Development of high-throughput phenotyping tools to enhance soybean yields

Abstract

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] Increasing yield is the primary objective of most crop breeding programs and yield is often the primary selection criterion. In drought-prone environments, progress is slowed by large genotype x environment interactions arising from unpredictable rainfall among seasons and locations. The assessment of traits associated with drought tolerance mechanisms can shape crop growth and yield under this stress. However, these traits are typically assessed by destructive sampling or wet chemistry methods, which are tedious and time-consuming to perform, especially when dealing with large segregating populations. Therefore, phenotyping tools that are easy, inexpensive, nondestructive and stable over a measurement period are necessary. To address this question three different studies are presented, the first study tested the usage of common digital cameras coupled with image analysis and statistical procedures as a potential phenotyping tool to predict yield and other crop traits in two soybean cultivars. This technique proved to effectively identify the dynamics of canopy growth and changes due to drought stress intensity and senescence; it also presented high correlation with several of the plant traits measured and yield. The second study validated the applicability of vegetation indices derived from multispectral aerial images as yield-prediction tools for soybean under different drought stress levels. This study found that the vegetation indices estimated are good predictors of soybean yield only under moderate drought or unstressed conditions, the regressions found were able to explain up to 80% of the data based on R2 values. The third study explored the genetic diversity in a group of soybean lines in terms of canopy chlorophyll concentration and tested the capabilities of vegetation indices derived from multispectral aerial images as predictors of this trait. In these experiments, a group of 21 high-Chlorophyll, high-yielding lines was found. The best predictor of chlorophyll was the Near-Infrared + Red vegetation index.