|
|
||||||||
University of California, Davis 95616,3
Abstract
Frequently, exotic, hybrid and native sires are evaluated based on progeny records, using additive models for direct effects and assuming common genetic variances across breed groups. In some cases, however, additive maternal and nonadditive direct and maternal fixed and random effects may be important. Also, heterogeneous variances across genetic groups may exist. Hence, sire proofs from additive models will be biased and not of minimum variance. Sire-maternal grandsire models that account for direct and maternal, additive and nonadditive fixed and random effects as well as different additive and nonadditive variances across genetic groups are presented. Direct and maternal effects are treated as separate characters. Sire proofs for direct and maternal traits contain group (fixed) and sire (random) additive and nonadditive effects. Problems of confounding and multicollinearity may arise due to empty cells in the design. Ridge regression may be used to help remove problems of multicollinearity. Extension to multitrait evaluation is straightforward.
1 Present address: Dept. of Anim. Sci., Cornell Univ., Ithaca, NY 14853.
2 The authors gratefully acknowledge A. P. Fenech for his statistical assistance, G. E. Bradford for many useful discussions and E. J. Pollak for editorial comments on this paper. The senior author also acknowledges the financial support of the Univ. of California, Davis, throughout his graduate studies.
This article has been cited by other articles:
![]() |
S. Newman, A. Reverter, and D. J. Johnston Purebred-crossbred performance and genetic evaluation of postweaning growth and carcass traits in Bos indicusxBos taurus crosses in Australia J Anim Sci, July 1, 2002; 80(7): 1801 - 1808. [Abstract] [Full Text] [PDF] |
||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |