Genotype-by-environment interactions affecting heterosis in maize

Abstract

The environment can influence heterosis, the phenomena in which the offspring of two inbred parents exhibits phenotypic performance beyond the inbred parents for specific traits. In this study we measured 25 traits in a set of 47 maize hybrids and their inbred parents grown in 16 different environments with varying levels of average productivity. By quantifying 25 vegetative and reproductive traits across the life cycle we were able to analyze interactions between the environment and multiple distinct instances of heterosis. The magnitude and rank among hybrids for better-parent heterosis (BPH) varied for the different traits and environments. Across the traits, a higher within plot variance was observed for inbred lines compared to hybrids. However, for most traits, variance across environments was not significantly different for inbred lines compared to hybrids. Further, for many traits the correlations of BPH to hybrid performance and BPH to better parent performance were of comparable magnitude. These results indicate that inbred lines and hybrids show similar trends in environmental response and both are contributing to observed genotype-by-environment interactions for heterosis. This study highlights the degree of heterosis is not an inherent trait of a specific hybrid, but varies depending on the trait measured and the environment where that trait is measured. Studies that attempt to correlate molecular processes with heterosis are hindered by the fact that heterosis is not a consistent attribute of a specific hybrid.

Fig 1. Better parent heterosis (BPH) comparisons for 25 traits and 47 hybrids across 16 environments.

A) Pearson correlation coefficients (r) of BPH between traits; gray shaded cells indicate missing data. B) Network visualization of Pearson correlation coefficients of BPH between traits. Only correlation coefficients less than -0.3 or greater than 0.3 are shown. Traits in yellow circles and green rectangles are reproductive and vegetative traits, respectively. Red line, r<-0.3; gray line, 0.3<r<0.5; blue line, r>0.5. C) Average BPH rank scaled with white (highest BPH rank) to dark blue (lowest BPH rank). Hybrid genotypes are followed by the parental identity by state value.

Fig 2. Correlation coefficient for percent better parent heterosis (%BPH), hybrid performance, and better-parent performance of plant height at different development stages in different environments.

The numbers of 14–49 in x-axis indicate days after planting and PHt is plant height at physiological maturity.

Fig 3. Coefficient of variation within and across environments for hybrid and inbred genotypes.

A) Coefficient of variation within plot (6 plants were phenotyped within each plot). B) Coefficient of variation across all available environments for each trait. In each figure blue and red colors indicate hybrid and inbred, respectively. * significant at p = 0.05; ** significant at p = 0.01; *** significant at p = 0.001 in a two-tail t-test between the inbred and hybrid genotypes.

Fig 4. Percent better parent heterosis (%BPH) for grain weight (GWT) and plant height at maturity (PHt) for 47 hybrids across 16 environments.

A and C) Heatmap of %BPH for GWT (A) and PHt (C); black shaded cells indicate missing data. The green and blue arrow in each plot indicates the hybrids that have the highest and lowest %BPH across all 16 environments. Environments and hybrids were clustered using hierarchical clustering (trees not shown). B and D) Highest (indicated by green arrows in A and C) vs. lowest (indicated by blue arrows in A and C) %BPH hybrids across all environments for GWT (B) and PHt (D). Red dots are the eight low-density environments and black dots are the eight high-density environments. H7 is PHP02 x DK3IIH6, H30 is LH185 x DK3IIH6, H32 is LH198 x LH185, H34 is LH82 x W606S.

Fig 5. Relationships among percent better parent heterosis (%BPH), hybrid, and better-parent performance.

Plots A-D are for grain weight (GWT) and E-H are for plant height at maturity (PHt). A and E) Hybrids with the highest %BPH across 16 environments. B and F) Hybrids with the highest standard deviation of the rank of %BPH among all 47 entries. C and G) Hybrids with the lowest standard deviation of the rank of %BPH among all 47 entries. D and H) Correlation coefficient of hybrid vs. %BPH and better-parent vs. %BPH across 16 environments for each hybrid. Colored dots represent the highest %BPH (red—A and E), highest standard deviation of the rank of %BPH (green—B and F), and lowest standard deviation of the rank of %BPH (blue—C and G). For A-C and E-G dots along the x-axis represent each of the 16 environments.