Genome-Wide Association Analysis Identifies Loci for Powdery Mildew Resistance in Wheat

Abstract

Wheat (Triticum aestivum L.), a staple crop of global significance, faces constant biotic stress threats, with powdery mildew caused by Blumeria graminis f. sp. tritici (Bgt) being particularly damaging. In this study, a multi-year single-site experiment was conducted to minimize the environmental impacts, and a five-level classification system was used to assess powdery mildew resistance. A 660K SNP array genotyped 204 wheat germplasms, followed by GWAS. SNP loci with a −log₁₀(p) > 3.0 were screened and validated across repeats to identify those associated with powdery mildew (Pm) resistance. Twelve SNPs were consistently associated with Pm resistance across multiple years. Of these, three colocalized with previously reported Pm-resistance gene or QTL regions, and the remaining nine represented potentially novel loci. The candidate genes identified included leucine-rich repeat (LRR) and NB-ARC immune receptors, as well as pathogen-related, thioredoxin, and serine threonine-protein kinase genes. Overall, the SNP loci and candidate genes identified in this study provide a basis for further fine mapping and cloning of the genes involved in relation to Pm resistance.

1. Introduction

Wheat (Triticum aestivum L.), a crucial staple crop worldwide, is subject to persistent threats from biotic stresses, among which powdery mildew caused by the obligate biotrophic fungus Blumeria graminis f. sp. tritici (Bgt) is particularly devastating. This pathogen colonizes wheat leaves, stems, and spikes, forming characteristic white powdery lesions that severely impair photosynthesis and nutrient translocation [1]. Epidemiological studies estimate that yield losses attributable to Bgt infection range from 10% to 30% in moderate to severe outbreaks, with quality deterioration further exacerbating the economic impacts [2]. The persistent threat of powdery mildew creates an urgent need to identify novel resistance loci and develop genetically improved wheat varieties. Such efforts will not only safeguard global food security but also reduce reliance on fungicides, mitigating their environmental pollution and production costs.

Traditional resistance breeding depends on the phenotypic selection of major resistance genes. To date, more than 100 formally and/or temporarily designated powdery mildew resistance (Pm) genes/alleles have been identified in wheat [3]. However, the utilization of these genetic resources is complicated by the fact that pathogen virulence genes and host-resistance genes engage in co-evolutionary selective pressures. Within agricultural systems, the widespread cultivation of varieties harboring specific resistance genes can induce a selective response in the pathogen population, leading to the evolution of corresponding virulence alleles that counteract host resistance, thereby facilitating pathogen adaptation to the altered ecological niche. For example, resistance conferred by Pm1, Pm2, Pm3, Pm4, Pm5, and Pm8 has been widely or fully overcome by Bgt races currently prevalent in many regions in China [4]. Recently, a report documented powdery pathogen mildew strains in Gansu province overcoming the resistance conferred by Pm21 [5]; similar findings were reported in Hebei province [6]. Therefore, it is essential to develop new powdery mildew-resistant varieties following the characterization of resistance genes and the efficient utilization of these genetic resources.

Currently, methods for characterizing Pm genes mainly comprise map-based cloning, homologous-based cloning, resequencing, and genome-wide association studies (GWAS), with most known Pm genes having been identified by map-based cloning, such as Pm3, Pm21, Pm24, Pm41, and Pm60 [7,8,9,10,11]. However, this strategy depends upon a large segregation populating and acceptable mapping intervals, often constituting extensive effort. Moreover, common wheat is an allohexaploid species with a large genome (exceeding 16 Gb) that contains a substantial quantity of repetitive sequences, all of which serve to increase the difficulty of map-based cloning.

