This work is a preliminary study to develop a selection index for fertility traits in the local Reggiana cattle breed. The study aimed to investigate non-genetic sources of variation for fertility and to identify the best model in terms of variance explained. Moreover, the variation of the target fertility traits in the different parities and months of first service have been considered. The fertility traits under investigation were the interval between calving and first insemination, the interval between calving and conception, the number of inseminations per conception, and the calving interval. The dataset included 22,731 records of 10,502 cows, collected between 1986 and 2019. Four different models were tested: Model 1 included the fixed effects of herd-year-month of first service and parity; Model 2 separately accounted for herd-year and month of first service, in addition to parity. Additionally, Model 1a and Model 2a presented the same effects of Model 1 and Model 2 with the addition of the age at first insemination as linear covariate. The best fertility performances were observed in March and April, whereas for parity effect the best performance was in third lactation. Regarding the Model 1, the coefficient of determination was on average 0.40 for the four traits, and this value increased to 0.48 in Model 1a. Moving to Models 2 and 2a, the coefficient of determination dropped to the average values of 0.21 and 0.27, respectively. Therefore, Model 1a had the best fitting performances for all studied traits, although data editing for age at first insemination was very strict. On the other hand, in Models 2 and 2a the variance absorbed by fixed effects was low, and this could potentially bias the final estimates. Model 1 was therefore the best one in terms of trade-off between data loss and predictivity.
Non-genetic effects affecting fertility traits in local Reggiana cattle
Mancin E.;Sartori C.;Guzzo N.;Mantovani R.
2020
Abstract
This work is a preliminary study to develop a selection index for fertility traits in the local Reggiana cattle breed. The study aimed to investigate non-genetic sources of variation for fertility and to identify the best model in terms of variance explained. Moreover, the variation of the target fertility traits in the different parities and months of first service have been considered. The fertility traits under investigation were the interval between calving and first insemination, the interval between calving and conception, the number of inseminations per conception, and the calving interval. The dataset included 22,731 records of 10,502 cows, collected between 1986 and 2019. Four different models were tested: Model 1 included the fixed effects of herd-year-month of first service and parity; Model 2 separately accounted for herd-year and month of first service, in addition to parity. Additionally, Model 1a and Model 2a presented the same effects of Model 1 and Model 2 with the addition of the age at first insemination as linear covariate. The best fertility performances were observed in March and April, whereas for parity effect the best performance was in third lactation. Regarding the Model 1, the coefficient of determination was on average 0.40 for the four traits, and this value increased to 0.48 in Model 1a. Moving to Models 2 and 2a, the coefficient of determination dropped to the average values of 0.21 and 0.27, respectively. Therefore, Model 1a had the best fitting performances for all studied traits, although data editing for age at first insemination was very strict. On the other hand, in Models 2 and 2a the variance absorbed by fixed effects was low, and this could potentially bias the final estimates. Model 1 was therefore the best one in terms of trade-off between data loss and predictivity.
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
