Comparison of Procedures Used for Evaluating Sires Used in Artificial Insemination

Abstract
The growth of artificial insemination has increased the importance of obtaining sires with high genetic merit for production since the potential number of daughters of a single sire is large. This development has increased the importance of evaluating sires according to estimated genetic worth. The most valuable tool by which sires can be evaluated is the progeny test. Many variations of the progeny test have been proposed. The "best" of these are very time consuming unless modern high speed computing equipment is available. The purpose of this study was to compare the evaluation of sires based on a method currently used at the New York Dairy Records Processing Laboratory (which was assumed to be the "best") with evaluation by various other methods, some of which require much less computational effort. The New York system is essentially a regression of daughter records expressed as a deviation from an adjusted stablemate average on true breeding value of the sire. Other methods included daughter averages; deviations from herd season averages, stablemate averages, and stable-mate averages adjusted for numbers; and percentage of records exceeding breed season averages, herd season averages, stablemate averages, and stablemate averages adjusted for numbers. All methods were employed for first records only and for average records of daughters. The analysis of daughter records of 198 Holstein A.I. sires indicates: (1) None of the alternative procedures is accurate enough for general use in sire evaluation if it is assumed that the New York procedure is "best". (2) If, however, at least 50 1st records are available for each sire group the procedures utilizing averages of daughter averages are nearly as good as the standard procedure. (3) The computationally easy percentage systems may have merit if computational costs are high and if at least 50 daughter records are available per sire group.

This publication has 1 reference indexed in Scilit: