Identification of Quantitative Trait Loci Hotspots Affecting Agronomic Traits and High-Throughput Vegetation Indices in Rainfed Wheat

High-throughput phenotyping using hyperspectral indicators supports the genetic dissection of yield in durum wheat grown under heat and drought stress

High-throughput phenotyping (HTP) provides new opportunities for efficiently dissecting the genetic basis of drought-adaptive traits, which is essential in current wheat breeding programs. The combined use of HTP and genome-wide association (GWAS) approaches has been useful in the assessment of complex traits such as yield, under field stress conditions including heat and drought. The aim of this study was to identify molecular markers associated with yield (YLD) in elite durum wheat that could be explained using hyperspectral indices (HSIs) under drought field conditions in Mediterranean environments in Southern Spain. The HSIs were obtained from hyperspectral imagery collected during the pre-anthesis and anthesis crop stages using an airborne platform. A panel of 536 durum wheat lines were genotyped by sequencing (GBS, DArTseq) to determine population structure, revealing a lack of genetic structure in the breeding germplasm. The material was phenotyped for YLD and 19 HSIs for six growing seasons under drought field conditions at two locations in Andalusia, in southern Spain. GWAS analysis identified 740 significant marker-trait associations (MTAs) across all the durum wheat chromosomes, several of which were common for YLD and the HSIs, and can potentially be integrated into breeding programs. Candidate gene (CG) analysis uncovered genes related to important plant processes such as photosynthesis, regulatory biological processes, and plant abiotic stress tolerance. These results are novel in that they combine high-resolution hyperspectral imaging at the field scale with GWAS analysis in wheat. They also support the use of HSIs as useful tools for identifying chromosomal regions related to the heat and drought stress response in wheat, and pave the way for the integration of field HTP in wheat breeding programs.

Phenotyping cotton leaf chlorophyll via in situ hyperspectral reflectance sensing, spectral vegetation indices, and machine learning

Cotton (Gossypium hirsutum L.) leaf chlorophyll (Chl) has been targeted as a phenotype for breeding selection to improve cotton tolerance to environmental stress. However, high-throughput phenotyping methods based on hyperspectral reflectance sensing are needed to rapidly screen cultivars for chlorophyll in the field. The objectives of this study were to deploy a cart-based field spectroradiometer to measure cotton leaf reflectance in two field experiments over four growing seasons at Maricopa, Arizona and to evaluate 148 spectral vegetation indices (SVI's) and 14 machine learning methods (MLM's) for estimating leaf chlorophyll from spectral information. Leaf tissue was sampled concurrently with reflectance measurements, and laboratory processing provided leaf Chl a, Chl b, and Chl a+b as both areas-basis (µg cm-2) and mass-basis (mg g-1) measurements. Leaf reflectance along with several data transformations involving spectral derivatives, log-inverse reflectance, and SVI's were evaluated as MLM input. Models trained with 2019-2020 data performed poorly in tests with 2021-2022 data (e.g., RMSE=23.7% and r2 = 0.46 for area-basis Chl a+b), indicating difficulty transferring models between experiments. Performance was more satisfactory when training and testing data were based on a random split of all data from both experiments (e.g., RMSE=10.5% and r2 = 0.88 for area basis Chl a+b), but performance beyond the conditions of the present study cannot be guaranteed. Performance of SVI's was in the middle (e.g., RMSE=16.2% and r2 = 0.69 for area-basis Chl a+b), and SVI's provided more consistent error metrics compared to MLM's. Ensemble MLM's which combined estimates from several base estimators (e.g., random forest, gradient booting, and AdaBoost regressors) and a multi-layer perceptron neural network method performed best among MLM's. Input features based on spectral derivatives or SVI's improved MLM's performance compared to inputting reflectance data. Spectral reflectance data and SVI's involving red edge radiation were the most important inputs to MLM's for estimation of cotton leaf chlorophyll. Because MLM's struggled to perform beyond the constraints of their training data, SVI's should not be overlooked as practical plant trait estimators for high-throughput phenotyping, whereas MLM's offer great opportunity for data mining to develop more robust indices.

