Quantitative genetic studies of A/B zeins using a new model to test non-allelic interactions.

A genetic model developed by Bogyo et al. (1988) for quantitatively inherited triploid endosperm characters (an extension of the well-known Mather-Jinks model) is not well-suited for estimating epistatic interaction effects because it requires the assumption that, in segregating loci, all alleles positively affecting a particular character are in one of the inbred parental lines. To better explain zein inheritance in maize, a new model was developed not relying on this assumption. This model was tested by quantitative analysis of A/B zeins, the predominant prolamin storage proteins of maize, using reversed phase high-performance liquid chromatography of two inbred lines, their reciprocal F1 crosses, the F2 generation, backcrosses, and reciprocal backcrosses to both parent lines. The model required epistatic components to be included for an excellent fit for most protein peaks.