High-Throughput Architectural Trait Phenotyping for Association Mapping.

Plant architectural traits, such as plant height, leaf angle and leaf area, are important factors in determining grain and biomass productivity of sorghum (Sorghum bicolor (L) Moench). However, collecting data on these architectural traits is labor intensive and time-consuming. A genome-wide association study (GWAS) is a method often used to identify genes/genomic regions controlling traits of interest. GWAS requires the use of hundreds of diverse lines, which increases the time of data collection. This study examined the reliability of machine vision to quickly phenotype a large number of lines from an association panel under field conditions. The use of a field-based robotic platform facilitated the quick collection of stereo images. These images were used to create a 3D reconstruction of plants from which the following phenotypic features were automatically extracted: plant height, vegetative area, hedge width, and vegetation volume.

Comparing image-derived descriptors to manual measurements of conventional plant architecture parameters allowed for the validation of their biological significance. Association mapping identified significant markers/genomic regions that control these plant architectural features in a panel of diverse sorghum lines. Markers associated with plant height localized to regions previously reported to control this trait. Markers explaining variation in other traits represent new discoveries that could be applied in sorghum breeding programs, after validation. The image processing used in this study contributes new knowledge to the development of high-throughput phenotyping techniques and methods. The completely automated machine vision process utilized herein represents a novel tool for plant breeders.