In order to evaluate alternative genetic models to the additive dominant model, weights at yearling (PS) of 35,931 animals, sired by 752 bulls and 30,535 cows of Aberdeen Angus (A) and Nellore (N) breeds and the genetic groups from their crosses were used. Five different genetic models (M) were tested: M1, containing the direct additive fixed genetic effect (DA), heterozygote direct (HD), epystatic direct (ED), and direct joint additive direct (DJA); M2 was equal to M1, excluding DJA effect; M3 was equal to M1, excluding ED effect; M4 was equal M1, excluding ED and ACD effects, and M5 was equal to M1, excluding HD, ED, and DJA effects. The models were analyzed by different methods: Least Square Means Method (MQM), Ridge Regression Method (RC), and Restricted Maximum Likelihood Method (REML). Estimated coefficients by RC showed magnitude and sign which were biologically explained. The estimates of the covariances, parameters, and genetic values varied among the models, indicating the importance of the correct choice of the model for analysis, being necessary a previous knowledge of the studied phenomenon and its biological interpretation. Besides, it should always be considered the relationship between the independent variables before choosing a multibreed genetic analysis model. Important additional effects to the DA effect were considered by the inclusion of the HD and ED effects to the models for analysis. The DJA math notation, currently used in the literature and tested in the present study, was not able to explain the breed complementarity, due to the multi colinearity among the studied effects
additive effect; between breeds complementarily; epystatic effect; heterozygote effect; yearling weight
