Escobar Luis E, Qiao Huijie, Lee Christine, Phelps Nicholas B D
Minnesota Aquatic Invasive Species Research Center, University of Minnesota, St. Paul, MN, United States.
Department of Fisheries, Wildlife, and Conservation Biology, University of Minnesota, St. Paul, MN, United States.
Front Vet Sci. 2017 Jul 17;4:105. doi: 10.3389/fvets.2017.00105. eCollection 2017.
Disease biogeography is currently a promising field to complement epidemiology, and ecological niche modeling theory and methods are a key component. Therefore, applying the concepts and tools from ecological niche modeling to disease biogeography and epidemiology will provide biologically sound and analytically robust descriptive and predictive analyses of disease distributions. As a case study, we explored the ecologically important fish disease Heterosporosis, a relatively poorly understood disease caused by the intracellular microsporidian parasite . We explored two novel ecological niche modeling methods, the minimum-volume ellipsoid (MVE) and the Marble algorithm, which were used to reconstruct the fundamental and the realized ecological niche of , respectively. Additionally, we assessed how the management of occurrence reports can impact the output of the models. Ecological niche models were able to reconstruct a proxy of the fundamental and realized niche for this aquatic parasite, identifying specific areas suitable for Heterosporosis. We found that the conceptual and methodological advances in ecological niche modeling provide accessible tools to update the current practices of spatial epidemiology. However, careful data curation and a detailed understanding of the algorithm employed are critical for a clear definition of the assumptions implicit in the modeling process and to ensure biologically sound forecasts. In this paper, we show how sensitive MVE is to the input data, while Marble algorithm may provide detailed forecasts with a minimum of parameters. We showed that exploring algorithms of different natures such as environmental clusters, climatic envelopes, and logistic regressions (e.g., Marble, MVE, and Maxent) provide different scenarios of potential distribution. Thus, no single algorithm should be used for disease mapping. Instead, different algorithms should be employed for a more informed and complete understanding of the pathogen or parasite in question.
疾病生物地理学目前是一个有望补充流行病学的领域,生态位建模理论和方法是其关键组成部分。因此,将生态位建模的概念和工具应用于疾病生物地理学和流行病学,将为疾病分布提供生物学上合理且分析上可靠的描述性和预测性分析。作为一个案例研究,我们探讨了具有重要生态学意义的鱼类疾病——异孢子虫病,这是一种由细胞内微孢子虫寄生虫引起的相对了解较少的疾病。我们探索了两种新的生态位建模方法,即最小体积椭球体(MVE)和大理石算法,它们分别用于重建该寄生虫的基础生态位和实际生态位。此外,我们评估了发病报告的管理方式如何影响模型的输出。生态位模型能够重建这种水生寄生虫的基础生态位和实际生态位的代理,识别出适合异孢子虫病的特定区域。我们发现,生态位建模在概念和方法上的进展提供了可用于更新当前空间流行病学实践的工具。然而,仔细的数据整理以及对所采用算法的详细理解对于明确建模过程中隐含假设的定义以及确保生物学上合理的预测至关重要。在本文中,我们展示了MVE对输入数据的敏感性,而大理石算法可能以最少的参数提供详细的预测。我们表明,探索不同性质的算法,如环境聚类、气候包络和逻辑回归(例如大理石算法、MVE和最大熵模型)会提供不同的潜在分布情景。因此,不应仅使用单一算法进行疾病绘图。相反,应采用不同的算法,以便更全面、深入地了解所研究的病原体或寄生虫。