description
- Late maturity alpha-amylase (LMA) is a grain defect causing a reduction in Hagberg falling number and potential failing to meet Australian receival or market specifications. Currently screening for LMA susceptibility during breeding relies on a time consuming, costly and relatively low-throughout benchmark assay that involves growing plants under controlled conditions and inducing the expression of LMA to generate a scorable phenotype. This project will develop an innovative LMA susceptibility screening method with increased throughput, scalability, repeatability and significantly reduced costs compared to the current benchmark assay. It implements a paradigm shift in thinking about how LMA susceptibility screening is approached by aiming to develop a method to identify wheat lines that are susceptible to LMA in non-induced seed, i.e. seed grown under non-inducive conditions. Machine and deep learning approaches will be used to develop a prediction model in which biomarkers derived from electromagnetic, biochemical and proteomic spectra of non-induced seed are used to predict LMA susceptibility. Specifically, we will look for biomarkers in non-induced grain that are highly predictive of LMA susceptibility, eliminating the need for inducive environmental conditions to screen for LMA susceptibility. This innovative approach will enable LMA susceptibility screening to be applied as an early generation selection tool in breeding to accelerate the speed and efficiency at which LMA susceptibility can be controlled and managed in breeding programs.