Biosecurity in Canadian dairy farms: Unraveling the epidemiology of bovine infectious diseases and refining diagnostic strategies to improve their detection, prevention and control

生物安全是加拿大乳制品行业利益相关者的优先事项，也是加拿大奶农行动倡议的重要组成部分。生物安全旨在限制对加拿大经济、环境和人类健康产生重大负面影响的传染病的传播。生物安全措施可减少将新发疾病引入人群的风险，减少地方病的流行，并防止食品受到传染性病原体污染的风险。关于生物安全措施的信息通常是通过风险评估问卷收集的，有人认为它们有可能预测鸟分枝杆菌亚种。副结核菌群状况。然而，他们的分析并不是直截了当的，因为数据集的维度很高，其中有大量的变量通常是相关的。因此，机器学习是处理这些复杂数据的合适选择。准确的诊断测试和筛查策略对于精确衡量生物安全措施的影响和监测疾病流行至关重要，但在实践中，通常没有完善的参考测试。因此，贝叶斯潜类模型是评估不完善测试诊断准确性的合适方法。对于这个特定的资助周期，我的NSERC-Discovery研究计划的目标是：1)估计MAP、都柏林沙门氏菌和牛白血病病毒在加拿大奶牛群中的流行程度，2)开发并验证分类模型，以确定奶牛群对这些传染病的状况；3)评估在奶牛场检测MAP和S. Dublin的诊断测试和筛查策略的准确性。从长远来看，我的研究计划旨在预防和控制加拿大乳制品行业利益相关者高度重视的传染病，以限制与它们在农场引入或存在相关的经济损失，并最终减少它们对动物健康和福利，公众健康和环境的负面影响。我的研究项目将推进生物安全、流行病学和动物健康方面的现有知识，并帮助乳制品生产商和兽医减少与生产限制疾病相关的经济损失。政府机构将使用新生成的数据来设计或更新结构化的预防和控制项目。我的研究项目是朝着基于传播途径而不是单一病原体的诊断和控制策略迈出的第一步。该框架可扩展到其他传播途径，如空气传播或病媒传播疾病，以促进控制规划的实施。最终，加拿大消费者将受益于获得最高质量的乳制品，这些乳制品是由在卓越的福利条件下饲养的动物生产的，使用的抗菌剂最少。

Biosecurity is a priority for Canadian dairy industry stakeholders and an important component of the Dairy Farmers of Canada ProAction Initiative. Biosecurity aims to limit the transmission of infectious diseases with an important negative impact on the Canadian economy, the environment, and human health. Biosecurity measures reduce the risk of introduction of emerging diseases into a population, decrease the prevalence of endemic diseases and prevent the risk of food contamination with infectious pathogens. Information on biosecurity measures is often collected through risk assessment questionnaires and, it has been suggested that they have the potential to predict Mycobacterium avium subsp. paratuberculosis (MAP) herd status. However, their analysis is not straightforward because of the high dimensionality of the datasets which have a high number of variables that are often correlated. Machine learning is therefore a suitable alternative to handle these complex data. Accurate diagnostic tests and screening strategies are crucial to precisely measure the impact of biosecurity measures and monitor disease prevalence but in practice, perfect reference tests are not often available. Thus, Bayesian latent class models are a suitable methodology to evaluate the diagnostic accuracy of imperfect tests. For this specific funding cycle, the objectives of my NSERC-Discovery research program are to: 1) estimate the prevalence of MAP, Salmonella Dublin and Bovine Leukemia Virus in Canadian dairy herds, 2) develop and validate a classification model to determine the status of dairy herds regarding these infectious diseases; 3) evaluate the accuracy of diagnostic tests and screening strategies to detect MAP and S. Dublin in dairy farms. In the long-term my research program aims the prevention and control of infectious diseases of high priority for the Canadian dairy industry stakeholders, to limit the economic losses associated with their introduction or presence in the farms and, ultimately, reduce their negative impact on animal health and welfare, public health, and the environment. My research program will advance the current knowledge in biosecurity, epidemiology, and animal health, and help dairy producers and veterinarians to reduce economic losses associated with production limiting diseases. Government agencies will use the newly generated data to design or update structured prevention and control programs. My research program is a first step towards the diagnostic and control strategies based on transmission routes rather than single pathogens. A framework that could be extended to other transmission routes such as airborne or vector-borne diseases to facilitate the implementation of control programs. Ultimately, Canadian consumers will benefit by having access to dairy products of the highest quality, produced by animals raised in outstanding welfare conditions with minimal use of antimicrobials.