The greatest difficulty in maize breeding is to generate and evaluate all possible hybrid combinations (complete diallel). To reduce this difficulty, several genetic mating designs have been proposed, such as the partial diallel, circulant diallel, and testcross. These designs can cause wide differences in the estimates of variance components and of the value of the best parents, mainly due to their principles and representativeness of crosses, and intersections between the heterotic groups. With the advent of Genomic Selection (GS), it has become possible to predict, with some accuracy, the behavior of individuals not assessed under field conditions. However, the best mating design to build the training population has not been defined. Such a design should maximize accuracy given constraints on costs and on the logistics of the crosses to be made. The objective of this project is to evaluate the accuracy of GS to predict tropical maize single-crosses obtained by different mating designs, and to identify the design with the best cost benefit ratio. Four different training populations will be considered, derived from complete diallel, partial diallel, circulant diallel and testcross designs. These designs were implemented on a set of 224 inbred lines. In total, we obtained approximately 3,000 simple-crosses. The inbreed lines will be subject to Genotyping by Sequencing (GBS), at Cornell University, USA. Phenotypic data on the hybrids is being obtained by a partnership between ESALQ-USP and Helix Sementes, a private company, under field conditions at five sites in Brazil over three years and two growing seasons (summer and winter). From the phenotypic and genotypic data, the genetic-statistical analysis will be carried out using mixed models equations. This step will also be performed at the Center of Genetics and Plant Breeding, Cornell University, with the collaboration of Dr. Jannink.
News published in Agência FAPESP Newsletter about the scholarship: