Multi-locus genome-wide association studies reveal the genetic architecture of Fusarium head blight resistance in durum wheat

Abstract

Durum wheat is more susceptible to Fusarium head blight (FHB) than other types or classes of wheat. The disease is one of the most devastating in wheat; it reduces yield and end-use quality and contaminates the grain with fungal mycotoxins such as deoxynivalenol (DON). A panel of 265 Canadian and European durum wheat cultivars, as well as breeding and experimental lines, were tested in artificially inoculated field environments (2019-2022, inclusive) and two greenhouse trials (2019 and 2020). The trials were assessed for FHB severity and incidence, visual rating index, Fusarium-damaged kernels, DON accumulation, anthesis or heading date, maturity date, and plant height. In addition, yellow pigment and protein content were analyzed for the 2020 field season. To capture loci underlying FHB resistance and related traits, GWAS was performed using single-locus and several multi-locus models, employing 13,504 SNPs. Thirty-one QTL significantly associated with one or more FHB-related traits were identified, of which nine were consistent across environments and associated with multiple FHB-related traits. Although many of the QTL were identified in regions previously reported to affect FHB, the QTL QFhb-3B.2, associated with FHB severity, incidence, and DON accumulation, appears to be novel. We developed KASP markers for six FHB-associated QTL that were consistently detected across multiple environments and validated them on the Global Durum Panel (GDP). Analysis of allelic diversity and the frequencies of these revealed that the lines in the GDP harbor between zero and six resistance alleles. This study provides a comprehensive assessment of the genetic basis of FHB resistance and DON accumulation in durum wheat. Accessions with multiple favorable alleles were identified and will be useful genetic resources to improve FHB resistance in durum breeding programs through marker-assisted recurrent selection and gene stacking.

Keywords: DON; FHB resistance; GDP; GWAS; KASP markers; durum wheat; multi-locus.

Figure 1

Phenotypic distribution of mean values for FHB incidence (INC; %), FHB severity (SEV; %), Fusarium-damaged kernels (FDK; %), deoxynivalenol concentration (DON; ppm), protein content (PRO; %), yellow pigment (YP, µg g⁻¹), visual rating index (VRI; %), INC-SEV-DON index (ISD), plant height (PH, cm), heading date (HD, days), anthesis date (AD, days), and maturity date (MAT, days).

Figure 2

Correlation heatmap of FHB and agro-morphological trait BLUP values in the GWAS panel. INC, FHB incidence; SEV, FHB severity; FDK, Fusarium-damaged kernels; VRI, visual rating index; DON, deoxynivalenol; ISD, INC-SEV-DON index; AD, days to anthesis; HD, days to heading; MAT, days to maturity; PH, plant height; PRO, grain protein content; YP, yellow pigment.

Figure 3

Population structure of the 265 durum wheat lines based on 13,504 SNP markers. (A) Cross-validation error showing the likely appropriate number of populations K to be 5. (B) Population structure based on genetic admixture for K = 5, where each bar represents a single line and the colored segments within each bar reflect the proportional contributions of each subgroup to that line. (C) Principal component analysis (PCA) plot of the first two principal components (PCs). Percentages in brackets indicate the variance explained by the PCs. (D) Topological view of the neighbor-joining phylogenetic tree.

Figure 4

Manhattan plot reveals QTL for FHB visual rating index (VRI) using the MLM model (upper panel, A) and phenotypic variations at large effect QTL QFhb-5A (IAAV3365) for VRI (lower panel, B) based on overall mean.

Figure 5

Allelic frequencies in the GDP panel and the geographic distribution of the 10 most frequent haplotypes. (A) Allelic frequencies of the six representative SNPs from the most consistent QTL in cultivars (CLT), domesticated emmer (DOM), landraces (LND), and wild emmer (WLD); the y-axis shows the percentage, and the labels show the absolute frequencies. (B) Haplotype frequencies based on the six loci. (C) Geographic distribution of the top 10 most frequent haplotypes. The haplotype pie chart circle size is proportional to the number of accessions from the corresponding country.

Figure 6

Pyramiding effects of QFhb-2A.3, QFhb-3B.2, QFhb-6A, QFhb-6B.1, and QFhb-7B.2 provide resistance to FHB SEV, INC, and DON accumulation. SEV, severity (%); INC, incidence (%); DON, deoxynivalenol (ppm).