GWAS in a Collection of Bulgarian Old and Modern Bread Wheat Accessions Uncovers Novel Genomic Loci for Grain Protein Content and Thousand Kernel Weight

Genetic enhancement of grain production and quality is a priority in wheat breeding projects. In this study, we assessed two key agronomic traits-grain protein content (GPC) and thousand kernel weight (TKW)-across 179 Bulgarian contemporary and historic varieties and landraces across three growing seasons. Significant phenotypic variation existed for both traits among genotypes and seasons, and no discernible difference was evident between the old and modern accessions. To understand the genetic basis of the traits, we conducted a genome-wide association study with MLM using phenotypic data from the crop seasons, best linear unbiased estimators, and genotypic data from the 25K Infinium iSelect array. As a result, we detected 16 quantitative trait nucleotides (QTNs) associated with GPC and 15 associated with TKW, all of which passed the false discovery rate threshold. Seven loci favorably influenced GPC, resulting in an increase of 1.4% to 8.1%, while four loci had a positive impact on TKW with increases ranging from 1.9% to 8.4%. While some loci confirmed previously published associations, four QTNs linked to GPC on chromosomes 2A, 7A, and 7B, as well as two QTNs related to TKW on chromosomes 1B and 6A, may represent novel associations. Annotations for proteins involved in the senescence-associated nutrient remobilization and in the following buildup of resources required for seed germination have been found for selected putative candidate genes. These include genes coding for storage proteins, cysteine proteases, cellulose-synthase, alpha-amylase, transcriptional regulators, and F-box and RWP-RK family proteins. Our findings highlight promising genomic regions for targeted breeding programs aimed at improving grain yield and protein content.

Genome wide association in Spanish bread wheat landraces identifies six key genomic regions that constitute potential targets for improving grain yield related traits

KEY MESSAGE

Association mapping conducted in 189 Spanish bread wheat landraces revealed six key genomic regions that constitute stable QTLs for yield and include 15 candidate genes. Genetically diverse landraces provide an ideal population to conduct association analysis. In this study, association mapping was conducted in a collection of 189 Spanish bread wheat landraces whose genomic diversity had been previously assessed. These genomic data were combined with characterization for yield-related traits, including grain size and shape, and phenological traits screened across five seasons. The association analysis revealed a total of 881 significant marker trait associations, involving 434 markers across the genome, that could be grouped in 366 QTLs based on linkage disequilibrium. After accounting for days to heading, we defined 33 high density QTL genomic regions associated to at least four traits. Considering the importance of detecting stable QTLs, 6 regions associated to several grain traits and thousand kernel weight in at least three environments were selected as the most promising ones to harbour targets for breeding. To dissect the genetic cause of the observed associations, we studied the function and in silico expression of the 413 genes located inside these six regions. This identified 15 candidate genes that provide a starting point for future analysis aimed at the identification and validation of wheat yield related genes.

Genome-Wide Association Study for Grain Protein, Thousand Kernel Weight, and Normalized Difference Vegetation Index in Bread Wheat (Triticum aestivum L.)

Genomic regions governing grain protein content (GPC), 1000 kernel weight (TKW), and normalized difference vegetation index (NDVI) were studied in a set of 280 bread wheat genotypes. The genome-wide association (GWAS) panel was genotyped using a 35K Axiom array and phenotyped in three environments. A total of 26 marker-trait associations (MTAs) were detected on 18 chromosomes covering the A, B, and D subgenomes of bread wheat. The GPC showed the maximum MTAs (16), followed by NDVI (6), and TKW (4). A maximum of 10 MTAs was located on the B subgenome, whereas, 8 MTAs each were mapped on the A and D subgenomes. In silico analysis suggest that the SNPs were located on important putative candidate genes such as NAC domain superfamily, zinc finger RING-H2-type, aspartic peptidase domain, folylpolyglutamate synthase, serine/threonine-protein kinase LRK10, pentatricopeptide repeat, protein kinase-like domain superfamily, cytochrome P450, and expansin. These candidate genes were found to have different roles including regulation of stress tolerance, nutrient remobilization, protein accumulation, nitrogen utilization, photosynthesis, grain filling, mitochondrial function, and kernel development. The effects of newly identified MTAs will be validated in different genetic backgrounds for further utilization in marker-aided breeding.

