The complexity of allele interactions constrains crop improvement and the prediction of disease susceptibility. Additive allele effects are the foundation for selection in animal and plant breeding, and complex genetic and environmental interactions contribute to inefficient detection of desirable loci. Manipulation and modeling of other sources of variation, such as environmental variables, have the potential to improve our prediction of phenotype from genotype. As an example of our approach to analysis of the network linking environmental input to alleles, we mapped the genetic architecture of single and combined abiotic stress responses in two maize mapping populations and compared the observed genetic architecture patterns to simple theoretical predictions. Comparisons of single and combined stress effects on growth and biomass traits exhibit patterns of allele effects that suggest attenuating interactions among physiological signaling steps in drought and ultraviolet radiation stress responses. The presence of attenuating interactions implies that shared QTL found in sets of environments could be used to group environment types and identify underlying environmental similarities, and that patterns of stress-dependent genetic architecture should be studied as a way to prioritize prebreeding populations. A better understanding of whole-plant interactor pathways and genetic architecture of multiple-input environmental signaling has the potential to improve the prediction of genomic value in plant breeding and crop modeling.
Keywords: abiotic stress; modifier networks; multiple stress; nested association mapping; recombinant inbred line.