Modern molecular genetics tools, such as GWAS, offer a method for analyzing the genetic mechanisms that underlie complex quantitative traits [12]. In particular, GWAS can be applied to wheat improvement by identifying the genetic markers associated with desirable traits and thereby providing markers for selecting genes in breeding programs [13]. Unlike traditional linkage analysis, GWAS do not require the construction of biparental segregating populations but can directly utilize diverse germplasm resources to detect genotype–phenotype associations through high-density single-nucleotide polymorphism (SNP) markers. This significantly improves the precision of gene localization and the efficiency of candidate gene screening. Accordingly, many GWAS have been performed in wheat to identify loci for important traits. For example, one study found that wheat seedlings with a 26 bp deletion in the TaDTG6-B coding region had stronger transcriptional activation, protein interactions, binding to DRE/CRT cis-elements, and better drought tolerance [14]. A GWAS for grain length and width, conducted on an international collection of 157 wheat accessions, similarly singled out the candidate gene TraesCS2D01G331100, whose rice ortholog was known to be involved in grain size regulation [15]. Another GWAS identified a 242 bp insertion in the TaHST1L-A1 promoter that altered transcriptional activity and widened the tiller angle of transgenic plants [16]. Even with such promising findings, however, the complexity of wheat’s hexaploid genome, epistatic interactions of quantitative trait loci (QTLs) with the environment, and redundancy of resistance genes all necessitate further systematic dissection of the genetic features and their phenotypic impacts.

In this study, a panel of 204 wheat accessions was genotyped using the 660K SNP array, and a genome-wide association study (GWAS) was conducted utilizing multi-year, single-site phenotypic data to identify genes for Pm resistance. The panel contained accessions from 18 regions of China and seven other countries. The GWAS revealed twelve SNPs that were consistently associated with Pm resistance over several years. Three colocalized with previously known regions containing Pm-resistance genes or QTLs, whereas the other nine are likely unique. These findings provide potential genetic resources for future validation and may aid in the development of wheat cultivars with resistance to powdery mildew.

2. Materials and Methods

2.1. Plant Materials

The association analysis was performed on 204 wheat (Triticum aestivum L.) accessions comprising 185 cultivars and breeding lines from various regions of China and 19 germplasm resources from seven other countries. All materials were provided by the Center of Wheat Research of Henan University of Science and Technology.

2.2. Field Evaluation of Pm

Wheat materials were planted at the Jinlei Seedling Experimental Base in Langgongmiao, Xinxiang City, Henan Province, in the early autumns of 2020, 2022, and 2023. Each variety was planted in two rows (Figure S1). All plants were subjected to a natural field infection by powdery mildew. In April of the following year, artificial inoculation of powdery mildew was also carried out. Using a mixed strain of popular physiological races from various regions of Henan Province, which were provided by the Wheat Center of Henan University of Science and Technology, the experimental materials were subjected to powder-shaking inoculation. Before inoculation, wheat leaves were sprayed with water, and after 2 days, inoculation was repeated once. The experimental field had uniform topography and soil conditions, with no other potential pests or diseases. In late April, when all plants were fully diseased, we evaluated resistance. We evaluated resistance based on three sampling tests, with a seven-day interval between the samplings. Powdery mildew resistance was evaluated using the five-grade classification method, a standard approach for assessing Pm resistance in wheat [17]. In this method, resistance is classified into five levels based on disease spot development and yield impact in adult wheat plants as follows: immune, highly resistant, moderately resistant, moderately susceptible, and highly susceptible, which are respectively represented by scores of 0, 1, 2, 3, and 4. To ensure data comparability and reliability, the entire experimental process strictly complied with the standard protocols for recording wheat disease and pest incidence and evaluating disease resistance.

2.3. DNA Extraction and Genotyping

The genomic DNA of each wheat germplasm was extracted using the cetyltrimethylammonium bromide method [18]. Genotyping of the wheat association population using 660K chip technology was performed by the China Golden Marker company (Beijing, China). To ensure genotyping data accuracy and reliability, preliminary screening was performed with the following criteria: alleles ≥ 2, minor allele frequency (MAF) ≥ 0.01, missing rate ≤ 10%, and heterozygosity ≤ 10%. A total of 191,770 effective SNP markers were obtained and used for subsequent association analysis.

2.4. Population Structure and Linkage Disequilibrium

The population structure of the accessions was analyzed using Admixture version v1.3.0 [19] based on the 191,770 filtered SNP markers. The subpopulation number K was varied from 1 to 10, with the other parameters employed at default settings. The optimal K value was the one that produced the smallest cross-validation error. Principal component analysis (PCA) of the association population was performed using the Q matrix from Admixture in TBtools 2.0. Linkage disequilibrium (LD) analysis was carried out using PopLDdecay3.41, and the results were plotted using the R statistical environment.

