INRA, UR346 Epidémiologie Animale, Saint Genès Champanelle, France.
PLoS One. 2013 Jun 20;8(6):e66167. doi: 10.1371/journal.pone.0066167. Print 2013.
In natural populations, individuals are infected more often by several pathogens than by just one. In such a context, pathogens can interact. This interaction could modify the probability of infection by subsequent pathogens. Identifying when pathogen associations correspond to biological interactions is a challenge in cross-sectional studies where the sequence of infection cannot be demonstrated.
METHODOLOGY/PRINCIPAL FINDINGS: Here we modelled the probability of an individual being infected by one and then another pathogen, using a probabilistic model and maximum likelihood statistics. Our model was developed to apply to cross-sectional data, vector-borne and persistent pathogens, and to take into account confounding factors. Our modelling approach was more powerful than the commonly used Chi-square test of independence. Our model was applied to detect potential interaction between Borrelia afzelii and Bartonella spp. that infected a bank vole population at 11% and 57% respectively. No interaction was identified.
CONCLUSIONS/SIGNIFICANCE: The modelling approach we proposed is powerful and can identify the direction of potential interaction. Such an approach can be adapted to other types of pathogens, such as non-persistents. The model can be used to identify when co-occurrence patterns correspond to pathogen interactions, which will contribute to understanding how organism communities are assembled and structured. In the long term, the model's capacity to better identify pathogen interactions will improve understanding of infectious risk.
在自然种群中，个体感染一种以上病原体的情况比只感染一种病原体的情况更为常见。在这种情况下，病原体之间可能会相互作用。这种相互作用可能会改变随后感染其他病原体的概率。在横断面研究中，无法证明感染的先后顺序，因此确定病原体之间的关联是否对应于生物学相互作用是一个挑战。
方法/主要发现：在这里，我们使用概率模型和最大似然统计来模拟个体被一种病原体感染然后再被另一种病原体感染的概率。我们的模型旨在应用于横断面数据、媒介传播和持续性病原体，并考虑混杂因素。我们的建模方法比常用的独立性卡方检验更有效。我们的模型被应用于检测感染田鼠种群分别为 11%和 57%的伯氏疏螺旋体和巴尔通体之间潜在的相互作用。没有发现相互作用。
结论/意义：我们提出的建模方法功能强大，可以识别潜在相互作用的方向。这种方法可以适用于其他类型的病原体，如非持续性病原体。该模型可用于确定共现模式是否对应于病原体相互作用，这将有助于理解生物群落是如何组装和构成的。从长远来看，该模型更好地识别病原体相互作用的能力将提高对感染风险的理解。
