Genome-wide association study of resistance to stripe rust (Puccinia striiformis f. sp. tritici) in Sichuan wheat

Abstract

Background: Stripe rust (also called yellow rust) is a common and serious fungal disease of wheat (Triticum aestivum L.) caused by Puccinia striiformis f. sp. tritici. The narrow genetic basis of modern wheat cultivars and rapid evolution of the rust pathogen have been responsible for periodic and devastating epidemics of wheat rust diseases. In this study, we conducted a genome-wide association study with 44,059 single nucleotide polymorphism markers to identify loci associated with resistance to stripe rust in 244 Sichuan wheat accessions, including 79 landraces and 165 cultivars, in six environments.

Results: In all the field assessments, 24 accessions displayed stable high resistance to stripe rust. Significant correlations among environments were observed for both infection (IT) and disease severity (DS), and high heritability levels were found for both IT and DS. Using mixed linear models, 12 quantitative trait loci (QTLs) significantly associated with IT and/or DS were identified. Two QTLs were mapped on chromosomes 5AS and 5AL and were distant from previously identified stripe rust resistance genes or QTL regions, indicating that they may be novel resistance loci.

Conclusions: Our results revealed that resistance alleles to stripe rust were accumulated in Sichuan wheat germplasm, implying direct or indirect selection for improved stripe rust resistance in elite wheat breeding programs. The identified stable QTLs or favorable alleles could be important chromosome regions in Sichuan wheat that controlled the resistance to stripe rust. These markers can be used molecular marker-assisted breeding of Sichuan wheat cultivars, and will be useful in the ongoing effort to develop new wheat cultivars with strong resistance to stripe rust.

Keywords: 55 K SNP; Genome-wide association study; Resistance; Sichuan wheat accessions; Stripe rust.

Fig. 1

The distribution of infection type (IT) and disease severity (DS) for stripe rust based on BLUP values. a exhibited the IT distribution. b exhibited the DS distribution. IT, infection type; DS, disease severity; HR, highly resistance; MR, moderately resistance; MS, moderately susceptible; S, susceptible; The data are supplied in Additional file 2

Fig. 2

The genetic diversity between landraces and cultivars using SNP markers. SNP, single nucleotide polymorphism; PIC, polymorphism information content index; The data are supplied in Additional file 3

Fig. 3

Population structure of 244 Sichuan wheat accessions. a Hierarchical clustering based on Ward method with K-matrix indicated the 244 accessions could be divided into three clusters (cluster 1, cluster 2 and cluster 3). b Heat map of the IBS (identity by state) relative K-matrix. c The population structure of 244 accessions with Bayesian clustering analysis. Two colors stand for 2 different compositions. Different color area of each line represented different proportion of composition. The Sub-population 1 (SP1) mainly showed as green color. The Sub-population 2 (SP2) mainly showed as red color. d Heat map of IT and DS (IT, infection type; DS, disease severity) with BLUP (best linear unbiased prediction) values across six environments in the field. Blue to yellow to red lines show resistance to intermediate to susceptibility to stripe rust of the corresponding accessions. e Percentage memberships of accessions from two subpopulations based on Q-matrix. The landraces showed as blue color and cultivars showed as red color. The red dashed divided the K-matrix into three parts according the hierarchical clustering and the black dashed divided the items into two parts according the Bayesian clustering

Fig. 4

Genome-wide average linkage disequilibrium (LD) decay over physical distances based on 44,059 SNP markers. The red curve represents the model fits to LD decay. The blue line represents the half LD decay distance was 2.12 Mb when the LD declined to 50% (r² = 0.65) of its initial value. Mb, million base pairs

Fig. 5

The P values of associated loci with infection type (IT) and disease severity (DS) displayed as Manhattan plots. The associated loci with IT and DS in six environments were displayed as Manhattan plots with P values across 21 wheat chromosomes. The significant associated locus was considered as –log₁₀(P) > 3 which upper the blue lines. IT, infection type; DS, disease severity. CZ15 = Chongzhou 2015, CZ16 = Chongzhou 2016, MY16 = Mianyang 2016, CZ17 = Chongzhou 2017, MY17 = Mianyang 2017, WJ17 = Wenjiang 2017

Fig. 6

The position of the potentially novel QTLs on the chromosome in this study. The chromosome lengths were all standardized to the same relative length. QTLs marked as red color on the left side of chromosomes were the potentially new QTLs in this study. The reported genes and QTLs were marked as black color and mapped on the left and right side of the chromosomes separately which referred Wang and Chen (2017) [25]