Leaf rust, caused by the fungus Puccinia triticina Eriks., is a prevalent disease of wheat that affects grain yield and quality. Leaf rust resistance is an important trait that is evaluated in the registration of wheat varieties in western Canada and is an effective strategy for sustainable disease management. However, the genetic basis of this resistance is not fully understood in Canada Western Red Winter (CWRW) wheat.

This study aimed to identify the quantitative trait loci (QTL) controlling leaf rust resistance using genome-wide association studies (GWAS) and develop genomic selection (GS) models for improving leaf rust resistance in winter wheat. The wheat GWAS panel consisted of approximately 300 western Canadian winter wheat breeding lines and cultivars and 100 winter wheat breeding lines and cultivars from the USA, eastern Canada, and Europe. The panel was evaluated for leaf rust resistance in seedling tests using five different races of P. triticina, namely 12-3 MBDS, 128-1 MBRJ, 74-2 MGBJ, 06-1-1 TDBG, and 77-2 TJBJ. The same population was also tested for resistance in inoculated field trials located in Winnipeg and Morden, Manitoba, following randomized complete block designs (RCBD), with two replicates per field trial in the 2022-23 and 2023-24 growing seasons. Genotyping was done using the Illumina Infinium Wheat Barley 40K SNP array and the 25K wheat Infinium array. The DNA marker and leaf rust datasets were used for GWAS. Putative Quantitative Trait Nucleotides (QTNs) associated with leaf rust resistance have been detected from the GWAS analyses, which indicated the presence of Lr genes in the GWAS panel. The results indicated that the leaf rust-resistant genes Lr16 and Lr24 are likely present in the GWAS panel.

The leaf rust and SNP marker data were also used to develop genomic selection (GS) models for estimating leaf rust resistance in CWRW breeding germplasm. The accuracy of these GS models for predicting the leaf rust resistance of wheat germplasm was assessed via five-fold cross-validation. Overall, prediction accuracies, indicated by Pearson’s correlation coefficient (r), ranged from 0.66 to 0.67 for disease severity (DS), 0.59 to 0.60 for infection type (IT), and 0.65 to 0.67 for the average coefficient of infection (ACI) in field trials. The r values for IT data from seedling tests varied between 0.45 and 0.86 for different GS models. Improved knowledge of the resistance genes in Canadian winter wheat and DNA markers for selecting these genes will improve the efficiency of wheat breeding programs.