"Residual feed efficiency (rFE) is an increasingly important novel trait for maintaining the economic and environmental sustainability of dairy cattle production. Given that rFE is a derived trait based on the key energy source trait, dry matter intake (DMI), and key energy sink traits like milk energy (MILKE) and metabolic body weight (MBW), a more thorough quantitative genetic study of rFE would provide important baseline knowledge for dairy cattle management. A dairy consortium dataset focused on eventually developing genetic evaluations for rFE in US Holsteins, has been collected on DMI, MILKE, and MBW records from nearly 7000 dairy cattle in four countries. My dissertation was designed to address some motivating quantitative genetic research questions that could be addressed from the analyses of this dataset. In Chapter 2, I reassessed the merit of using residual feed intake (RFI), defined as the estimated residual from regressing DMI on energy sink traits, as a measure of rFE, given that it is most commonly used in dairy cattle breeding research. We proposed a multiple trait (MT) modeling strategy involving all of the component traits, demonstrating that the use of the Cholesky Decomposition (CD) on the genetic and residual variance-covariance matrices lead to a more potentially flexible quantitative genetic approach to modeling rFE compared to RFI. The advantages of this MT approach were confirmed by simulation when the genetic versus residual relationships between energy sink and source traits were rather divergent. However, there appeared to be no meaningful differences in quantitative genetic inferences when applied to the dairy consortium dataset. Similar conclusions on non-distinctions between the two models for genome wide association (GWA) analyses were also drawn in Chapter 4 although the MT GWA analyses shed further light that quantitative genetic inferences on DMI are distinctly independent of those on rFE. I also investigated heterogeneous genetic relationships across environments for rFE with two broadly different approaches. In Chapter 3, I extended the MT model to discover substantial heterogeneity in genetic and residual partial efficiencies and variance components defining rFE as functions of environmental and management factors. Subsequently, I inferred upon genotype by environment interaction at the genomic level across environment conditions in Chapter 5, determining that some genomic regions are sensitive to environmental covariates such as average production and temperature."--Pages ii-iii.
Read
