We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Modelling non-additive genetic effects using ridge regression for an Angus–Nellore crossbred population.
- Authors
Bertoli, C. D.; Neto, J. Braccini; McManus, C.; Cobuci, J. A.; Campos, G. S.; Piccoli, M. L.; Roso, V.
- Abstract
Data from 294 045 records from a crossbred Angus × Nellore population were used to estimate fixed genetic effects (both additive and non-additive) and to test different non-additive models using ridge regression. The traits studied included weaning gain (WG), postweaning gain (PG), phenotypic scores for weaning (WC) and postweaning (PC) conformation, weaning (WP) and postweaning (PP) precocity, weaning (WM) and postweaning (PM) muscling and scrotal circumference (SC). All models were compared using the likelihood-ratio test. The model including all fixed genetic effects (breed additive and complementarity, heterosis and epistatic loss non-additive effects, both direct and maternal) was the best option to analyse this crossbred population. For the complete model, all effects were statistically significant (P < 0.01) for weaning traits, except the direct breed additive effects for WP and WM; direct complementarity effect for WP, WM, PP and PM and maternal epistatic loss for PG. Direct breed additive effect was positive for weaning traits and negative for postweaning. Maternal breed additive effect was negative for SC and WP. Direct complementarity and heterosis were positive for all traits and maternal complementarity and heterosis were also positive for all traits, except for PG. Direct and maternal epistatic loss effects were negative for all traits. We conclude that the fixed genetic effects are mostly significant. Thus, it is important to include them in the model when evaluating crossbred animals, and the model that included breed additive effects, complementarity, heterosis and epistatic loss differed significantly from all reduced models, allowing to infer that it was the best model. The model with only breed additive and heterosis was parsimonious and could be used when the structure or amount of data does not allow the use of complete model. The identification of the best crossbred sires is not an easy task as these animals are composed of distinct genetic materials from different breeds. The objective of the present study was to compare different models under ridge regression methodology to identify the best method for selecting crossbred bulls. This selection is a requisite to guide future multiracial genetic evaluations. The model that contemplated all tested effects (breed additive, complementary, heterosis and epistatic loss, both direct and maternal) proved to be the best.
- Subjects
BEEF cattle; WEIGHT gain; HETEROSIS
- Publication
Animal Production Science, 2019, Vol 59, Issue 5, p823
- ISSN
1836-0939
- Publication type
Article
- DOI
10.1071/AN17439