Comparison of mixed linear model analysis and combining ability analysis of genetic effect of main traits in maize

Fourteen maize inbred lines and forty-nine crosses mated by NC II design were used in the experiment. The genetic effects for seven characters were analyzed with mixed linear model analysis and combining ability analysis. The mixed liner model analysis indicated: as for kernel weight plant~(-1), kernel number row~(-1), 100 kernels weight, ear diameter, and filling period, thre were dominant and A-A epistatic genetic effect. As for rows/ear and growth period, three were additive and dominant and A-A epistatic genetic effect, and A-A epistatic effect was not significant. But based on additive-dominant model, combining ability analysis just revealed the additive effect and dominant effect. The differences of genetic variance components between mixed model analysis and combining ability analysis were signfiicant. It was suggested by combining ability analysis that the difference of genetic variance component in some characters was mainly caused by less estimating genotype-environment interaction, and/or by more estimating additive effect in the others. The difference of heritability in the broad sense was caused by genotype-environment interaction. The difference of heritability in the narrow sense was caused by genotype-environment interaction and/or additive effect/A-A effect estimated by the two different methods.