A diverse panel of 755 bread wheat accessions harbors untapped genetic diversity in landraces and reveals novel genetic regions conferring powdery mildew resistance

Abstract

A bread wheat panel reveals rich genetic diversity in Turkish, Pakistani and Iranian landraces and novel resistance loci to diverse powdery mildew isolates via subsetting approaches in association studies. Wheat breeding for disease resistance relies on the availability and use of diverse genetic resources. More than 800,000 wheat accessions are globally conserved in gene banks, but they are mostly uncharacterized for the presence of resistance genes and their potential for agriculture. Based on the selective reduction of previously assembled collections for allele mining for disease resistance, we assembled a trait-customized panel of 755 geographically diverse bread wheat accessions with a focus on landraces, called the LandracePLUS panel. Population structure analysis of this panel based on the TaBW35K SNP array revealed an increased genetic diversity compared to 632 landraces genotyped in an earlier study and 17 high-quality sequenced wheat accessions. The additional genetic diversity found here mostly originated from Turkish, Iranian and Pakistani landraces. We characterized the LandracePLUS panel for resistance to ten diverse isolates of the fungal pathogen powdery mildew. Performing genome-wide association studies and dividing the panel further by a targeted subsetting approach for accessions of distinct geographical origin, we detected several known and already cloned genes, including the Pm2a gene. In addition, we identified 22 putatively novel powdery mildew resistance loci that represent useful sources for resistance breeding and for research on the mildew-wheat pathosystem. Our study shows the value of assembling trait-customized collections and utilizing a diverse range of pathogen races to detect novel loci. It further highlights the importance of integrating landraces of different geographical origins into future diversity studies.

Fig. 1

LandracePLUS panel diversity and powdery mildew symptoms: a World map with semitransparent purple dots representing the origin of each of the 744 of the 755 wheat accessions of the LandracePLUS panel with known origin, b number of accessions from the LandracePLUS panel per country of origin, c selection of diverse wheat spikes from the LandracePLUS panel. Scale bar, 5 cm, and d leaf of wheat cultivar “Kanzler” with powdery mildew symptoms. Scale bar, 1 cm (Color figure online)

Fig. 2

Genetic diversity and kinship analysis of the LandracePLUS panel: a PCA from 27,337 SNPs including high-quality sequenced wheat accessions labeled in the figure. PC1 = 8.4%, PC2 = 5.1%. Colors refer to the geographical origin of accessions and are indicated in b. b Dendrogram of a hierarchical clustering analysis from 27,337 SNPs including high-quality sequenced accessions indicated with stars. Colors represent the geographical origin of wheat accessions. Circled numbers on nodes refer to groups 1 to 4 when dividing into four clusters. c Alignment of Admixture plot, kinship matrix and dendrogram for the 755 wheat accessions based on 29,965 SNPs. The Admixture plot shows K = 5, where colors represent ancestral populations. In the kinship matrix, more saturated shades of blue indicate stronger relatedness. Dashed lines separate the four groups based on hierarchical clustering (Color figure online)

Fig. 3

Virulence patterns of powdery mildew isolates and resistance distribution in the LandracePLUS panel. Virulence pattern of the ten powdery mildew isolates on differential lines with known Pm genes are shown as a heatmap on the top. On the bottom, the virulence pattern of the ten powdery mildew isolates on the LandracePLUS panel is depicted via violinplots and boxplots, highlighting median, 25 and 75 percentiles. The defined resistance threshold of 20 is indicated by a dashed line. Differential lines for the 27 Pm genes are listed in Table S2

Fig. 4

Analysis of the region associated with powdery mildew resistance on chromosome 5D. Manhattan plot for GWAS with all wheat accessions infected with isolate ISR_106 showing a all 21 chromosomes, b chromosome 5D and c the region of 40 to 50 Mb on chromosome 5D. The locus of the Chinese Spring version of Pm2 with a partially deleted gene (Sánchez-Martín et al. 2016) is indicated with a line. d Manhattan plot for GWAS with ISR_106 when adding a covariate for Pm2 presence in wheat accessions. e Manhattan plot for the Pm2 virulent isolate CHE_98230. Solid lines represent the threshold for false discovery rate, and dashed lines for Bonferroni correction. f Map with semitransparent brown dots depicting the origin of landraces that contain Pm2. Countries of origin are abbreviated with the three-letter country code of ISO 3166

Fig. 5

Manhattan plots of novel resistance-associated regions: a all phenotyped LandracePLUS panel accessions, b all phenotyped LandracePLUS panel accessions using a Pm2 covariate, c subset of accessions from Pakistan and Iran, and d subset of accessions from Turkey with CHE_97266 representing the peak occurring for multiple isolates. Isolates and chromosomes are depicted in the top-left corner. Solid lines represent the threshold for false discovery rate and dashed lines for Bonferroni correction