Modeling maternal effects in later traits in broilers

TY - ABST

T1 - Modeling maternal effects in later traits in broilers

AU - Rome, Helene Julie Sophie

AU - Marois, Danye

AU - Huang, Chyong-Huoy

AU - Jensen, Just

PY - 2020

Y1 - 2020

N2 - The additive genetic effect and the permanent environmental effect of the dam (pem) might affect performances of the offspring. That is particularly true for juvenile performance. Often those effects are ignored in chicken for breeding value estimation (EBV) even if it can affect accuracy of EBV and most likely the inflation of those EBV. In this study focus, we focused on different way to model maternal effects to predict breeding value of late body weight in a commercial line of broilers. Population consisted of males and females farmed in commercial environment. Body weight (BW) have been recorded weekly (week1 to week 5). In a previous study, we showed the importance of fitting into evaluation models fixed effect for dam (i.e. age of the dam) and the pem. Adding the additive genetic effect did not help at later age but was needed at earlier age. In this study we focused on BW at week 5. Our conventional model (“BW5”) was considering fixed and random genetic effects for the individuals as well as fixed effects for the dam, the pem and the maternal genetic effect. Since maternal effect is bigger at earlier age, we studied the possibility to control maternal effect in week5 by adding information from week 1. For two model were tried. The first one (“BW5-BW1”) included the fixed and random effect for maternal effect and a regression of BW at week 1 on BW at week 5. The second model (“BW5-EBV1”) was similar but instead of BW at week 1, the EBV estimated at week 1 was fitted into the regression. The model “BW5” showed an accuracy of 0.49 and an inflation of EBV of 0.77). Model “BW5-BW1” did not show any improvement in term of accuracy and inflation of EBV (accuracy of 0.46 and inflation of 0.74), compare to model “BW5”. On the contrary, model BW5-EBV1 showed an improvement in term of accuracy and most importantly a reduction of the inflation of EBV (accuracy of 0.53and inflation of 0.99). This showed that it is possible to account for maternal effect in later BW by considering juvenile trait into the evaluation model. Other way to improve modeling of maternal effects should be explore. For example a better modeling of the age of the dam may help.

AB - The additive genetic effect and the permanent environmental effect of the dam (pem) might affect performances of the offspring. That is particularly true for juvenile performance. Often those effects are ignored in chicken for breeding value estimation (EBV) even if it can affect accuracy of EBV and most likely the inflation of those EBV. In this study focus, we focused on different way to model maternal effects to predict breeding value of late body weight in a commercial line of broilers. Population consisted of males and females farmed in commercial environment. Body weight (BW) have been recorded weekly (week1 to week 5). In a previous study, we showed the importance of fitting into evaluation models fixed effect for dam (i.e. age of the dam) and the pem. Adding the additive genetic effect did not help at later age but was needed at earlier age. In this study we focused on BW at week 5. Our conventional model (“BW5”) was considering fixed and random genetic effects for the individuals as well as fixed effects for the dam, the pem and the maternal genetic effect. Since maternal effect is bigger at earlier age, we studied the possibility to control maternal effect in week5 by adding information from week 1. For two model were tried. The first one (“BW5-BW1”) included the fixed and random effect for maternal effect and a regression of BW at week 1 on BW at week 5. The second model (“BW5-EBV1”) was similar but instead of BW at week 1, the EBV estimated at week 1 was fitted into the regression. The model “BW5” showed an accuracy of 0.49 and an inflation of EBV of 0.77). Model “BW5-BW1” did not show any improvement in term of accuracy and inflation of EBV (accuracy of 0.46 and inflation of 0.74), compare to model “BW5”. On the contrary, model BW5-EBV1 showed an improvement in term of accuracy and most importantly a reduction of the inflation of EBV (accuracy of 0.53and inflation of 0.99). This showed that it is possible to account for maternal effect in later BW by considering juvenile trait into the evaluation model. Other way to improve modeling of maternal effects should be explore. For example a better modeling of the age of the dam may help.

M3 - Conference abstract for conference

T2 - 71st Annual Meeting of European Federation of Animal Science

Y2 - 1 December 2020 through 4 December 2020

ER -