Optimization of progeny tests with prior information on young bulls

Algorithms are derived to optimize the number of progeny records per young bull (or boar) in a progeny test when prior EBVs are available. The prior EBV may come from a pedigree index, own record or indicator traits. Progeny testing of the highest ranking young bulls on prior EBV, with an equal number of progeny records per bull is close to optimal even when: (1) some bulls are more promising than others based on prior EBV; (2) the bulls are related; and (3) the cost prices of the bulls differ. The latter depends on the assumed distribution of the cost prices, and the algorithm that was derived to account for the differences in cost prices may still prove a useful safeguard against buying too expensive bulls. The algorithm was extended to the situation where inbreeding was reduced by putting a cost on the average relationship among the selected bulls, i.e., G - kĀ was maximized, where G and Ā are the genetic merit and the average relationship of the selected bulls and k is the cost factor. In an example, the algorithm yielded 11% higher G - kĀ than simply progeny testing the highest ranking bulls based on prior EBV.