Separation of the effects of two reduced height (Rht) genes and genomic background to select for less Fusarium head blight of short-strawed winter wheat (Triticum aestivum L.) varieties

Abstract

FHB resistance shared pleiotropic loci with plant height and anther retention. Genomic prediction allows to select for genomic background reducing FHB susceptibility in the presence of the dwarfing allele Rht-D1b. With the high interest for semi-dwarf cultivars in wheat, finding locally adapted resistance sources against Fusarium head blight (FHB) and FHB-neutral reduced height (Rht) genes is of utmost relevance. In this study, 401 genotypes of European origin without/with dwarfing alleles of Rht-D1 and/or Rht24 were analysed across five environments on FHB severity and the morphological traits such as plant height (PH), anther retention (AR), number of spikelets per ear, ear length and ear density. Data were analysed by combined correlation and path analyses, association mapping and coupling single- and multi-trait genome-wide association studies (ST-GWAS and MT-GWAS, respectively) and genomic prediction (GP). All FHB data were corrected for flowering date or heading stage. High genotypic correlation (r_g = 0.74) and direct path effect (0.57) were detected between FHB severity and anther retention (AR). Moderate correlation (r_g = - 0.55) was found between FHB severity and plant height (PH) with a high indirect path via AR (- 0.31). Indirect selection for FHB resistance should concentrate on AR and PH. ST-GWAS identified 25 quantitative trait loci (QTL) for FHB severity, PH and AR, while MT-GWAS detected six QTL across chromosomes 2A, 4D, 5A, 6B and 7B conveying pleiotropic effects on the traits. Rht-D1b was associated with high AR and FHB susceptibility. Our study identified a promising positively acting pleiotropic QTL on chromosome 7B which can be utilized to improve FHB resistance while reducing PH and AR. Rht-D1b genotypes having a high resistance genomic background exhibited lower FHB severity and AR. The use of GP for estimating the genomic background was more effective than selection of GWAS-detected markers. We demonstrated that GP has a great potential and should be exploited by selecting for semi-dwarf winter wheat genotypes with higher FHB resistance due to their genomic background.

Fig. 1

Representation of the nature of the interactions among traits. a = path diagram illustrating direct and indirect effects; b = simultaneous structured equations matrix showing relationships between correlations and path coefficients. Double-arrowed lines indicate mutual association as measured by genotypic correlations (r_ij), and single-arrowed lines represent direct path effects (p_ij)

Fig. 2

Pearson’s product-moment correlation analysis among traits *, **, ***significant at p  < 0.05, 0.01 and 0.001, respectively. PH = plant height, FHB = FHB severity, AR = anther retention, EL = ear length, NS = number of spikelets per ear, ED = ear density.

Fig. 3

Manhattan plots highlighting significant marker–trait associations (MTAs) for single-trait genome-wide association studies (ST-GWAS): a = ST-GWAS₁ including all markers and b = ST-GWAS₂ without markers linked to plant height on chromosomes 4D (Rht-D1) and 6A (Rht24). The blue dotted line corresponds to an exploratory threshold of − Log 10(p) = 6, while the red plain line represents the Bonferroni-corrected threshold cut-off of alpha = 0.01 (color figure online)

Fig. 4

Boxplots showing variation of traits among Rht groups. (a) = plant height, (b) = FHB severity, and (c) = anther retention. n = number of genotypes, Min = minimum, Max = maximum. Boxplots with the same superscripts are statistically not significant at p < 0.05

Fig. 5

Comparison of prediction accuracies of genomic prediction (GP) and marker-assisted selection (MAS)

Fig. 6

Contribution of Rht genes and genomic background (GB) markers to plant height, FHB severity and anther retention. a = total genetic variation explained by groups of markers, and b = Genetic variation explained by individual markers

Fig. 7

Comparison of additive effects of Rht alleles and reduction effects of alleles by genomic background (GB) markers on: a = plant height, b = FHB severity and c = anther retention

Fig. 8

Scatter plot showing the strength of relationship between stacking of additive effects (SAE) of genomic background (GB) markers from single-trait genome-wide association study (ST-GWAS) and genomic estimated breeding values (GEBV) from genomic prediction (GP). SAE was estimated based on GB markers with p_G ≥ 5%. ***significant at p < 0.001