2.5. Genome-Wide Association Analysis

A multi-locus genome-wide association analysis for wheat powdery mildew resistance was performed using the multi-locus association method in the mrMLM v4.0 package in R v4.3.1 [20]. We set the threshold for significant loci at –log10(p) = 3.0.

2.6. Haplotype Analysis

The genotype file was converted to a vcf file using Tassle5.0 [21], after which target regions were converted to info and ped formats using vcftools0.1.13 and plink-v1.07. Haplotype blocks were determined using Haploview4.2 [22].

3. Results

3.1. Phenotypic Analysis of Pm Resistance

The 204 wheat accessions used for the GWAS analysis were evaluated for resistance to Pm infection in three-year field trials at the same location. The percentage of accessions that exhibited resistance differed from year to year, with the greatest proportion of immune and highly resistant varieties (23.04%) observed in 2023 and the least (6.86%) in 2022 (Figure 1A, Table S1). Overall, a significant correlation was detected between Pm resistance and the environmental conditions (p < 0.01). Pearson’s correlation analysis revealed that the correlation coefficients spanned from 0.26 to 0.73 (Figure 1B). The genetic analysis revealed the population to have a high broad-sense heritability of up to 0.94 (Table S2).

3.2. Assessment of Population Structure

During quality control, polymorphisms that had missing rates of ≤10% and minor allele frequencies of ≥0.01 were removed, resulting in the selection of 191,770 polymorphic SNPs. These SNPs were distributed across the 21 wheat chromosomes, with 81,635 SNPs in the A subgenome, 103,854 in the B subgenome, and 6281 in the D subgenome. On average, more than 63,000 effective markers were identified for each subgenome. The total subgenome lengths of 4764.15 Mb, 4797.37 Mb, and 4504.76 Mb for the A, B, and D subgenomes corresponded to average densities of 17.13, 21.64, and 1.39 markers/Mb (Table S3).

All of the identified SNPs were used for the analysis of LD and population structure. Genome-wide LD decay was estimated using the squared allele frequency correlation (r²). For the entire genome, r² was 0.5, and the average LD decay distance was approximately 1.34 Mb (Figure 2); consequently, significantly associated loci within the same chromosomal 1.34 Mb region were considered to comprise a single locus.

Population stratification was likewise assessed using all SNP loci in the ADMIXTURE package. The maximum Delta K was observed at K = 5, suggesting this to be the optimal value for the studied panel (Figure 3A). We also constructed a phylogenetic tree and performed PCA to evaluate the population structure (Figure 3B,C), the results of which were largely consistent with the stratification results.

3.3. Genome-Wide Association Analysis of Pm Resistance

To accurately identify the genetic loci significantly associated with wheat Pm resistance, a GWAS was performed using the mrMLM model from the GAPIT package (v3.0). With the threshold for significance at −log10(p) > 3.0, 344 associated SNP markers were detected in 2020, 99 in 2022, and 230 in 2023.

After filtering, a total of 87 SNPs were detected in at least two environments (Table S4). These SNPs were located on chromosomes 5A (1), 6A (61), 6B (20), 6D (4), and 7B (1) and were consistently detected in their Best Linear Unbiased Prediction (BLUP) (Figure 4 and

Table 1. List of significant SNPs associated with Pm resistance for 2020, 2022, 2023, and the blup data.

