Linear and logistic models for multiple-breed genetic analysis of heifer fertility in Mexican Simmental–Simbrah beef cattle

Abstract

ABSTRACTThe aim of this study was to compare alternative models for the genetic evaluation of heifer fertility in Simmental–Simbrah cattle. The analyses were conducted using a database with 37,390 female birth information recorded from 1984 to 2007, and 59,018 individuals in the pedigree. Three generalized mixed models were adjusted for a single trait in a multiracial population: linear animal, linear sire and logistic sire. The models were analysed by restricted maximum likelihood procedure with Average Information algorithm. Two strategies of cross-validation were carried out to evaluate the predict ability of the models. The heritability from linear animal and sire models had lower values than that estimated with the logistic model, 0.04 ± 0.00, 0.05 ± 0.00 and 0.20 ± 0.03, respectively. High Spearman and Kendall correlations were observed between the ranks of breeding values (BV) estimated from the linear and logistic sire models, 0.99 and 0.94, respectively. In contrast, these correlations were lower between the animal and sire models, 71% and 54%, respectively. The logistic sire model was the best estimating the BV with ancestors’ information, while the linear animal model (LAM) was the best predicting with scattered information. In general, it was considered that best fit and prediction was produced with the LAM.