Combining Threshold, Thurstonian and Classical Linear Models in Horse Genetic Evaluations for Endurance Competitions.
Authors: Cervantes Isabel, Gutiérrez Juan Pablo, García-Ballesteros Silvia, Varona Luis
Journal: Animals : an open access journal from MDPI
Summary
# Editorial Summary Genetic evaluation of endurance horses traditionally relies on race completion times and finishing ranks, but this approach systematically excludes the 35% of horses that fail to finish—a practice that can bias breeding value predictions and skew selection decisions. Spanish researchers analysed 6,135 competition records from 1,419 horses (predominantly Arab crosses) using a novel multitrait genetic model that simultaneously incorporated finishing time, placement rank, and whether horses completed the race at all, employing threshold, Thurstonian, and classical linear methodologies to capture the full spectrum of performance outcomes. Heritabilities were modest (0.06–0.17 depending on trait and model), with race time and rank showing near-perfect genetic correlation and therefore limited independent value, whilst completion status emerged as the most genetically distinct trait. The findings suggest that endurance breeding programmes should prioritise completion capacity as a separate selection criterion rather than relying exclusively on finishing metrics from completers, and that modern statistical approaches can accommodate both categorical and continuous performance data to provide more robust breeding value estimates than traditional methods that discard non-finishers entirely. For practitioners advising breeders, this work emphasises that genetic potential for endurance includes the ability to start and finish, not merely the speed of those horses that do.
Read the full abstract on PubMed
Practical Takeaways
- •Genetic evaluations for endurance horses should include placing (completion) status alongside performance traits to avoid bias from excluding non-finishers
- •Breeders can focus on rank or time alone rather than both, as these traits are genetically very similar in endurance horses
- •Selection for endurance performance should treat completion ability as a separate genetic trait from speed/ranking performance
Key Findings
- •Heritabilities for endurance traits were moderately low, ranging from 0.06-0.17 depending on trait and model used
- •Racing time and rank showed high genetic correlation (inverse relationship), suggesting rank could replace time in future evaluations
- •Placing (completion status) was genetically independent from time and rank traits
- •35% of endurance competition records involved non-placed horses, indicating high non-completion rate requiring threshold model consideration