Mate selection for multi-breed beef cattle populations

Mate selection is an attractive breeding strategy for multi-breed beef cattle populations, as it allows selection and crossbreeding to be exploited simultaneously. Two components are required for mate selection; an objective function called a mate selection index (MSI) which describes economic gain as a function of selection and mate allocations (or 'mate selections'), and a mate selection algorithm which finds the mating set which maximises the MSI.

One concern with mate selection is that as breeding decisions are based on progeny merit only (the next generation), short term genetic merit may be maximised at the cost of longer term genetic merit. To address this concern, LAMS (Look Ahead Mate Selection) schemes have been proposed. In LAMS schemes, some weight is given in the MSI to the merit of predicted grand-progeny in mate selection decisions.

To date, there have been few studies predicting the benefits of mate selection in multi-breed beef cattle populations. The impact of mate selection, especially LAMS schemes, on population structure and longer term genetic gain is unknown. Further, the advantages of mate selection when the genetic model of multi-breed populations is extended beyond selection and heterosis to include within breed dominance variance is unknown.

This thesis was designed to investigate three issues:

1. What population structures will emerge when mate selection (including LAMS schemes) is applied in multi-breed populations?
2. What is the effect of longer term implementation of LAMS schemes on genetic merit of multi-breed populations
3. Does mate selection significantly improve the genetic value of a multi-breed population when individual dominance is included in the MSI (relative to other breeding strategies, such as selection followed by mate allocation)?

To investigate the first issue, a deterministic evaluation of mate selection was considered. Mate selection was at the level of genetic groups, where genetic groups were defined by breed composition and selection history. The initial population structure consisted of breeds A and B. Genetic merit of genetic group l was y_l =  µ_l + g^A_lm_A + g^B_lm_B + g^A_jd_ABg^B_k + g^B_jd_ABg^A_k, where g^A_l, the proportion of breed A in progeny genotype l, is calculated as g^A_l = 1/2(g^A_j + g^A_k), where g^A_j and g^A_k are proportions of breed A in parental genotypes j and k respectively. The proportion of breed B in progeny genotype l is calculated similarly. Parameter d_AB, is the F₁ heterosis when breeds A and B are crossed, and µ_l is the mean additive breeding value for progeny genotype l.

The aim of each breeding schemes was to create a new herd in the first generation by importing sires and dams of breeds A and B, and then breed a further generation of progeny in the home herd. The second generation of progeny could be created either by importing more sires and dams, or using sires and dams bred in the home herd in the previous generation. It was assumed no selection was occurring within the foreign populations. Breeding schemes were compared by using the sum of progeny merit over the two generations, for a range of heritabilities and heterosis values. Schemes evaluated were structured crossbreeding (F₁ and F₂ designs), progeny merit each generation as mate selection criteria (PROG_DET), or cumulative merit over two generations as mate selection criteria (CUMUL_DET). CUMUL_DET was a LAMS scheme, as merit of the generation beyond the progeny generation was considered in the mate selection decisions. At all parameter values, PROG_DET and CUMUL_DET gave better cumulative merit than structured crossbreeding designs. This was a result of selecting sires and dams from the whole population rather than within genotypes or genetic groups. The advantage of mate selection strategies over structured strategies was greatest when h²=0.6; up to 1.9% for cumulative merit from CUMUL_DET over the best F₁ cross.

CUMUL_DET gave slightly greater cumulative merit than PROG_DET in all scenarios, with the advantage increasing with greater heritability or greater selection intensities. The advantage of CUMUL_DET over PROG_DET was from allocation of a proportion of purebred matings in the first generation in the home herd in CUMUL_DET. In the second generation, purebred progeny from these selected purebred parents could be crossed to breed F₁ grandprogeny with maximum heterosis and cumulative additive merit from selection. The proportion of purebred matings in the home herd in the first generation increased with heritability (1% of all matings in the first generation when h²=0.1 and 11% when h²=0.6).

To investigate the second issue, response from implementation of tactical (at the individual animal level) LAMS schemes was investigated. The genetic model used included additive breed and maternal effects, and direct and maternal heterosis. For this model, progeny merit of an individual progeny from a mating was PM_i = 1/2(p^T_s + p^T_d)a + p^T_sDp_d + p^T_da_m + p^T_mgsD_mp_mgd where vector p represents the proportion of genes of each breed in sire (s) or dam (d), maternal grand sire (mgs) or maternal granddam (mgd), a is a vector of fixed breed effects (breed means for each breed), a_m is a vector of fixed maternal breed effects, D is a breed x breed matrix of direct heterosis effects, and Dm is a breed x breed matrix of maternal heterosis effects.

Schemes investigated included mate selection with progeny merit only in the MSI (PROG strategy) and mate selection with progeny merit and grandprogeny merit equally weighted in the MSI (LAMS strategy). An additional scheme, COMP, was evaluated which used the contribution of the mating set to progeny with an optimal composite genotype as the MSI. Schemes were assessed using simulation of a single trait (yearling weight) over six generations of a four breed population. It was assumed no selection was occurring in foreign populations. With LAMS, a proportion of matings each generation in all simulations were allocated to breed F1 dams, improving the merit of the next generation, as a three breed cross could be created. LAMS gave the highest cumulative progeny merit over six generations; 20kg greater than PROG, and almost 70kg greater than COMP.

COMP created a population of optimum composite animals after six generations of breeding, with no variation in breed composition among individuals in the population (all animals had the optimum composite genotype). This strategy would be useful for commercial beef producers aiming to produce uniform lines of turnoff progeny.

The performance of LAMS was also evaluated when the genetic model was extended to include individual additive breeding values...