Skip to main content
ARS Home » Southeast Area » Raleigh, North Carolina » Soybean and Nitrogen Fixation Research » Research » Research Project #441343

Research Project: Drought Resiliency for the Farm- Yield Limitations of Commercial Soybean Varieties under Drought (ISU)

Location: Soybean and Nitrogen Fixation Research

Project Number: 6070-21220-069-118-A
Project Type: Cooperative Agreement

Start Date: Oct 1, 2021
End Date: Sep 30, 2022

Develop drought-tolerant cultivars adapted to locations in Iowa and other drought-prone U.S. environments that will aid farmers with variety choices and guide new public-private breeding efforts to fix the drought problem.

A time series controlled environment experiment that will include several hundred accessions (MG late I to MG III), to study the genetic variation for major indicators of drought tolerance including reflectance and temperature variation. This test will also be repeated in the field. Steps of analysis prior to genomic studies (genome wide association study – GWAS) will include a data pipeline to build an imaging pipeline. Due to the high complexity and dimensionality of the imagery collected, a spatiotemporal deep learning algorithm will be utilized to extract important phenotypic features. Extracted canopy reflectance will be used as an indicator trait for early detection of stress onset and gene discovery. Early detection of stress onset will utilize the extended wavelengths captured to identify pre-visual symptoms of water stress, which will allow farmers using irrigation to more efficiently control water use and conserve resources. Leaf and canopy traits will be extracted from the 3D laser scanning point cloud to identify physical changes in response to drought and flooding. This information in conjunction with other important phenotypic traits will help inform on plastic response to environmental changes that can be implemented in stable line development. Genomic prediction (GP) models combing phenotyping traits and molecular data will be created to determine accuracy improvements in GP using high dimensional phenotyping.