Carlos D Messina, Carla Gho, Graeme L Hammer, Tom Tang, Mark Cooper

J Exp Bot.; 2023 Sep 2; 74(16):4847-4861. doi: 10.1093/jxb/erad231.

Abstract

We review approaches to maize breeding for improved drought tolerance during flowering and grain filling in the central and western US corn belt and place our findings in the context of results from public breeding. Here we show that after two decades of dedicated breeding efforts, the rate of crop improvement under drought increased from 6.2 g m-2 year-1 to 7.5 g m-2 year-1, closing the genetic gain gap with respect to the 8.6 g m-2 year-1 observed under water-sufficient conditions. The improvement relative to the long-term genetic gain was possible by harnessing favourable alleles for physiological traits available in the reference population of genotypes. Experimentation in managed stress environments that maximized the genetic correlation with target environments was key for breeders to identify and select for these alleles. We also show that the embedding of physiological understanding within genomic selection methods via crop growth models can hasten genetic gain under drought. We estimate a prediction accuracy differential (Δr) above current prediction approaches of ~30% (Δr=0.11, r=0.38), which increases with increasing complexity of the trait environment system as estimated by Shannon information theory. We propose this framework to inform breeding strategies for drought stress across geographies and crops.

See https://pubmed.ncbi.nlm.nih.gov/37354091/

Fig. 6. Simulated yields define theoretical maize yield response to evapotranspiration for 80 and 99 percentiles (A, lines), maize yield variability for hybrids with contrasting response to water deficit across a range of evapotranspiration (B), and spatial distribution of yields for 2 years with different drought patterns at 30 × 30 km resolution (C). Yield observations shown in (A) are for a single cross-hybrid grown at three locations in the western US corn belt for maize grown under rainfed and irrigated conditions, and under normal (filled symbols) and increased plant population by 1 plants m^–2 (open symbols). Data are from Messina et al. (2020a).