Genetic architecture of fusarium head blight disease resistance and associated traits in Nordic spring wheat

Abstract

This study identified a significant number of QTL that are associated with FHB disease resistance in NMBU spring wheat panel by conducting genome-wide association study. Fusarium head blight (FHB) is a widely known devastating disease of wheat caused by Fusarium graminearum and other Fusarium species. FHB resistance is quantitative, highly complex and divided into several resistance types. Quantitative trait loci (QTL) that are effective against several of the resistance types give valuable contributions to resistance breeding. A spring wheat panel of 300 cultivars and breeding lines of Nordic and exotic origins was tested in artificially inoculated field trials and subjected to visual FHB assessment in the years 2013-2015, 2019 and 2020. Deoxynivalenol (DON) content was measured on harvested grain samples, and anther extrusion (AE) was assessed in separate trials. Principal component analysis based on 35 and 25 K SNP arrays revealed the existence of two subgroups, dividing the panel into European and exotic lines. We employed a genome-wide association study to detect QTL associated with FHB traits and identify marker-trait associations that consistently influenced FHB resistance. A total of thirteen QTL were identified showing consistent effects across FHB resistance traits and environments. Haplotype analysis revealed a highly significant QTL on 7A, Qfhb.nmbu.7A.2, which was further validated on an independent set of breeding lines. Breeder-friendly KASP markers were developed for this QTL that can be used in marker-assisted selection. The lines in the wheat panel harbored from zero to five resistance alleles, and allele stacking showed that resistance can be significantly increased by combining several of these resistance alleles. This information enhances breeders´ possibilities for genomic prediction and to breed cultivars with improved FHB resistance.

Fig. 1

Histogram distributions based on the mean phenotypic data over all testing environments for AE, DH, FHB, PH and tDON for the NMBU spring wheat panel

Fig. 2

Principal component analysis based on the 25 K data showing the population structure of NMBU spring wheat panel, which is divided mainly into two groups—European and others (lines from outside the Europe such as from CIMMYT, China and USA)

Fig. 3

QQ plots of a AEmean, b DONcPHDHmean and c FHBcPHDHmean for different GWAS models—GLM, CMLM, ECMLM and FarmCPU, using the 25 K genotype data

Fig. 4

Boxplot showing the haplotype effect of QTL on chromosome 7A based on the a mean AE, b mean corrected DON content, c mean corrected FHB disease severity of complete panel, and d mean AE, e mean corrected DON content, f mean corrected FHB disease severity of European panel. Wilcoxon method is used for pair-wise comparisons (***P < 0.0001, **P < 0.001, *P < 0.05, ns > 0.05). Nonsignificant comparisons are not shown in this figure

Fig. 5

Boxplot showing the haplotype effect of QTL on chromosome 7A based on the a AE, b corrected DON content, c corrected FHB disease severity of validation panel at Vollebekk in 2020 and d AE, e corrected FHB disease severity of validation panel at Tulln in 2021. Wilcoxon method is used for pair-wise comparisons (***P < 0.0001, **P < 0.001, *P < 0.05, ns > 0.05). Nonsignificant comparisons are not shown in this figure

Fig. 6

Boxplot showing the effect of number of stacked resistant alleles based on the a mean AE, b mean corrected DON content, c mean corrected FHB disease severity of complete panel, and d mean AE, e mean corrected DON content, f mean corrected FHB disease severity of European panel. Wilcoxon method is used for pair-wise comparisons (***P < 0.0001, **P < 0.001, *P < 0.05, ns > 0.05). Nonsignificant comparisons are not shown in this figure