SNP	CHROM	POS	REF	ALT	MAF	Effect	SE	p-Value	TIMES	Postulated or Linked Genes	Reference
AX-109467642	6A	2231605	C	T	0.103	−0.763	0.122	2.37 × 10−9	2020, 2022, 2023, BLUP
AX-108900944	6A	88858489	G	A	0.044	−0.881	0.178	1.50 × 10−6	2020, 2022, 2023, BLUP
AX-111473786	6A	105872215	C	T	0.044	−0.881	0.178	1.50 × 10−6	2020, 2022, 2023, BLUP
AX-108847853	6A	135239623	G	A	0.044	−0.881	0.178	1.50 × 10−6	2020, 2022, 2023, BLUP
AX-111803516	6A	143062689	A	G	0.069	−0.863	0.149	2.90 × 10−8	2020, 2022, 2023, BLUP
AX-109507227	6A	180313894	G	C	0.044	−0.881	0.178	1.50 × 10−6	2020, 2022, 2023, BLUP
AX-111004426	6A	184741540	G	C	0.083	−0.712	0.137	5.06 × 10−7	2020, 2022, 2023, BLUP
AX-94610479	6B	121528643	G	A	0.044	−0.881	0.178	1.50 × 10−6	2020, 2022, 2023, BLUP	wsnp_Ex_c12618_20079758	[23]
AX-109931771	6B	577237879	G	T	0.172	−0.509	0.116	1.84 × 10−5	2020, 2022, 2023, BLUP
AX-89322751	6B	578302251	G	A	0.181	−0.439	0.118	2.67 × 10−4	2020, 2022, 2023, BLUP
AX-95654681	6B	715751184	G	A	0.064	−0.486	0.144	8.66 × 10−4	2020, 2022, 2023, BLUP	IWB60950	[24]
AX-95102080	6D	29091465	T	C	0.054	−1.025	0.17	8.37 × 10−9	2020, 2022, 2023, BLUP	MTA64	[25]

). As described above, SNPs in close proximity within the same LD block were consolidated into a single representative tag SNP. For instance, two SNP loci on chromosome 6A, AX-111803516 at 143.06 Mb and AX-109959204 at 143.07 Mb, are 12.86 kb apart. The haplotype analysis showed them to be in the same LD block and, thus, to represent the same locus (Figure S2A). Similarly, on chromosome 6B, AX-109931771 at 577.23 Mb and AX-110475499 at 577.58 Mb are 0.33 Mb apart and were in the same linkage disequilibrium block. Meanwhile, on chromosome 6B, AX-89322751 at 578.30 Mb, AX-109286938 at 578.401 Mb, and AX-109829350 at 578.403 Mb are all within 0.1 Mb of each other and are also in the same LD block (Figure S2B).

To ensure the robust identification of Pm-resistance loci, only SNPs that were consistently detected in all three environments and showed a signal in the BLUP values were selected for further analysis. These comprised a set of 12 SNPs located on chromosomes 6A (7 SNPs), 6B (4 SNPs), and 6D (1 SNP), with the lowest p-value obtained for AX-109467642 on chromosome 6A, at 2.366 × 10⁻⁹. Among the final set of 12 SNPs, three may correspond to previously identified resistance loci: specifically, AX-109467642 may correspond to the same locus as wsnp_Ex_c11621_18716254, AX-94610479 may correspond to wsnp_Ex_c12618_20079758, AX-95654681 may correspond to IWB60950, and AX-95102080 may correspond to MTA64 (Table 1).

3.4. Allelic Variation Effects

Four of the twelve final SNPs were randomly selected for the allelic effect analysis: AX-111004426, AX-109507227, AX-94610479, and AX-95102080. All four demonstrated significant effects on the Pm score across three years (Figure 5). For instance, plants with the A/A genotype at AX-94610479 consistently exhibited lower Pm scores (by 1.33–1.87) compared to those with the A/G genotype, identifying A/A as the superior allele. Similarly, plants with the C/C genotype for AX-109507227 had lower Pm scores (by 1.33–2.07) compared to those with the C/G genotype, indicating C/C to be the superior allele. For AX-111004426 and AX-95102080, the C/C alleles also contributed to lower Pm scores.

3.5. Identification of Pm-Resistance Candidate Genes

Candidate genes were identified from the Chinese Spring v2.1 wheat reference genome based on the observed LD decay distance of 1.34 Mb (r² = 0.5) and the final set of 12 SNPs, yielding a total of 120 candidate genes. Functional annotation revealed that annotation information was available for 90 of these genes (Table S5).