Uncovering the genetic basis for quality traits in the Mediterranean old wheat germplasm and phenotypic and genomic prediction assessment by cross-validation test

The release of new wheat varieties is based on two main characteristics, grain yield and quality, to meet the consumer's demand. Identifying the genetic architecture for yield and key quality traits has wide attention for genetic improvement to meet the global requirement. In this sense, the use of landraces represents an impressive source of natural allelic variation. In this study, a genome-wide association analysis (GWAS) with PCA and kinship matrix was performed to detect QTLs in bread wheat for fifteen quality and agronomic traits using 170 diverse landraces from 24 Mediterranean countries in two years of field trials. A total of 53 QTL hotspots containing 165 significant marker-trait associations (MTAs) were located across the genome for quality and agronomical traits except for chromosome 2D. The major specific QTL hotspots for quality traits were QTL_3B.3 (13 MTAs with a mean PVE of 8.2%) and QTL_4A.3 (15 MTAs, mean PVE of 11.0%), and for yield-related traits were QTL_2B.1 (8 MTAs, mean PVE of 7.4%) and QTL_4B.2 (5 MTAs, mean PVE of 10.0%). A search for candidate genes (CG) identified 807 gene models within the QTL hotspots. Ten of these CGs were expressed specifically in grain supporting the role of identified QTLs in Landraces, associated to bread wheat quality traits and grain formation. A cross-validation approach within the collection was performed to calculate the accuracies of genomic prediction for quality and agronomical traits, ranging from -0.03 to 0.64 for quality and 0.46 to 0.65 for agronomic traits. In addition, five prediction equations using the phenotypic data were developed to predict bread loaf volume in landraces. The prediction ability varied from 0.67 to 0.82 depending on the complexity of the traits considered to predict loaf volume.

Discovering Loci for Breeding Prospective and Phenology in Wheat Mediterranean Landraces by Environmental and eigenGWAS

Knowledge of the genetic basis of traits controlling phenology, differentiation patterns, and environmental adaptation is essential to develop new cultivars under climate change conditions. Landrace collections are an appropriate platform to study the hidden variation caused by crop breeding. The use of genome-wide association analysis for phenology, climatic data and differentiation among Mediterranean landraces led to the identification of 651 marker-trait associations that could be grouped in 46 QTL hotspots. A candidate gene analysis using the annotation of the genome sequence of the wheat cultivar 'Chinese Spring' detected 1097 gene models within 33 selected QTL hotspots. From all the gene models, 42 were shown to be differentially expressed (upregulated) under abiotic stress conditions, and 9 were selected based on their levels of expression. Different gene families previously reported for their involvement in different stress responses were found (protein kinases, ras-like GTP binding proteins and ethylene-responsive transcription factors). Finally, the synteny analysis in the QTL hotspots regions among the genomes of wheat and other cereal species identified 23, 21 and 7 ortho-QTLs for Brachypodium, rice and maize, respectively, confirming the importance of these loci.

High-Density Linkage Mapping of Agronomic Trait QTLs in Wheat under Water Deficit Condition using Genotyping by Sequencing (GBS)

Improvement of grain yield is the ultimate goal for wheat breeding under water-limited environments. In the present study, a high-density linkage map was developed by using genotyping-by-sequencing (GBS) of a recombinant inbred line (RIL) population derived from the cross between Iranian landrace #49 and cultivar Yecora Rojo. The population was evaluated in three locations in Iran during two years under irrigated and water deficit conditions for the agronomic traits grain yield (GY), plant height (PH), spike number per square meter (SM), 1000 kernel weight (TKW), grain number per spike (GNS), spike length (SL), biomass (BIO) and harvest index (HI). A linkage map was constructed using 5831 SNPs assigned to 21 chromosomes, spanning 3642.14 cM of the hexaploid wheat genome with an average marker density of 0.62 (markers/cM). In total, 85 QTLs were identified on 19 chromosomes (all except 5D and 6D) explaining 6.06–19.25% of the traits phenotypic variance. We could identify 20 novel QTLs explaining 8.87–19.18% of phenotypic variance on chromosomes 1A, 1B, 1D, 2B, 3A, 3B, 6A, 6B and 7A. For 35 out of 85 mapped QTLs functionally annotated genes were identified which could be related to a potential role in drought stress.