Breeding for Longevity in Dairy Cattle

Abstract

<p>In a genetic evaluation of dairy cattle, it is usually assumed that genetic correlations between traits are constant over time. However, dairy producers cull cows due to various reasons and these reasons could vary over time or across lactations. Due to a change of trend of culling reasons, there might be a change of the survival trait over time and across lactations. Subsequently, there is also a change of genetic correlation of survival with production, fertility, type and health traits over time. As a result, accuracy of genetic evaluation might be affected. Thus, we have designed a genetic evaluation study to address changes of the survival trait over time and thereby improve survival in dairy cattle breeding programs. Therefore, this thesis outlined four experimental chapters.</p><p> The first chapter of the thesis describes the major culling reasons and thereby analyses trends of culling reasons and age at culling over time. A total of 2,452,124 records (2,140,337 Holstein and 311,787 Jersey cows) were used to identify main reasons of culling, trends of culling reasons over time, using a binary logit model, and trends in age at culling of Australian dairy cows. The study showed that low production has sharply declined as a culling reason over the 21 years period whereas infertility had an increasing trend over time. The change of culling reasons over time revealed that there could be a change of genetic correlations of survival with production, fertility and type traits over time.</p><p> The second chapter explored the effect of changing correlations of survival with production, fertility and type traits on genetic evaluation of dairy cattle using simulated data. A constant genetic correlation and a changing genetic correlation scenarios were used in a multi-trait sire model to estimate variance components, heritability, correlations, accuracy and bias over time. A suitable model for practical application was selected based on accuracy and bias of genetic evaluation. The genetic correlation of survival with protein changed considerably over generations. The genetic correlation of fertility and survival also changed over generations. A model that accounted for changing genetic correlations over time provided less biased estimates of residual survival breeding value over generations. Further, this model increased the efficiency of the residual survival breeding value in predicting breeding value for involuntary culling.</p><p> Chapter 3 of the thesis was a follow up simulation study and proposed method to account for changing genetic correlations over time of survival with other traits using a random regression model (RRM). The RR model was compared against a model that does not account for a change of correlations over time. High accuracy and less bias were obtained for production, fertility and type traits over time using RRM. The correlation between true breeding value of involuntary culling and estimated breeding value of residual survival was significantly improved with RRM compared with a model that does not account for a change in the trait over time. The slope of the regression of true breeding value on estimated breeding value of residual survival substantially improved when using a RRM.</p><p> The last chapter dealt with optimizing dairy cattle breeding program using survival as a different trait across lactations. Index selection was compared where survival was modelled as either the same trait or as different traits across lactations. Based on these models, selection response was evaluated based on recording survival information at first lactation only versus survival over the first three lactations. A small difference in response was achieved when survival was assumed to be same trait or as different trait across lactations. However, the response to selection was significantly improved when information on later lactations was included. Higher economic gain per unit index was achieved with the inclusion of information on survival in later lactations than survival at first lactation only and more response was achieved particularly for survival in later lactations, which might be a trait more closely related to the ability to avoid involuntary culling. In conclusion, our study suggested that survival and its correlation with other traits changed over time and across lactations. Accounting for this change with appropriate genetic model is important for genetic improvement of survival in a breeding program.</p>