Effective control of avian diseases in domestic populations requires understanding of the transmission dynamics facilitating viral emergence and spread. In 2016–17, Italy experienced a significant avian influenza epidemic caused by a highly pathogenic A(H5N8) virus, which affected domestic premises housing around 2.7 million birds, primarily in the north‐eastern regions with the highest density of poultry farms (Lombardy, Emilia‐Romagna and Veneto). We perform integrated analyses of genetic, spatiotemporal and host data within a Bayesian phylogenetic framework. Using continuous and discrete phylogeography, we estimate the locations of movements responsible for the spread and persistence of the epidemic. The information derived from these analyses on rates of transmission between regions through time can be used to assess the success of control measures. Using an approach based on phylogenetic–temporal distances between domestic cases, we infer the presence of cryptic wild bird‐mediated transmission, information that can be used to complement existing epidemiological methods for distinguishing transmission within the domestic population from incursions across the wildlife–domestic interface, a common challenge in veterinary epidemiology. Spatiotemporal reconstruction of the epidemic reveals a highly skewed distribution of virus movements with a high proportion of shorter distance local movements interspersed with occasional long‐distance dispersal events associated with wild birds. We also show how such inference be used to identify possible instances of human‐mediated movements where distances between phylogenetically linked domestic cases are unusually high.

中文翻译：

---

2016-17年意大利H5N8高致病性禽流感疫情时空重建及传播动态

有效控制国内禽类疾病需要了解促进病毒出现和传播的传播动态。2016-17 年，意大利经历了由高致病性 A(H5N8) 病毒引起的严重禽流感疫情，影响了大约 270 万只家禽的住宅，主要集中在家禽养殖场密度最高的东北部地区（伦巴第、艾米利亚-罗马涅和威尼托）。我们在贝叶斯系统发育框架内对遗传、时空和宿主数据进行综合分析。使用连续和离散的系统发育地理学，我们估计了导致流行病传播和持续的移动位置。从这些区域间传播率随时间变化的分析中得出的信息可用于评估控制措施的成功与否。使用基于国内病例之间的系统发育-时间距离的方法，我们推断存在隐秘的野生鸟类介导的传播，这些信息可用于补充现有的流行病学方法，以区分国内人群内的传播与跨野生动物-国内界面的入侵，兽医流行病学中的一个常见挑战。流行病的时空重建揭示了病毒传播的高度倾斜分布，其中短距离局部传播的比例很高，偶尔还夹杂着与野生鸟类相关的长距离传播事件。我们还展示了如何使用这种推论来识别人类介导的运动的可能实例，其中系统发育相关的国内病例之间的距离异常高。