Suitability of simplified milk recording methods for genetic evaluations using test-day models in dairy sheep

Abstract

Nowadays in dairy sheep, test-day records are collected monthly under an alternate morning/evening system. This alternate system was implemented because of the steady increase in recording cost per sheep observed in the last few decades. Furthermore, additional modifications to the testing schemes have been proposed to reduce the costs of milk recording. In dairy sheep, the cost for milk recording is indeed too high compared both with individual outputs and with other dairy species. Therefore, the purposes of the present study were: i) to compare estimated breeding values (EBVs) for milk yield using different simplified testing schemes with a test-day sire model; ii) to evaluate the effect of different testing schemes on the ranking of top rams; and iii) to evaluate the effect of different testing schemes on the genetic progress, based on sire selection. A total of 28,304 test-day records from 4,968 lactations of 2,429 Valle del Belice ewes, collected in 15 flocks between 1994 and 2006, were used for the analysis. To have the same ewes through different testing schemes, it was requested that a ewe have at least six test-day records included in the dataset. The pedigree file included a total of 206 sires. The response variable was the daily milk production. Four different recording schemes were used to estimate breeding values and to make comparisons. The reference scheme considered all data obtained from ewes in the present recording system in which ewes in lactation were tested every month. Three alternative schemes that used less information were considered. A random regression sire model that used a Legendre function of days in milk was fitted. EBVs obtained with alternative recording schemes and under different intensities of selection showed different degrees of Spearman correlation with EBVs obtained using the monthly recording scheme. These correlations ranged from 0.68 to 0.98. The consequence of moving the test-day schedule to one of these schemes will produce a decrease in genetic response to selection due to the associated loss of intensity and accuracy. Consequently, before given any recommendation in this regard, further research should be done to better elucidate the consequences of a change in the scheme of test-day collection.