Practical and statistically sound methods are required to provide reliable assessments of the tolerance to P. thornei of large numbers of wheat genotypes for plant breeders’ selections and growers’ choice of cultivars to sow in the subtropical grain region of eastern Australia. We showed from six experiments that tolerance indices for wheat genotypes as ratios of grain yield from paired high and low P. thornei treatments were highly significantly related to grain yield on P. thornei infested field sites. We then applied factor analytic multi-environment trial (MET) analysis to 29 field experiments that tested 784 unique wheat genotypes for tolerance to P. thornei based on grain yield on infested sites. Tolerance to P. thornei was effectively modelled with three factors rotated to a principal components solution, which accounted for 84.0, 4.7 and 3.6% respectively of the genotype x experiment variance and covariance. We then validated the estimated best linear unbiased predictions based on these three components (PA(1 + 2 + 3)-eBLUPs) as a quantitative tolerance index to predict relative grain yield of wheat genotypes in independent experiments on P. thornei infested sites. The range of PA(1 + 2 + 3)-eBLUPS was subdivided into nine arithmetically equal subranges to provide nine ordinal alpha ratings from very intolerant to tolerant for growers’ sowing guides and to select nine genotypes as P. thornei tolerance checks in future experiments. There was a 60% loss of grain yield between genotypes in the very intolerant category compared with the tolerant category emphasising the need for continued characterisation of wheat for tolerance to P. thornei.
Read full abstract