Predictability of adult Show Jumping ability from early information: Alternative selection strategies in the Spanish Sport Horse population

The purpose of this study was to estimate the genetic parameters and to evaluate whether performance results of young horses are good predictors for later sport performance, in order to propose alternative selection strategies to assess adult show jumping (SJ) ability. The data included 84,600 performance records from 4011 Spanish Sport Horses. Two different bivariate animal models with genetic groups were compared, using rankings of young and adult horses as the dependent variables. The models included the age as covariate, and gender, event, type of penalty scale and competition level as fixed effects. The horse*rider interaction and the animal permanent environment random factors were tested to find the most suitable model (A vs B, respectively). The heritability values obtained were in the low range, between 0.11 +/- 0.01 in adult horses and 0.17 +/- 0.01 in young horses. The results also showed high genetic correlations (>= 0.84 +/- 0.04) between results early and later on in life. With model A, the prediction of the adult SJ ability was higher (Pearson correlation between predicted and real data: 0.33 +/- 0.01) than with model B (0.31 +/- 0.01). The implementation of alternative strategies, such as two-stage selection, could increase the expected genetic gain significantly (over 50%) compared to other strategies. This strategy could promote genetic progress, without any loss of accuracy and little additional cost.