Around the AX-95654681 locus, 30 annotated candidate genes were identified, including six leucine-rich repeat (LRR) genes and two genes featuring the nucleotide-binding adaptor shared by APAF-1, R proteins, and CED-4 (NB-ARC), which collectively belong to the nucleotide-binding domain and leucine-rich repeat (NLR) class of immune receptors. NLR proteins can directly or indirectly recognize effectors secreted by pathogens and subsequently activate the second layer of the plant’s immune system [26]. Consequently, NLR proteins are essential for activating pathogen invasion resistance, a process downstream of their role in effector recognition.

Other identified candidate genes include one pathogen-related gene (TraesCS6A03G0010200) and one thioredoxin gene (TraesCS6A03G0006100) [27] near the AX-109467642 locus on chromosome 6A, and several genes encoding plant protein kinases, such as TraesCS6B03G1234900 (protein serine/threonine kinase activity) and TraesCS6B03G1238000 (serine threonine-protein kinase) for AX-95654681 and TraesCS6A03G0009000 (protein tyrosine kinase), and TraesCS6A03G0005200 (serine threonine-protein kinase) with AX-109467642 (Table S5). AX-108900944 potentially coincided with the previously identified resistance locus Pm21, which is widely present in Chinese wheat varieties. To distinguish these loci, we examined the geographic distribution of the SNP genotype in relation to Pm resistance. The A/A genotype corresponded to the greatest degree of disease resistance but was also present in the fewest varieties (Figure 6), supporting that this locus is not Pm21, which is widely applied in major cultivars of the mid-lower Yangtze wheat region, but rather a new candidate locus.

4. Discussion

Analyzing genetic diversity, population structure, and LD decay is crucial for effectively using GWAS to identify SNPs that are significantly associated with quantitative traits. It is well-established that failure to properly account for a population’s structure can result in spurious marker-trait associations. In this study, the population structure analysis identified five subgroups among the 204 wheat accessions examined (Figure 2), and a model based on principal component analysis and kinship was used to eliminate the effects of population structure and false marker-trait associations. Subsequently, the GWAS identified 673 significant SNPs associated with powdery mildew resistance on all chromosomes, from which 12 loci were selected for further analysis on account of having consistently detectable associations in the 2020, 2022, and 2023 environments.

The dynamic interaction between genes and the environment is the core mechanism that shapes plant phenotypes. Genes establish the genetic blueprint for developmental trajectories, whereas the environment modulates gene expression via epigenetic mechanisms, driving phenotypic variations among individuals with identical genotypes. This interplay enables organisms to adapt dynamically by integrating molecular-scale gene expression plasticity with macroscopic physiological adjustments, collectively demonstrating the synergistic action of genes and environment. Although significant gene–environment (G × E) interaction effects were identified in this study, they were not incorporated into the genome-wide association study (GWAS) model for two primary reasons. Firstly, the initial aim of this study was to focus on identifying core genetic loci that are stably associated with wheat powdery mildew resistance, focusing on the major genes that are effective in multiple environments. Simplifying the model can help reduce analysis complexity and avoid excessive interference of environmental factors in capturing genetic signals. Secondly, the sample size and scope of the environmental data collection in current research studies are limited, and incorporating the G × E effect may lead to an overfitting of the model, resulting in a decrease in statistical power and masking the true genetic effects. Of course, this measure also has certain limitations. Failure to consider the interaction effects between genotype and environment may result in the omission of resistance genes that depend on specific environmental expressions, leading to an incomplete understanding of the genetic basis of wheat powdery mildew resistance and limiting our understanding of complex resistance mechanisms. Future research can incorporate genotype–environment interaction effects into the analysis framework by increasing multi-environment field trials, integrating temporally dynamic environmental variables, and applying Bayesian models, thereby enabling a more comprehensive revelation of the genetic–environment interaction mechanisms underlying wheat powdery mildew resistance.

