Canopy coverage phenotyping and field spatial variability adjustment as an efficient selection tool in soybean breeding

Abstract

AbstractAccurate estimation of grain yield potential in soybean [Glycine max (L.) Merr.] progeny rows (PRs) by measuring yield itself is challenging due to the small number of seeds available. To obtain more precise estimates of soybean yield potential and control nongenetic sources of variability, soybean breeders in the United States use secondary traits, visual selection (VIS), adjustments for field spatial variation, pedigree (PED) information, and unmanned aerial systems–based plant phenotypes; however, there are limited comparisons among the different PR testing procedures. We conducted a selection experiment in 2018 PR populations developed for yield and diversity from four soybean breeding programs. Then, we compared the performance of the lines selected using 13 selection categories in 2019 preliminary yield trials (PYTs). The sources of information used across categories included spatially adjusted (SP) traits, PED information, the canopy of the plant measured by aerial and ground digital images, reproductive length (RL), and grain yield (YLD). SP trait covariates and canopy data were the information sources most highly associated with lower yield ranks and higher yield performance of PYT. The most effective secondary trait was average canopy coverage (ACC) measured by high‐throughput phenotyping (HTP) platforms. Our selection experiment shows that ACC used as a secondary trait in combination with SP trait covariates effectively selects high‐yielding lines from non‐replicated experiments. Based on the scenarios considered in this study, it may be possible to increase the gain from selection by phenotyping secondary traits using HTP and implementing spatial variation adjustments in PR trials, which could help enhance crop productivity.