Genomic prediction (GS) studies using diversity panels are essential to identify genetic variations associated with traits of interest in maize. Unfortunately, most of these studies have been conducted on temperate germplasm and on a global germplasm collection in which tropical genotypes are under-represented. Nonetheless, a continuous effort has been directed to improving the accuracy of GS. While genotyping is currently a precise and efficient mechanized process, phenotyping is still laborious, low-throughput, and highly sensitive to environmental variations. Also, extreme shifts in the weather pattern due to climate change complicates the selection of superior genotypes with broad adaptability. In this context, adopting GS models that account for genotypes x environments reaction norms, crop growth models, and environmental covariates should increase the accuracy of genomic predictions. Thus, the objective of this project is to develop a diversity panel of tropical maize for genomic prediction studies that incorporate high-throughput phenotyping, plant growth models, and environmental covariables. For that, 360 tropical maize lines from ESALQ-USP, IAPAR, IAC, and CIMMYT will be genotyped and phenotyped using traditional methods and multispectral imaging in eight environments (two locations, two years, and two seasons). With this data, several GS models will be tested and compared for prediction accuracy and selection coincidence. Besides the development of novel GS models and high-throughput phenotyping protocols in tropical maize, we will also organize, characterize, and publicize a panel of tropical maize lines (data and genetic material) to the scientific community that will serve as the benchmark for new studies of this nature in tropical maize.

FAPESP (2017/24327-0)