Compare analysis to genetic characteristics and potential in breeding of maize(Zea mays L.) inbred lines between mixed linear model and combining ability

Fourteen maize inbred lines and forty-nine crosses mated by NCⅡdesign were used in the experiment. The genetic effects of kernel weight/plant were analyzed with mixed linear model analysis and combining ability analysis. And the results were compared. The major results are as following: There are dominant and A-A epistatic genetic effect in the mixed linear model analysis. But combining ability analysis is based on additive-dominant model. It reveals that just have additive effect and dominant effect. The differences of genetic variance component and heritability between mixed model analysis and combining ability analysis are not significant. Array sum of parents (including parents effect) comply with GCA in number order (from high to low). Correlation coefficient between the dominant effect of LUP (linear unbiased predict) and SCA (special combining ability)/heterosis are 0.70 and 0.93 respectively at 0.01 significant level. Correlation coefficient between SCA and heterosis is just 0.56 at significant 0.01 level.