Precision medicine, which is the application of high dimensional data to refine disease classification, diagnosis, and treatment, is the future of health care in the developed world. This is no less true for veterinary patients, including horses. Horses are a uniquely positioned species to apply precision medicine approaches, with a continually improving annotated genome and a number of syndromic diseases shared with humans. However, additional resources are needed to enhance the ability to move equine medicine into the precision age. This takes on three broad areas: improved phenotype data, expanded omic analysis and big data/computational approaches. This thesis addresses the areas of need to promote progress towards precision medicine through resource development and utilization to demonstrate the application in equine disease.

We begin this thesis by introducing, in Chapter 1, the Pioneer 100 Horse Health Project, a deep phenotype and multiomic resource developed specifically for use in precision medicine studies of equine health and disease. This is followed with Chapter 2 that details the application of this resource for improving the understanding of drivers of disease in Equine Metabolic Syndrome (EMS). Through deep endocrine phenotyping, married with plasma metabolomic and fecal microbiota, we establish a clear relationship between environment-host-microbiota factors that promote EMS. Next, in Chapter 3, we further investigate the role of environmental interaction and disease in the neurologic condition, equine neuroxaonal dystrophy (eNAD). Here, we evaluate the impact of vitamin E deficiency on axonal health in juvenile horses to refine the use of a previously validated biomarker, phosphorylated neurofilament heavy, in the supportive diagnosis of eNAD. We show that vitamin E deficiency alone is enough to drive subclinical axonal damage, even if clinical eNAD does not ensue. Finally, in Chapter 4, we apply machine learning models to proteomic data sets for serum and cerebrospinal fluid (CSF) to discover disease informative biomarkers for equine neurodegenerative disease. Here, we demonstrate the use of two- and three-plex protein biomarkers that can accurately predict horses with eNAD from both neurologically normal horses, and those with the disease mimic condition, cervical vertebral compressive myelopathy.

The development of equine precision medicine resources, detailed in this thesis, are a key step forward in this discipline. To that end, availability of these data and methods through publication and repository submissions are an important step towards that goal. While validation in external cohorts will be necessary in the future, the data presented in Chapters two, three and four independently advances equine health and welfare.

Lastly, this thesis includes an addendum providing the details for biobank generation in the Functional Assembly of Animal Genomes (FAANG) consortium using two Thoroughbred stallions. Improvements to the annotation run parallel to the development of precision medicine approaches in the horses. By more precisely annotating the equine genome to include detailed information about regulatory element and tissue specific gene expression, we greatly enhance the resolution for health, disease and trait investigation in the horses.

Collectively this work demonstrates the both the approach and application of precision medicine in the horses and acts as a blueprint for future progress in the field.