Net feed intake (NFI) is the residual portion of daily feed intake (DFI) not explained by growth or maintenance requirements. The NFI phenotype (NFIp) is based on a 70-day test period where DFI and fortnightly weights (to calculate average daily gain (ADG) and maintenance as metabolic mid-weight (MMWT)) are measured. Recording NFIp is costly, and shortening the test length would be advantageous. However, research has shown that ADG cannot be accurately measured from a shortened test. Genetic NFI EBVs (NFIg) were calculated using DFI EBV adjusted for ADG and MMWT EBV and were shown to have a Pearson correlation of 0.99 with the NFIp EBV from 3,088 Angus steers. The regression slope between NFIg and NFIp EBVs was 1.14. Alternative NFIg models where growth and maintenance requirements were obtained from BREEDPLAN live weight traits instead of live weights recorded in the test period, demonstrated high Pearson correlations (r=0.87 to 0.93) and regression slopes between 0.63 and 0.97 with NFIp EBVs. Results suggest that genetic NFI EBVs can be obtained, with growth and maintenance requirements being determined from BREEDPLAN live weight traits. This provides the opportunity to determine if the length of the test to measure DFI can be shortened, reducing the cost of recording NFI per animal.

BREEDPLAN net feed intake (NFI) EBV is derived from a phenotypic regression based on a 70-day feed intake test. Genetic NFI (NFIg) EBVs have been proposed as an alternative EBV and this recent development may also allow for a shortened feed intake test period. This study used feed intake records of 3,088 Angus steers from the full 70-day test and compared them to daily feed intake (DFI) from shortened test periods. Results showed DFI from shortened test periods had similar means but increased phenotypic variation. Phenotypic correlation with DFI from the full test period decreased as the test period decreased in weekly intervals and ranged between 0.75 and 0.99. NFIg EBVs were predicted using DFI from different length tests. The mean of all NFIg EBVs was close to zero, but the EBV standard deviation increased as the test period decreased. Pearson correlations between NFIg EBVs from a full test period and reduced test periods ranged between 0.73 and 0.99, the regression slope of NFIg from reduced test periods on NFIg from the full test period ranged between 0.73 and 0.95, and the bias ranged between 0.00 and 0.02. These results indicate that as the test period decreases, the spread of EBVs increases, resulting in extreme animals having overestimated NFIg EBVs. A shortened DFI test period could be used to estimate NFIg EBVs.

Optimum polygenic and genomic weights enhance the accuracy of breeding value estimates in single-step genomic evaluations. This study estimated the contribution from marker information to total additive genetic variation referred as λ using an extended single-step model in a multi-trait variance component estimation based procedure using data for six Australian Angus carcase traits. The λ for these traits ranged from 0.54 (for carcass intramuscular fat) to 0.79 (for carcass eye muscle area). Heritabilities were similar between the pedigree only and the extended single-step multi-trait model when using the total genetic variance, and ranged from 0.37 (for carcass rib fat) to 0.53 (for carcass weight), suggesting that the single-step model did not explain more genetic variance than pedigree based models. Results suggest that the scalar λ in the current single-step routine evaluation could be replaced by an extended single-step model allowing for different proportions of the additive genetic co-variance explained by markers for all elements of the genetic co-variance matrix.

The conversion of feed into usable products, also known as feed efficiency (FE), is important given the necessity to increase quality food production. This concept is also important given the environmental sustainability and profitability of the beef production systems. The present thesis analysed different aspects involving FE and their inclusion in beef cattle breeding programs. Thus, this work aimed to increase the understanding of genetic variability of the FE traits and the inclusion of residual feed intake (RFI) into genomic assisted beef cattle breeding programs.

