Genetic evaluation of Swedish standard-bred trotters for racing performance traits and racing status

Abstract

The currently applied AM-BLUP methodology for genetic evaluations of the standard-bred trotters in Sweden was reviewed. The computational procedure for multiple trait evaluation of the all-or-non trait racing status and five continuous racing performance traits was described in detail. The procedure involves generating expected values by the EM algorithm for unraced horses, which are updated in each round of iteration, in order to facilitate canonical transformation of records. The validity of the procedure was confirmed by comparing selection based on the multiple trait procedure with selection on ST-BLUP (single trait animal model for racing performance), based on records of raced horses only. The MT-BLUP (multiple trait animal model procedure including racing status and performance as two correlated traits) process was shown to remove bias and increase both the accuracy and precision of selection considerably. The genetic parameters of the current model were evaluated by regressing BLUP solutions from the whole data on solutions from subsampled data (method R). The results showed that a slight increase of the currently used heritabilities, maintaining the same covariance structure, resulted in a very close fit to the expected R-values of one.