The wheat production areas of South Africa are affected by several diseases. Among them, the three rusts, namely stem rust, leaf rust and stripe rust caused by Puccinia graminis f. sp. tritici Eriks. & E. Henn., Puccinia triticina Eriks and Puccinia striiformis West. f. sp. tritici Eriks. & E. Henn. respectively, are the most prominent. Their severity and importance in a specific area is largely influenced by climatic factors and they can lead to yield losses of more than 50% (Terefe et al. 2024).

Wheat rusts can effectively be controlled by the deployment of multiple resistance genes that confer durable resistance. The utility and durability of resistance genes can be extended considerably if multiple genes are combined in new varieties, thus creating more complex genetic barriers that will less likely be overcome by the pathogen (Mapuranga et al. 2022). One of the most effective strategies to incorporate resistance genes is by a structured process of pre-breeding based on biotechnology tools such as molecular marker-assisted selection (MAS), the production of doubled haploids (DH) and genomic selection (GS).

With the rapid decline in genotyping costs and the development of statistical methods to accurately predict marker effects, GS has been accepted as a molecular breeding tool for the improvement of complex traits where many loci of small effects control the trait (Crossa et al. 2014; Crossa et al. 2017). It has the potential to reduce the number of cycles in a breeding process and decrease phenotyping costs. Moreover, it facilitates the use of recurrent selection in wheat, a well-documented breeding method that increases favourable alleles in a population (Rutkoski et al. 2015).

Therefore, this study aims to implement genomic selection in a recurrent selection pre-breeding program and evaluating its ability to rapidly and accurately predict rust resistant genotypes to be selected as crossing parents for crop improvement and potential release as cultivars in breeding programs.