35 Refined phenotyping methods for the equine cardiac startle response

Abstract

As a prey animal, the startle response in equids evolved as a survival mechanism. In domesticated horses the startle response has the potential to cause harm to the animal or human handlers. The cardiac response to a novel object due to startle negatively impacts the health and welfare of a diverse range of livestock. Our previous analyses in horses demonstrated significant heritability of 0.615 for change in startle heart rate with maturity. Measures of the cardiac startle response may provide quantitative phenotypes amenable to future genome-wide association studies to identify loci contributing to this important trait. To better understand the startle reaction in horses, we employed novel analytical approaches comparing the cardiac response, recorded by monitors, to behavioral changes following introduction of a novel object in young horses. Following acclimatization in a visually isolated round pen, young horses approach a feed pan where an umbrella is suddenly opened. Data from 74 UF bred horses, primarily AQHA registered, were tested as weanlings. Cardiac phenotypes were constructed using the raw inter-beat intervals recorded 2 s before presentation of the novel object to 10 s post event by a k-means clustering approach. Video recorded behaviors were quantified using a 6-category ethogram including flight distance, activity, orientation to the novel object, and latency to return to the feed pan. Following transformation, behavioral variables were combined into a single score using principal component analysis followed by factor rotation which were compared with the heart rate clusters (accelerator or decelerator categories). PCA and factor rotation yielded 2 scores correlated most strongly to the latency and flight distance variables and captured 55.6% and 32.5% of the total variation across all behavioral variables. For example, horses with a high Factor 1 score displayed a larger flight distance and take longer to return to the pan. In contrast, horses with a high factor 2 score walked and trotted more frequently while in the pen. Cardiac startle response categories (accelerator versus decelerator) were significantly correlated to the behavioral reactions captured in Factor 1 ( P = 0.0003) but not for Factor 2 ( P = 0.5155). Future work will utilize this novel equine cardiac startle phenotype in heritability estimates, and in a Genome Wide Association Study to identify loci contributing to the trait. Developing genetic tests to assess startle response would be useful for breeding selection or specific individual needs of horses and handlers.