Epistasis in Related and Unrelated Maize Hybrids Determined by Three Methods

Abstract

The objectives of this study were (1) to estimate genetic parameters in maize (Zea mays L.), (2) develop and apply a genetic model capable of measuring variance associated with additive, dominance, and two factor epistatic effects in a diallel cross including segregating generations, (3) to measure epistasis as difference of three‐way crosses from the nonparental single crosses, and (4) to compare different methods measuring the importance of epistasis. Three inbreds derived from Stiff Stalk Synthetic and three inbreds related to the Lancaster variety were used to make all possible crosses, their segregating, selling and backcross generations.Heterosis for yield and other traits was larger in crosses involving unrelated lines than in crosses involving related ones. No apparent association of genetic effects estimated from the generation mean analysis with crosses involving related or unrelated lines was found. Dominance effects were the most important genetic effects controlling the performance of all traits. Generation mean analysis was a poor method for measuring additive effects. Significant epistatic effects were found in several crosses. Dominance × dominance effects were consistently negative for all traits except yield, suggesting that selection for yield prior to release of the inbreds might have reduced the diminishing effect of this type of epistasis in cross combinations. Dominance effects for yield were larger than the mean of the pooled additive effects of the lines, suggesting that homozygous beneficial genes (with + effects) and homozygous deleterious (with ‐ effects) were present in lines.In the diallel analysis, the mean squares for dominance effects were larger than for additive effects for yield and plant leaf area, but smaller for plant height, ear height, ear length, and ear diameter. Epistatic variation was significant but small compared to additive and dominance variation. Interactions of genetic effects with locations were significant. When tested against the appropriate interaction mean squares only the dominance mean squares were significant for grain yield. Epistasis was significant in 13 of 54 three‐way crosses. The sign of epistasis was predominately positive for yield, suggesting that lines selected for yield may have certain combinations of genes that are expressed favorably in hybrids. Mean epistatic effects over all crosses were significant for yield, ear height, ear length, and ear diameter.