Titel: Modeling traits in broilers
Projektbeskrivelse
Practical experience have shown that excluding or including maternal effects in models for genetic evaluation in broilers can have dramatic effects on estimated variance components. Inclusion of different types of maternal effects can lead to highly complex models with parameters that are difficult to estimate. This may lead to inaccuracy of model parameters and this may in turn lead to less accurate predicted breeding values. On the other hand exclusion of important maternal effects may lead to biases in predicted breeding values even when such effects are not significant using classical testing methods such as likelihood ratio tests. This is because source of information for different maternal effects and breeding values for direct additive genetic effects are highly unbalanced and often the effects are non-significant, but need to be included in the model to avoid biases or inflation in the predicted breeding values.
Model selection in genetic evaluation, therefore, need to be carried out using extensive model validation procedures based on cross validation in order to assess accuracy, inflation and bias in the predicted breeding values. Maternal effects is an indirect effect and this makes cross validation more complicated that in more standard models used in animal breeding problems. Inflation is an effect that changes the variance of the predicted breeding values away from what should be expected from quantitative genetic theory. This create problems if groups with different amount of inflation are compared and also create problems with proper weighting of different traits in computing multiple trait selection indices. Bias can occur if the average breeding values of groups of individuals differ significantly from what would be expected from quantitative genetic theory.
Breeding programs for broilers obviously focuses on traits measured early in life at or before commercial slaughter age. Such traits include body weight, feed efficiency, carcass composition, meat quality, and survival where the latter can include survival of the embryo during the incubation period. Here we collectively call these traits broiler traits (BT). All BT are affected by genetic and environmental effects and this needs to be taken into account when conducting genetic evaluation in order to ensure the most accurate selection of future parents. Maternal effects are effects on the offspring that are mediated from the dam to the offspring via the egg. This can be due to effects such as egg size, nutrient content of the egg, or antibodies protecting against infections etc. Under practical circumstances the precise cause of maternal effects in specific cases is, generally, unknown.
