Adequacy of genetic evaluation of dairy cows for milk yield using different testing schemes

Alternative milk recording schemes, as alternate a.m.-p.m. (AT) procedures, have been proposed in order to reduce recording costs per cow thereby increasing the number of cows and young bulls that can be tested. The objective of this paper is to compare predicted breeding values using alternative recording schemes for estimating lactation yields, specifically for dams of sires. Over 4 million test day records for milk yield of 306,148 Italian Friesian cows collected in 649 herds in the province of Milan (Italy) between 1990 and 1997 were used. The simulation of alternative testing schemes was carried out using A4 single milking data. The alternative schemes to estimate daily milk yield from AM and PM data, were: a) alternate scheme using a doubling; b) alternate scheme using a correction for milking interval applying (MI) direct or inverse factors; c) non-alternate scheme using a correction for MI applying inverse factors; d) random scheme using a correction for MI applying inverse factors. A mixed model procedure to estimate breeding values was used. The rank correlation between EBV with A4 method and other testing schemes resulted around 99% at the population level and around 89% with regard to the top 1% of cows. About 10% of cows ranked with A4 method were excluded from the list of the top 1% when alternative testing schemes were applied. A negligible effect on the reduction of selection intensity for the dams of sires path was found.