Maize’s (Zea mays L.) leaf angle has changed over the last 60 years because of intense selection pressure for high yielding hybrids under increasing planting densities. Leaf angle plays a crucial role in distributing sunlight to different canopy leaves and its optimization across the canopy is essential for increasing productivity per unit of land. This dissertation seeks to explain and predict leaf angle variation across multiple canopy levels in maize through genetic mapping, meta-analysis, and genomic prediction.

Two genetic mapping populations were developed using inbred lines B73, PHW30, and Mo17 that represent important maize heterotic groups. The two populations were genotyped using genotyping-by-sequencing (GBS) and phenotyped for lower canopy leaf angle in selected lines from the F2 and F2:3 generations to explain this portion of the canopy’s genetic variation. A total of 12 quantitative trait loci (QTL) were detected across both populations and generations, including one consistently detected on chromosome 1. The detected QTL were mapped into genomic bins along with QTL from 19 previous studies to gain a comprehensive understanding of the natural variations underlying leaf angle. A meta-analysis with genomic bins revealed 58 genomic hotspots that contained 33 candidate genes. Together, these results enrich our understanding of the genetic control of lower canopy leaf angle in inbred lines representing the major heterotic groups in maize and provide a roadmap for future researchers to investigate the molecular basis for these natural variations.

Doubled haploids were developed from the selected F2 lines and used to explore leaf angle variation observed in other portions of the canopy. These were genotyped with GBS and phenotyped for leaf angle across four canopy levels over multiple years for genetic mapping. The four leaf angle phenotypes were regressed on their canopy position to derive three additional traits and understand how leaf angle fluctuates across the canopy. The relationship between the detected QTL’s effects and canopy position was explored to improve the understanding of how leaf angle is controlled across the canopy. Genetic mapping revealed 59 QTL across the seven traits, including two major effect QTL on chromosomes 1 and 5. Developmental reaction norms with QTL detected across the canopy revealed QTL with genetic effects that were stable and dynamic in response to canopy position. Phenotyping and genetically mapping leaf angle for individual leaves across the canopy revealed new insights into the genetic control of leaf angle across the canopy, and the selection of leaf angle QTL with dynamic effects may be beneficial for developing lines adapted to high planting densities.

It is challenging to predict phenotypes utilizing the knowledge explained from genetic mapping populations to other mapping populations or diverse germplasm. Estimating allelic effects from diverse germplasm can overcome this challenge and be useful in developing prediction models. The predictability of leaf angle at two canopy levels and 34 other traits was investigated using the Maize Association Population, which is a collection of diverse germplasm from breeding programs around the world. This panel was assessed for its predictability for these traits through cross-validation and used to train a prediction model for generating genomic estimated breeding values for the US maize national seed bank. Using two empirical populations, predictions for nine traits were empirically validated across multiple environments. An upper bound for prediction reliability was calculated for each predicted line and compared with prediction accuracy to investigate their relationship. Prediction accuracy, assessed by cross-validation for the 36 traits, ranged from weak to strong and was highly dependent on a trait’s repeatability. For the two empirically validated leaf angle traits, prediction accuracy was similar to the values obtained during cross-validation. In lines with high prediction reliability values, higher prediction accuracy was observed. These results suggest the genomic prediction models developed using the Maize Association Population will be a valuable tool to predict leaf angle and other traits for the diverse germplasm at the US maize national seed bank.

The genetic mapping and meta-analysis results provided new insights into explaining the genetic control of leaf angle across multiple canopy levels including identifying candidate genes and revealing QTL effects that are stable or dynamic in response to canopy position. Genomic prediction provided a thorough analysis of the Maize Association Population’s potential for predicting two leaf angle traits along with 34 other traits for the US maize national seed bank. Together, this knowledge will be essential for optimizing leaf angle across the maize canopy in order to increase productivity per unit of land.