Relative utility of agronomic, phenological, and morphological traits for assessing genotype-by-environment interaction in maize inbreds

Celeste M. Falcon, Shawn M. Kaeppler, Nicholas Haase, Naser AlKhalifah, Darwin A. Campbell, Ignacio Ciampitti, David Ertl, Sherry Flint-Garcia, Michael A. Gore, Candice N. Hirsch, James B. Holland, Nick Lauter, Carolyn J. Lawrence-Dill, Jonathan P. Lynch, Torbert Rocheford, Nathan Springer, Randall J. Wisser, Edgar P. Spalding, Nathan D. Miller, Martin Bohn, Lisa Coffey, Edward S. Buckler, Jode Edwards, Christopher Graham, Joseph Knoll, Diego Jarquín, Elizabeth C. Lee, Aaron Lorenz, M. Cinta Romay, Margaret Smith, Teclemariam Weldekidan, Rebecca Nelson, Patrick S. Schnable, Brian Scully, Mitchell R. Tuinstra, Renee Walton, Wenwei Xu, Natalia de Leon

Plant breeders face the challenge of genotype x environment interaction (G x E) in comprehensively breeding for expanded geographic regions. An improved understanding of G x E sensitivity of traits and the environmental features that effectively discriminate among genotypes will enable more efficient breeding efforts. In this study of 31 maize (Zea mays L.) inbreds grown in 36 environments that are part of the Genomes to Fields Initiative, we measured 14 traits, including flowering date, height, and yield components (i.e., ear and kernel dimensions) to (i) identify traits that are the most sensitive indicators of G x E; (ii) determine how geographic location and weather factors influence environments' discriminability of inbreds; and (iii) detect patterns of stability in better and worse discriminating environments. Genotype x environment interaction explained between 9.0-20.4% of the phenotypic variance with greater effects in the yield-component traits. Discriminability of environments varied by trait. Midwest locations (where 26 of the 31 inbreds were developed) were among the most discriminating environments for more traits, while environments in the West and East tended to be less discriminating. Weather factors during silking were significantly different between the most and least discriminating environments more often than average weather across the season or during the period from planting to silking. Stability of genotypes varied by trait, and performance was usually not correlated with stability. The dissection of complex traits, such as yield into component traits, appears to be a useful approach to understand how environmental factors differentially affect phenotype.