Estimation of Additive Maternal and Cytoplasmic Genetic Variances In Animal Models

Abstract

Two pathways for maternal genetic effects to act on production traits are additive maternal and cytoplasmic. Estimation of these variances have generally assumed that the correct maternal pathway is known. A method is described that allows simultaneous estimation of additive maternal and cytoplasmic genetic variances as well as additive direct genetic variances and error variances using an animal model. Data were simulated assuming an additive maternal model, cytoplasmic model, or both. The method was tested by analyzing the data under the true or incorrect maternal model using REML with an animal model that accounted for additive genetic relationships. Additive direct, additive maternal, cytoplasmic, and error variances were correctly partitioned when estimated under the true model. However, when additive maternal or cytoplasmic genetic effects were ignored in the model used for analysis, estimates of additive direct variance were significantly inflated. Also, under an additive maternal genetic model, a small cytoplasmic variance was estimated although none was present in the data.