The first research chapter was oriented to better understand the genetic variability of the main FE traits and their association with growth and meat quality traits in an Angus population. The analysed FE component traits were daily weight gain, metabolic midweight, average of daily, feed intake, feed conversion ratio, RFI, and residual gain (RG), evaluated during a 70-day feedlot test period. In this chapter, it was concluded that selection on RFI would have a negative impact in growth and carcass traits. Therefore, it was suggested that selection only for RFI would have negative impacts on growth and meat quality in the studied population.

The second research chapter compared repeatability (REPM) and random regression models (RRM) on feed efficiency component traits during the feedlot test period. Unlike the REPM, the RRM can accommodate changes in parameters of those traits over duration of the test period. First-order RRM applied to body weight (BW) and average daily feed intake (ADFI), shown that genetic parameters tend to change during the feedlot period. By comparing these models, it was concluded that ignoring the change in parameters, regardless of feed costs, resulted in a loss of selection response of approximately 3%

The third research chapter analysed REPM and RRM when using genomic information. Genomic variants associated with BW and ADFI variation were identified and a change of their effect during the feedlot test period was evaluated. For both traits, RRM presented the best fit and only one genomic region was detected with a constant effect throughout the 70-d feedlot test period. For BW and ADFI, the strongest associated variants were rs43350564 and rs109326204, located on chromosomes 20 and 5, respectively. These identified SNP may help to unravel the biology of FE and can be used for more accurate genomic prediction of breeding values.

The final research chapter evaluated various realistic multi-trait selection strategies incorporating RFI, as well as including the use of genomic selection (GS). Here it was concluded that selecting on RFI via a genomic test yielded the largest increase in selection accuracy and overall responses of 0.48 and 64.9%, respectively. Finally, it was concluded that including feed efficiency increased the $ net return without compromising meat quality and cow condition score in the beef cattle breeding programs.

It is important in single-step genetic evaluations to use appropriate lambdas (λ) for calculating weighted average of NRM (numerator relationship matrix) and GRM (genomic relationship matrix) in joint relationship matrix. λ is usually estimated using a single-trait cross-validation procedure. However, it can be shown that a univariate single-step model applying a scalar λ is simply a condensed form of an extended model containing two genetic factors, factor H~N(0, H) and factor A~N(0, A), where the partitioning of the total genetic variance reflects λ. For multivariate single-step genetic evaluation, this model condensation implies that all involved genetic variances may yield the same λ, which is highly unlikely. Hence, it is required to estimate λ by accounting for its heterogeneity using the extended model for variance component estimation. This study used an extended single-step model to estimate variances and λs for calving difficulty (CD), gestation length (GL), and birth weight (BW) using Australian Angus data. A total of 129,851 animals with 45,575 genotypes were analysed. Initial variances obtained from a pedigree-only model were then used as starting values for the extended single-step model assigning 90% of the genetic variance to factor A and 10% to factor H. Since CD is a categorical trait with three categories, a threshold model-Gibbs sampling method was used to estimate variances. Heritability estimates for the extended single-step model were very similar to those from the pedigree only model implying that the single-step model was not explaining more variation in the data than the pedigree only model. For CD, GL, and BW, the total heritability estimates were 0.39 ± 0.04, 0.68 ± 0.02, and 0.44 ± 0.01, respectively. For the same traits, the total maternal heritability estimates were 0.17 ± 0.02, 0.11 ± 0.01, and 0.09 ± 0.01, respectively. In contrast, to the Gibbs sampling starting values, the genetic variance was partitioned between A and H such that direct genetic λ estimates for CD, GL, and BW were 0.36 ± 0.05, 0.62 ± 0.03, 0.75 ± 0.03, respectively. Maternal genetic λ estimates ranged from 0.01 ± 0.01 (for BW) to 0.05 ± 0.01 (for CD). The results imply that λ values are heterogeneous in multivariate single-step genomic evaluation. Further studies are needed to investigate the consequences of using heterogenous λ values for direct genetic and maternal genetic components in multivariate single-step evaluation in terms of model dimensions, solver convergence rate, and model forward predictive ability.