcrossing, cool season grass species currently undergoing direct domestication as a perennial grain crop. Though many traits are selection targets, understanding the genetic architecture of those important for local adaptation may accelerate the domestication process. Nested association mapping (NAM) has proven useful in dissecting the genetic control of agronomic traits many crop species, but its utility in primarily outcrossing, perennial species has yet to be demonstrated. Here, we introduce an intermediate wheatgrass NAM population developed by crossing ten phenotypically divergent donor parents to an adapted common parent in a reciprocal manner, yielding 1,168 F₁ progeny from 10 families. Using genotyping by sequencing, we identified 8,003 SNP markers and developed a population-specific consensus genetic map with 3,144 markers across 21 linkage groups. Using both genomewide association mapping and linkage mapping combined across and within families, we characterized the genetic control of flowering time. In the analysis of two measures of maturity across four separate environments, we detected as many as 75 significant QTL, many of which correspond to the same regions in both analysis methods across 11 chromosomes. The results demonstrate a complex genetic control that is variable across years, locations, traits, and within families. The methods were effective at detecting previously identified QTL, as well as new QTL that align closely to the well-characterized flowering time orthologs from barley, including Ppd-H1 and Constans. Our results demonstrate the utility of the NAM population for understanding the genetic control of flowering time and its potential for application to other traits of interest.

Genomic prediction enables rapid selection of high-performing genets in an intermediate wheatgrass breeding program

In an era of constrained and depleted natural resources, perennial grains could provide sustainable food production along with beneficial ecosystem services like reduced erosion and increased atmospheric carbon capture. Intermediate wheatgrass (IWG) [Thinopyrum intermedium (Host) Barkworth & D. R. Dewey subsp. intermedium] has been undergoing continuous breeding for domestication to develop a perennial grain crop since the 1980s. As a perennial, IWG has required 2-5 yr per selection generation, but starting in 2017, genomic selection (GS) was initiated in the breeding program at The Land Institute, Salina, KS (TLI), enabling one complete cycle per year. For each cycle, ∼4,000 seedlings were profiled using genotyping-by-sequencing (GBS) and genomic estimated breeding values (GEBVs) were calculated. Selection based on GEBVs identified ∼100 individuals to advance as parents each generation, while validation populations of 1,000-1,200 genets for GS model training were also selected using the genomic relationship matrix to represent genetic diversity in each cycle. The selected parents were randomly intermated in a greenhouse crossing block to form the subsequent cycle, while the validation populations were transplanted to irrigated and nonirrigated field sites for phenotypic evaluations in the following years. For priority breeding traits of seed mass, free threshing, and nonshattering, correlations between predicted values and observed data were >.5. The realized selection differential ranged from 11-23% for selected traits, and the expected genetic gains for these traits, including spike yield, ranged from 6 to 14% per year. Genomic selection is a powerful tool to speed the domestication and development of IWG and other perennial crops with extended breeding timelines.

A classic approach for determining genomic prediction accuracy under terminal drought stress and well-watered conditions in wheat landraces and cultivars

The present study aimed to improve the accuracy of genomic prediction of 16 agronomic traits in a diverse bread wheat (Triticum aestivum L.) germplasm under terminal drought stress and well-watered conditions in semi-arid environments. An association panel including 87 bread wheat cultivars and 199 landraces from Iran bread wheat germplasm was planted under two irrigation systems in semi-arid climate zones. The whole association panel was genotyped with 9047 single nucleotide polymorphism markers using the genotyping-by-sequencing method. A number of 23 marker-trait associations were selected for traits under each condition, whereas 17 marker-trait associations were common between terminal drought stress and well-watered conditions. The identified marker-trait associations were mostly single nucleotide polymorphisms with minor allele effects. This study examined the effect of population structure, genomic selection method (ridge regression-best linear unbiased prediction, genomic best-linear unbiased predictions, and Bayesian ridge regression), training set size, and type of marker set on genomic prediction accuracy. The prediction accuracies were low (-0.32) to moderate (0.52). A marker set including 93 significant markers identified through genome-wide association studies with P values ≤ 0.001 increased the genomic prediction accuracy for all traits under both conditions. This study concluded that obtaining the highest genomic prediction accuracy depends on the extent of linkage disequilibrium, the genetic architecture of trait, genetic diversity of the population, and the genomic selection method. The results encouraged the integration of genome-wide association study and genomic selection to enhance genomic prediction accuracy in applied breeding programs.

New Food Crop Domestication in the Age of Gene Editing: Genetic, Agronomic and Cultural Change Remain Co-evolutionarily Entangled

The classic domestication scenario for grains and fruits has been portrayed as the lucky fixation of major-effect "domestication genes." Characterization of these genes plus recent improvements in generating novel alleles (e.g., by gene editing) have created great interest in de novo domestication of new crops from wild species. While new gene editing technologies may accelerate some genetic aspects of domestication, we caution that de novo domestication should be understood as an iterative process rather than a singular event. Changes in human social preferences and relationships and ongoing agronomic innovation, along with broad genetic changes, may be foundational. Allele frequency changes at many loci controlling quantitative traits not normally included in the domestication syndrome may be required to achieve sufficient yield, quality, defense, and broad adaptation. The environments, practices and tools developed and maintained by farmers and researchers over generations contribute to crop yield and success, yet those may not be appropriate for new crops without a history of agronomy. New crops must compete with crops that benefit from long-standing participation in human cultural evolution; adoption of new crops may require accelerating the evolution of new crops' culinary and cultural significance, the emergence of markets and trade, and the formation and support of agricultural and scholarly institutions. We provide a practical framework that highlights and integrates these genetic, agronomic, and cultural drivers of change to conceptualize de novo domestication for communities of new crop domesticators, growers and consumers. Major gene-focused domestication may be valuable in creating allele variants that are critical to domestication but will not alone result in widespread and ongoing cultivation of new crops. Gene editing does not bypass or diminish the need for classical breeding, ethnobotanical and horticultural knowledge, local agronomy and crop protection research and extension, farmer participation, and social and cultural research and outreach. To realize the ecological and social benefits that a new era of de novo domestication could offer, we call on funding agencies, proposal reviewers and authors, and research communities to value and support these disciplines and approaches as essential to the success of the breakthroughs that are expected from gene editing techniques.