Mapping rust resistance in European winter wheat: many QTLs for yellow rust resistance, but only a few well characterized genes for stem rust resistance

Abstract

Stem rust resistance was mainly based on a few, already known resistance genes; for yellow rust resistance there was a combination of designated genes and minor QTLs. Yellow rust (YR) caused by Puccinia striiformis f. sp. tritici (Pst) and stem rust (SR) caused by Puccinia graminis f. sp. tritici (Pgt) are among the most damaging wheat diseases. Although, yellow rust has occurred regularly in Europe since the advent of the Warrior race in 2011, damaging stem rust epidemics are still unusual. We analyzed the resistance of seven segregating populations at the adult growth stage with the parents being selected for YR and SR resistances across three to six environments (location-year combinations) following inoculation with defined Pst and Pgt races. In total, 600 progenies were phenotyped and 563 were genotyped with a 25k SNP array. For SR resistance, three major resistance genes (Sr24, Sr31, Sr38/Yr17) were detected in different combinations. Additional QTLs provided much smaller effects except for a gene on chromosome 4B that explained much of the genetic variance. For YR resistance, ten loci with highly varying percentages of explained genetic variance (pG, 6-99%) were mapped. Our results imply that introgression of new SR resistances will be necessary for breeding future rust resistant cultivars, whereas YR resistance can be achieved by genomic selection of many of the detected QTLs.

Fig. 1

Correlation plot and histograms for yellow rust and stem rust responses in all populations based on best linear unbiased estimators (BLUEs) on logit scale. Blue line indicates BLUE of Parent 1, red lines of Parent 2. ***P < 0.001 alpha level (colour figure online)

Fig. 2

Manhattan plots of the phenotypic means for stem rust for the biparental populations (headers from top to bottom). In the Manhattan plot the − log₁₀ of the P-value is displayed for all markers on the chromosomes of each genome according to a consensus map calculated from separate linkage maps based on marker data of the biparental populations. Markers that could not (uniquely) be assigned to a chromosome are displayed on the chromosome “un” with arbitrary positions. Manhattan plots of the biparental populations were interpolated with a spline function (colour figure online)

Fig. 3

Manhattan plots of the phenotypic means for the trait yellow rust for the biparental populations (headers from top to bottom). In the Manhattan plot the − log₁₀ of the P-value is displayed for all markers of each genome according to a consensus map calculated from separate linkage maps based on marker data of the biparental populations. Markers that could not (uniquely) be assigned to a chromosome are displayed on the chromosome “un” with arbitrary positions. Manhattan plots of the biparental populations were interpolated with a spline function (colour figure online)

Fig. 4

Boxplots of genotypes grouped by alleles of markers linked with three stem rust (Sr) genes in the respective populations. Allele A indicates the allele of parent1 and allele B of parent2. Number of genotypes in the respective group are reported above boxes. Genotypes with missing marker alleles were not included

Fig. 5

Boxplots of genotypes grouped for individual and combined alleles of different genes/QTL found for yellow rust resistance in seven populations. The alleles and allele combinations of respective groups are reported on the x-axis. Allele A belongs to the allele of parent 1, allele B to parent 2 and H to heterozygous. As only single QTLs were found on separate chromosomes, chromosome names can be used to deduce respective markers used for clustering from Table 4. To keep reasonable group sizes, for some populations only two or three most significant QTLs were used for clustering. Numbers of genotypes in respective group are reported above the boxes

Fig. 6

Boxplots of genotypes grouped by alleles of marker IAAV8501 linked to Sr38/Yr17 on chromosome 2A for stem rust and yellow rust response analyzed as best linear unbiased estimators (BLUEs) for five populations (separated by dashed lines). Allele A indicates the allele of Parent 1, allele B of Parent 2. Numbers of genotypes in respective groups are reported above boxes. Genotypes with missing marker alleles were not included