Estimation of genetic value of showjumping horses from the ranking of all performances in all competitions

SUMMARY: In the genetic evaluation of showjumping horses, the trait of interest is showjumping ability. Since this cannot be observed directly, traditional genetic evaluation systems have used various measures of the relative ability of horses, such as money won or points awarded in competition. However, such analyses have limitations for the genetic evaluation of a population of performing horses, as they fail to take account of any performances which do not earn points (or money), so that there is no record of poor performances in competition and no adjustment is made for the total number of competitions in which each horse performed. There is no satisfactory way to accommodate the number of ‘starts' in analysis so that the effect of the total number of starts is appropriately reflected in the accuracy of the genetic evaluations. This paper presents a method which overcomes the above limitations by using information from all horses in all competitions to estimate the variance components (for random effects) and the BLUP breeding values using an animal model. Horses were ranked from first to last in each competition and these ranks were then converted to normal scores, which served as a measure of the relative showjumping ability. The method was applied to 160084 records from 6786 horses in 7313 showjumping competitions in Ireland in the 4-year period 1993-1996. Components of variance for additive genetic, permanent environment and temporary environment were estimated using REML-VCE (Groeneveld 1994). The heritability of individual performances was estimated as 0.08, the repeatability within animal was 0.20, and single trait animal model BLUP estimates were obtained for all performing animals and their relatives using the PEST package (Groeneveld and Kovac 1990; Groeneveld et al. 1992). The BLUP estimates of breeding values have a distribution which is approximately normal, and more importantly, the use of all records from all competitions enables a high accuracy to be achieved for horses that have a large number of performances and/or performing relatives.