TY - JOUR
T1 - Novel methods in disease biogeography
T2 - A case study with heterosporosis
AU - Escobar, Luis E.
AU - Qiao, Huijie
AU - Lee, Christine
AU - Phelps, Nicholas B.D.
N1 - Publisher Copyright:
© 2017 Escobar, Qiao, Lee and Phelps.
PY - 2017/7/17
Y1 - 2017/7/17
N2 - 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 Heterosporis sutherlandae. 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 H. sutherlandae, 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.
AB - 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 Heterosporis sutherlandae. 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 H. sutherlandae, 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.
KW - Disease biogeography
KW - Ecological niche modeling
KW - Heterosporosis
KW - Minimum-volume ellipsoid
KW - Risk map
UR - http://www.scopus.com/inward/record.url?scp=85038224580&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85038224580&partnerID=8YFLogxK
U2 - 10.3389/fvets.2017.00105
DO - 10.3389/fvets.2017.00105
M3 - Article
C2 - 28770215
AN - SCOPUS:85038224580
SN - 2297-1769
VL - 4
JO - Frontiers in Veterinary Science
JF - Frontiers in Veterinary Science
IS - JUL
M1 - 105
ER -