The effect of different milk pricing schemes on a selection index for South African Holstein cattle

English: A total of 150673 first lactation records of South African Holstein lactating cows were used to estimate the genetic parameters of milk (MY), butterfat (BFY), and protein (PRY) yields, and butterfat (BFP) and protein (PRP) percentages. The data comprising 113 056 dams and 1 429 sires, distributed over 1 205 herds, were collected over a period of 21 years, 1980 to 2000. The study was aimed at the following aspects: to determine the non - genetic factors influencing milk traits for use in the model for a subsequent genetic analysis; to estimate the variance-covariance components and heritabilities of all traits. Finally, to calculate the weighting factors for a selection index and predict the expected response to selection when using different milk pricing systems. Five fixed effects were defined, viz: milking frequency, age at calving, herd, month of calving and year of calving. Analysis of variance indicated that all effects were highly significant (p < 0.0001) explaining on average 64.60%, 62.96%, 64.36%, 15.66% and 23.86% of the total variation for MY, BFY, PRY, BFP and PRP, respectively. Considering the R2 and RMSE as an option, month of calving for all traits, milking frequency for BFP and PRP, and age at calving for PRP showed a minor contribution to the variation and were, thus, excluded from the model for the subsequent analysis. Heritability estimates varied from medium for BFY and PRY to high for MY, BFP and PRP. This indicates that faster genetic progress will be possible through selection of MY, BFP and PRP. The highly positive genetic correlation among yield traits indicates selection for milk yield will result in a favourable response with butterfat and protein yield. However, it will lead to an unfavourable response to selection for the percentage traits. This is due to a negative genetic correlation between milk yield with percentage traits. Using the Desire program, weighting factors and expected response per generation for milk, butterfat and protein yield were calculated under three milk buyers (A, B and C) pricing systems. The standard deviations of alternative breeding objectives and selection indices were also calculated. The weighting factors for milk, butterfat and protein yields were 0.16, 2.41 and 3.24 for milk buyer A, 0.04, 3.43 and 5.70 for milk buyer Band 0.03, 3.48 and 4.37 for milk buyer C, respectively. While the expected response for the corresponding traits, when the intensity of selection equals one, were +546kg, + 16.6kg and + 15.6kg for MY, BFY, PRY respectively. in the ease of milk buyer A, +509kg, +17.1kg and +15.7kg for milk buyer B and +504kg, + I7.3kg and + I5.5kg for milk buyer C. The correlations of indices of B and C with respect to index A were 0.97 and 0.84, respectively. This indicates that the loss of efficiency when selection index Band C were used is 3 % and 16 %, respectively. Rankings of Holstein cows under the three milkpricing systems were different and this indicates that depending on the choice of their breeding goals South African Holstein breeders should follow different selection programs.