Expected responses to different strategies of selection to increase heat tolerance assessed by changes in litter size in rabbit.

Response on litter size (LS) to selection based on EBV of the own trait and several indexes including EBV of 1 or more parameters of a reaction norm model (RNM) was evaluated by simulation. The RNM described animal's performance under different environmental temperatures assuming that this has an animal-specific effect on LS beyond an animal-specific temperature threshold (i.e., it had 3 parameters: intercept [Int], slope [Slp], and threshold [Thr]). Litter size data were generated assuming this model under 2 environmental conditions (comfort [CC] and comfort-to-heat [CtHC]). Variance components for Int, Slp, Thr, and LS were taken from literature. The initial base population consisted of 125 females and 25 males. Ten generations of selection were conducted keeping constant the population size. Eight different selection criteria were considered, depending on both the evaluation model and the index combining EBV for 1 or more parameters of the RNM. In 1 case selection was based on EBV of the own trait predicted by using a repeatability animal model. In the other 7 cases the genetic evaluation was conducted using the same model as that used in the simulation. For each scenario 25 replicates were conducted and response to selection was assessed within replicate as the difference between generations in the average of trait. Results indicate that, under the studied conditions and for the used genetic parameters, selection based on the observed trait seems to have the same effect as selection based on some index, including EBV of the RNM parameters. In addition, response to selection could be greater under CtHC rather than CC. Animals selected exclusively for EBV of the Slp and Thr are not expected to have good performance under CC. Under CtHC, selection for LS has a major response on those parameters, whereas response on Int was very small. Under CC, response to selection on LS is mainly determined by a change in Int, whereas Slp does not change and Thr slightly increases. Selection based on EBV of Slp, Thr, or on an index including both seems to have the same effect on the trait and it was doubly effective in modifying the shape of the RNM under CtHC than under CC. Selection based on EBV of Thr and Slp does not seem to lead to any response in LS. Selection based on EBV of Int seems to have no effect on the trait or on animal's tolerance to heat under CtHC, but it would lead to a positive response in LS under CC.