Integrating genetic and epidemiological data to determine transmission pathways of foot-and-mouth disease virus
Integrating genetic and epidemiological data to determine transmission pathways of foot-and-mouth disease virus
复制标题
DOI：
10.1098/rspb.2007.1442
复制
发表时间：
2008-04-22
影响因子：
4.7
通讯作者：
Haydon, Daniel T.
中科院分区：
文献类型：
--
作者：
Cottam, Eleanor M.;Thebaud, Gael;Haydon, Daniel T.
估计详细的传播树，反映疾病暴发期间受感染的个人或人群之间的关系，往往提供了宝贵的见解，疾病传播的性质和潜在的流行病学过程的整体动态。这些树可能基于与感染时间和传染性有关的流行病学数据，或显示从受感染个体分离的病原体的遗传相关性的遗传数据。由于其固有的高突变率，遗传数据在估计病毒病原体的传播树中变得越来越重要。在这里，我们提出了一种最大似然方法，允许流行病学和遗传数据结合在同一分析中，以推断可能的传播树。我们采用这种方法，从20个农场感染的数据在2001年英国口蹄疫爆发，使用完整的病毒基因组序列从每个感染的农场和信息时，农场首次估计已开发的临床疾病，当这些农场的牲畜被扑杀。在实施国家运动禁令之前，由于动物运动而引起的已知感染联系导致与遗传数据一致的树木数量从41，472减少到1，728，其中只有4棵代表使用模型计算的总可能性的95%以上。这些树在几个方面不同于在遗传数据可用之前构建的树。
Estimating detailed transmission trees that reflect the relationships between infected individuals or populations during a disease outbreak often provides valuable insights into both the nature of disease transmission and the overall dynamics of the underlying epidemiological process. These trees may be based on epidemiological data that relate to the timing of infection and infectiousness, or genetic data that show the genetic relatedness of pathogens isolated from infected individuals. Genetic data are becoming increasingly important in the estimation of transmission trees of viral pathogens due to their inherently high mutation rate. Here, we propose a maximum-likelihood approach that allows epidemiological and genetic data to be combined within the same analysis to infer probable transmission trees. We apply this approach to data from 20 farms infected during the 2001 UK foot-and-mouth disease outbreak, using complete viral genome sequences from each infected farm and information on when farms were first estimated to have developed clinical disease and when livestock on these farms were culled. Incorporating known infection links due to animal movement prior to imposition of the national movement ban results in the reduction of the number of trees from 41 472 that are consistent with the genetic data to 1728, of which just 4 represent more than 95% of the total likelihood calculated using a model that accounts for the epidemiological data. These trees differ in several ways from those constructed prior to the availability of genetic data.