Enabling breeding of spring wheat for optimisation of flowering time for current and future climates by linking genetic maps to simulation model parameters

Abstract

In Australian wheat production, optimizing wheat phenology is essential to reach yield potential and to avoid within‐season stress at critical periods, especially around flowering. Identifying loci that determine heading date of wheat cultivars and developing models to predict genotype performances under different pedo‐climatic scenarios will assist optimization of phenology through breeding. In this study, association genetics for heading date (in thermal units °Cd), earliness components (PS, photoperiod sensitivity; VR, vernalization requirement; EPS, earliness per se) and simulation model (APSIM) phenology parameters from a panel of Australian cultivars and breeding lines identified loci with stable effects (i.e. repeatable over two years) on heading date and its components, providing new insights into the genetic architecture of this trait in commercial Australian germplasm. Few major regions with stable effects on earliness components were detected: the Ppd‐D1 region on 2D for photoperiod sensitivity (PS) and earliness per se (EPS), one region on 5B for photoperiod sensitivity (PS), one region on 6B for EPS and the Vrn‐A1 region on 5A for vernalization requirement (VR). Other regions with stable but lower effects were detected on 1A and 2D for PS, on 5A and 6B for EPS and on 1A and 5D for VR. Finally, other regions with stable effects on heading date and different earliness component were located on chromosomes 1A, 2B, 4B, 5B, 6B and 7B (PS and EPS), 2A, 3A and 7A (EPS and VR). QTL‐based model parameters were used to simulate heading dates in different location × sowing dates across the Australian wheat belt for a validation dataset of independent genotypes. Relationships between average observed and predicted heading dates for four main regions of the Australian wheat belt showed good performance in prediction of independent lines from QTL information alone (r2 0.61 to 0.83). This QTL‐based model requires only genetic information. It allows the testing of different putative genotypes under various pedo‐climatic scenarios, and opens the way to the identification of target ideotypes for these scenarios. This allows breeders to optimize heading date and develop adaptation to anticipated climate changes.