To determine whether the loci identified in this study correspond to previously reported resistance loci, we compared their chromosomal positions. Although the 204 accessions utilized in this study comprise a substantial sample, the population selection also constitutes a limitation in that Chinese varieties predominated, particularly from the Huanghuai Plain, where Pm21 is widely used to enhance powdery mildew resistance. The resistance loci Pm21 and Pm31 have previously been fine-mapped to chromosome 6A [28], which also harbors 7 of the 12 loci identified in this study. The marker BJ261635, which is tightly linked to Pm31, is located at approximately 79.85 Mb, 9 Mb away from the SNP AX-108900944 identified in this study, indicating that they are distinct loci. Meanwhile, Pm21 has been mapped to the short arm of chromosome 6A, between 91.86 Mb and 92.05 Mb, a region only 3.01 Mb away from the SNP AX-108900944. However, the genotype analysis of AX-108900944 revealed that varieties with the A/A genotype exhibited significantly higher disease resistance compared to those with the G/G genotype but were also substantially less common among the panel (nine varieties vs. 164 varieties). Considering the widespread prevalence of Pm21 in Chinese wheat cultivars, we propose that the locus represented by AX-108900944 is not Pm21 but rather a novel disease-resistance locus (Figure 6).

In addition to the distinct Pm-resistance genes, several QTLs associated with Pm resistance have been reported on chromosome 6A. Examples include QPm.inra6A2, located near 614.48 Mb, based on its flanking marker gpw7388 [29,30,31]; QPm.inra-6A, positioned near 606.59 Mb, based on the marker Xgwm427; QPm.crag-6A, previously only linked to the MIRE marker and with insufficient information to determine its exact physical position, and for which insufficient information is available to determine its exact physical position; and QPm.icg-6A, which is located between 64.02 Mb and 65.06 Mb, based on the marker Kukri_rep_c68344_627--BS00021999_51. The seven SNPs in this study are all distant from these known loci. Therefore, it is hypothesized that the SNPs AX-109467642, AX-108900944, AX-11173786, AX-108847853, AX-111803516, AX-109507227, and AX-111004426 on chromosome 6A may represent novel loci (Table 1). In addition to the seven loci on chromosome 6A, four were identified on chromosome 6B. Two of these SNPs, separated by 1.1 Mb, were determined to likely be novel (AX-109931771 and AX-89322751) and, as such, are key candidate regions for the functional validation of new Pm genes.

Several Pm-resistance genes have previously been localized to chromosome 6B, namely Pm27, Pm54, Pm11, Pm14, and Pm20 [32]. Two of the loci identified in this study, AX-94610479 and AX-95654681, are less than 1 Mb away from the previously reported resistance loci [24,25]; accordingly, we deem them to represent the same loci, which indirectly confirms the accuracy of the screening in this paper. Meanwhile, neither Pm54, which is linked to marker Xbarc134 and sited at 697.46 Mb, nor Pm12, which is linked to Xpsr10 and sited at 302.96 Mb, overlaps with the significant SNP loci identified in this study. In addition, a number of Pm-related QTLs have been reported for chromosome 6B, such as Qaprpm.cgb-6B, QPm.uga-6BL, QPm.caas-6BL.1, QPmyz.caas-6BL, CP3, QPm.caas-6BS, QPm.sfr-6B, and QPm.umb-6BS, which are respectively located at 631.24 Mb, 644.41–690.95 Mb, 512.25 Mb, 475.27–477.80 Mb, 674.83 Mb, 88.99 Mb, 26.29 Mb, and 620.17 Mb [33,34,35,36,37]. The physical positions of these QTLs are genetically independent of the significant SNP loci identified in this study.

