Guideline to define breeding objective functions, construct selection  indexes and deal with uncertain sires in sheep and goat breeding  programmes

Joaquín Mueller, Tesfaye Getachew, Barbara Rischkowsky, Aynalem Haile. (1/6/2021). Guideline to define breeding objective functions, construct selection indexes and deal with uncertain sires in sheep and goat breeding programmes.

Abstract

The first and crucial step in any breeding programme is to define the breeding goal. Small ruminant breeding programmes usually aim at simultaneous improvement of several traits. Guidelines are given to define multi-trait breeding objective functions by identifying goal traits and calculating their economic values using profit functions. Partial budgeting procedures are explained and exemplified with spreadsheet templates. The concept of discounted genetic expressions (DGE), which allow for goal trait expressions in future generations of a selection candidate, is introduced and a computer tool is provided for their calculation. 2) Genetic and economic improvement through selection of sires and dams for multi trait breeding objectives require the choice of appropriate selection criteria, these are best formalized in selection indexes. Guidelines are given to construct selection indexes which improve or restrict genetic and economic gain in goal traits. Examples of index construction and index evaluation with different measurements on selection candidates are derived and replicated in spreadsheet files. A computer tool is provided to calculate the index weights and their accuracies, the relevance of traits included in the index, the expected genetic and economic gains in breeding goals as well as the outcomes when selection is in stages or when traits are dropped from the index. 3) In many low input breeding systems, the sire of a selection candidate is unknown and the true sire is known to be amongst a limited number of possible sires. In such cases, an uncertain sire relationship matrix can be used in the solution of the mixed models to predict breeding values and thereby increasing the accuracies of breeding values and increasing the eventual genetic gains. A computer tool is provided to construct the uncertain sire relationship matrix and a computer aided procedure with examples is described to calculate breeding values from field data in populations with up to three uncertain sires.