Balancing breeding goals in Lacaune sheep
– by T. Byrne
This task within WP7, describes the outcomes of a breeding program simulation in Lacaune sheep using a general selection index model framework. The objectives of this simulation are to:
- Assess the long-term impact of breeding for resilience and efficiency traits in Lacaune sheep populations.
- Test how new genomics data and tools can improve breeding programs and populations faster.
- Produce 20-year forecasts of the productivity gains that can be expected when resilience and efficiency traits are included in the breeding programme.
This simulation forms part of Task 7.4 of the H2020 Small Ruminants Breeding for Efficiency and Resilience (SMARTER) project. Adding Functional Longevity (FL) and Feed Efficiency (FE) to the breeding program results in a significant long-term (20 year) response to selection for both traits, with 0.3 more lactations and an increase of 13.2% in feed efficiency (13.2% less feed). Genomic evaluations increase these responses to 1.3 lactations and 21.3% in feed efficiency. As expected, when adding more traits to the index, the relative trait emphasis and response to selection in other traits in the index reduces, with most noticeable reductions for protein yield (PY) and fat yield (FY). The sensitivity analyses showed that outcomes were most sensitive to changes in the accuracy of genomic breeding values for FL and FE. The emphasis of FL and FE in the index increased significantly, from 11.8% to 17.7%, and 12.7% to 15.3%, respectively, when the accuracy of genomic breeding values increased from 50% to 70%. The emphasis on FL and FE in the index decreased significantly, by similar proportions, when the accuracy of genomic breeding values decreased from 50% to 30%. Adjusting these accuracies resulted in, on average, a 5.2% change in response to selection for PY and FY. For most sensitivity analyses, only slight differences were observed in responses to selection and emphasis of other traits in the breeding program. Genetic parameters like heritability, and genetic and phenotypic correlations affect responses to selection. Considering that for FE a genetic and phenotypic correlation was only estimated with MY, estimation of the correlations between FE and other index traits and re-running of simulations would provide a more complete picture of the implications of adding FE and of the sensitivity of responses to the genetic relationship between FE and other traits. Simulation outcomes from the current breeding program (base scenario) show that LSCS, TA, UD and TP have an unwanted long-term (20 year) response to selection; increasing LSCS and TA, whilst decreasing UD and TP. By changing the FL and FE weights in the index, these unwanted responses to selection can reduce. Revisiting the weights applied to the traits in the TMI, including when traits are added to the breeding goal and carefully setting index weights for these additional traits (like FL and FE), would inform a balanced approach to selection. Furthermore, the use of economic values in the index instead of weights based on desired responses, would deliver real world value of breeders and commercial farmers. These economic values result in an index that represents the profit (or loss) to the farmer and are based on price and cost data and other sources of information.