Estimation of heritability by parent-offspring regression.

The implications of bias due to previous inbreeding of parents and genotype x environmental interaction on narrow sense heritability (h(2)) estimates by parent-offspring regression are enumerated. To remove the bias caused by genotype x environment interaction, an analysis of covariance model could be used. In special cases, where phenotypic expression is a result of two organisms interacting, such as in symbiotic N2 fixation, an analysis of covariance model with a test of heterogeneity of slopes is recommended. When host genotype x strain interactions are significant, separate heritability estimates for each strain are suggested to take advantage of genotype x strain interaction, which may be a major factor contributing to the expression of N2 fixation traits.