Computational protocol: Integrated Analysis of Environment, Cattle and Human Serological Data: Risks and Mechanisms of Transmission of Rift Valley Fever in Madagascar

Similar protocols

Protocol publication

[…] A landscape map of Madagascar was obtained from Globcover project []. The GlobCover 2009 landscape product is a 300-m global landscape map produced from an automated classification of Medium Resolution Imaging Spectrometer (MERIS) time series. The global landscape map included 22 landscape classes defined with the United Nations (UN) Land Cover Classification System (LCCS). Among these 22 classes, we identified 5 relevant LCCS categories: “Cultivated Terrestrial Areas and Managed Lands” (so-called Crops), “Woody/ Trees”, “Shrubs”, “Herbaceous”, “Artificial Surfaces (so-called Urbanization)”. To reflect the availability of potential breeding habitats of RVF vectors in Madagascar such as artificial, irrigated, permanent and temporary water bodies, we needed to combine different data sources extracted from several GIS databases. The first one described inland permanent water point, such as lake, and was available from DIVA-GIS (http://www.diva-gis.org/). Marshland data representing temporary water bodies were obtained from Geographical Information Systems at the Royal Botanic Gardens, Kew []. Wetland locations representing temporary water bodies were extracted from the International Panel on Climate Change (IPCC; []). Irrigated area locations came from Global Map of Irrigation Areas (GMIA) from AQUASTAT-FAO []. [...] As a first step univariate analyses of association between suspected risk factors and cattle or human RVFV serological status were undertaken using Chi square tests for categorical factors and generalized linear models for quantitative factors. Risk factors with significance level ≤0.20 were then included as explanatory variables in GLMMs, with cattle or human individual serological status as the binomial response. In these models, it was assumed that the relationships between serological prevalence and quantitative factors were linear on the logit scale. To account for interdependency of serological status of individuals sampled in the same locality, the smallest administrative unit—the commune for the cattle model and the city/village for human model- were included in the models as a random effect. Multicollinearity among explanatory variables was assessed using Variance Inflation Factors (VIF) and correlation tests. Collinear factors were not included in a same model. The selection of the best models was based on the Akaike Information Criterion (AIC). When needed, a multi-model inference approach was used to estimate model-averaged fixed effects (mafe) and the relative importance (RI) of each explanatory variable []. Within the set of models tested, only those with an AIC within 2 units difference from the best model were considered [].Internal validity of sets of models was evaluated using the Receiver Operating Characteristic (ROC) curve method [].In addition, we calculated the 10-fold cross-validation prediction. Because, it is not possible to perform 10-fold cross-validation on GLMM, this procedure was applied to Generalized Linear Models that were similar to the selected GLMM except that did not include the site of sampling as random effect. Firstly, the cattle seroprevalence dataset was split randomly into 10 parts. Then, the model was fitted to 90% of the data and used to predict the serological status of the remaining 10% individuals as validation step. The procedure was performed 10 times, each time with 1 of the 10 parts as validation step. [].Finally, parameter estimations derived from the best cattle model were used to predict and map cattle seroprevalence at the commune scale for the whole island.Data analyses were performed using R software version 3.0.1 [–]. […]

Pipeline specifications

Software tools DIVA-GIS, MuMIn
Application Phylogenetics
Organisms Bos taurus, Homo sapiens
Diseases Infection, Q Fever, Rift Valley Fever