Combining Threshold Thurstonian and Classical Linear Models in Horse Genetic Evaluations for Endurance Competitions

摘要

Endurance competitions are carried out across the countryside worldwide, where the adapted physical and metabolic conditions of the horses are essential to satisfactorily end the competition. Horse breeding programs usually use the rank and time to improve the performance of horses in competitions. It is also relevant to analyse the placing trait i.e., whether or not a horse finishes the race. The discontinuous nature of rank and placing traits require special methodologies to deal with them. Here we have used 6135 endurance records from 1419 horses, with a pedigree containing 10,868 animals, to develop a multitrait model with a new free software tool (GIBBSTHUR). The obtained results suggest that it is possible to ignore the race time and use rank and placing to perform the genetic evaluation in endurance horse populations.