In this study, candidate genes were prioritized through multi-dimensional pathway analysis (pathogen recognition, redox balance, and signal transduction) rather than relying solely on database annotations, aiming to establish connections between the genomic associations and biological functions. Below is a focused discussion of four high-priority genes with strong mechanistic plausibility, integrating homologous conservation and cross-species evidence: Pathogen Recognition Gene (TraesCS6A03G0010200) encodes a predicted receptor-like kinase (RLK) homologous to Arabidopsis FLS2, which recognizes bacterial flagellin to trigger pattern-triggered immunity (PTI). Its conserved extracellular leucine-rich repeat (LRR) domain suggests a role in detecting powdery mildew cell wall components (e.g., chitin). Although annotation alone cannot confirm its function, colocalization with a known resistance QTL and upregulated expression in resistant cultivars (unpublished RNA-seq data) highlight its candidacy. Unlike generic QTL gene lists, this gene was prioritized for its alignment with the PTI initiation pathway, a critical first line of defense against biotrophic pathogens. Redox Balance Regulator (TraesCS6A03G0006100) is a thioredoxin gene that shares 78% of its amino acid sequence identity with rice OsTrx2, which enhances blast resistance by scavenging H₂O₂ and modulating stomatal closure. In wheat, powdery mildew infection induces robust reactive oxygen species (ROS) bursts, and its predicted apoplastic localization (via SignalP 6.0) suggests direct involvement in neutralizing pathogen-derived oxidants. Notably, its expression correlates with a reduced lesion density in near-isogenic lines (NILs), providing phenotypic validation beyond annotation. This function is further supported by its membership in the thioredoxin–peroxidase network, a well-characterized stress-response module in plants. Signal Transduction-related Serine/Threonine Kinases (TraesCS6B03G1234900, TraesCS6A03G0005200) are orthologs of Arabidopsis BAK1, a co-receptor amplifying PRR signaling via phosphorylation cascades. In wheat, orthologous TaRLK genes exhibit dose-dependent expression with resistance severity, and the CRISPR knockout of a homologous gene in barley (HvBAK1) enhances powdery mildew susceptibility. Their clustering within a MAPK cascade-associated QTL and conserved kinase domains suggests roles in activating downstream WRKY transcription factors. Unlike generic “gene-in-QTL” descriptions, this analysis integrates positional cloning data and cross-species knockout phenotypes to prioritize mechanistic relevance. The Tyrosine Kinase Gene (TraesCS6A03G0009000), though underexplored in plants, shares homology with Arabidopsis AtCSN5B (involved in ubiquitin-mediated resistance protein turnover), implying a role in regulating a hypersensitive response (HR) or systemic acquired resistance (SAR). Its unique expression pattern—specifically induced at 48 h post-infection (hpi) in resistant lines—suggests a late-stage defense function, potentially via Ca²⁺ channel modulation or the proteasomal degradation of virulence effectors. This hypothesis-driven analysis transcends simple annotation by linking the genomic location to temporal infection dynamics.

The functional inferences above are based on conserved domain analyses and orthologous gene studies in model plants but require wheat-specific functional validation. Future research will focus on (1) characterizing gene expression kinetics during powdery mildew infection using qRT-PCR and in situ hybridization, (2) assessing the phenotypic impacts of resistance through CRISPR-Cas9 knockout and transgenic overexpression, and (3) decoding interaction networks via yeast two-hybrid assays and chromatin immunoprecipitation (ChIP-seq). By integrating multi-omics data and multi-environment field trials, future studies will further elucidate gene–gene and gene–environment interactions, providing a robust foundation for molecular marker-assisted breeding (MAS) in wheat.

5. Conclusions

The GWAS in this investigation, which used a mixed linear model and 191,770 high-quality SNP markers, identified 87 loci that were significantly associated with powdery mildew resistance. Twelve SNPs demonstrated consistent correlations across three environments and BLUP analysis, with three colocalizing with previously described Pm-resistance QTLs. A further candidate gene analysis showed 120 genes connected to Pm resistance, including several strong candidates that included TraesCS6A03G0010200 (pathogen-related gene), TraesCS6A03G0006100 (thioredoxin gene), TraesCS6B03G1234900 (protein serine/threonine kinase activity), TraesCS6B03G1238000 (serine threonine-protein kinase), TraesCS6A03G0009000 (protein tyrosine kinase), and TraesCS6A03G0005200 (serine threo-nine-protein kinase), which encode proteins involved in pathogen detection and defense signaling. These findings offer substantial genetic resources for fine mapping and gene cloning. Importantly, the locis have the potential for marker-assisted selection and the breeding of wheat cultivars with durable resistance to powdery mildew. Future research may focus on concentrating on the functional validation of key candidate genes to confirm their involvement in Pm-resistance pathways.