Maximizing Genetic Response in Breeding Schemes of Dairy Cattle with Constraints on Variance of Response

¹ Research Institute for Animal Production ‘Schoonoord’, PO Box 501, 3700 AM Zeist, The Netherlands
² Roslin Institute (Edinburgh), Roslin, Midlothian EH25 9PS, Scotland

Predicted genetic progress in dairy cattle breeding schemes was maximized with the variances of selection responses constrained, i.e., restricted effective population sizes. This restriction would also lead to schemes with acceptable rates of inbreeding (<.5%/yr).

If the required coefficient of variation of the annual selection response was reduced from .32 to .16, numbers of animals selected, openness of schemes, and generation intervals increased. When elite cows produced 8 offspring annually, this reduction of the coefficient of variation tended toward a conventional progeny-testing scheme. If the number of donor cows was optimized, responses increased 2%, and the breeding schemes became virtually closed. Variances of responses were reduced by selecting fewer, but proven, bulls, as is done in hybrid multiple ovulation and embryo transfer schemes, which select progeny-tested bulls and young elite cows.

In spite of the constrained coefficients of variation, maximized genetic gains were high and were only reduced from .300 to .293 genetic standard deviations per year, when coefficients of variation were reduced from .32 to .16. Adoption of breeding schemes with low coefficients of variation is recommended, because responses are high and coefficients of variation are sensitive to accidental changes in the breeding structure.

Key Words: breeding schemes • restricted maximization of selection response • variance of selection response • simulated annealing

Submitted on June 21, 1993
Accepted on February 16, 1994