Fasciola hepatica is a parasitic trematode that infects wild and domesticated mammals, particularly cattle and sheep, and causes significant economic losses to global livestock production. In the present study, we used codominant genetic markers to define and build, for the first time, complex genotype networks for F. hepatica isolated from cattle and sheep in NW Spain. We generated three types of random networks with a number of nodes and edges as close as possible to the observed networks, and we then calculated 14 node centrality measures for both observed and random networks. Finally, using Linear Discriminant Analysis (LDA) and these measures as inputs, we constructed a quantitative structure-property relationship (QSPR)-like model able to predict the propensity of a specific genotype of F. hepatica to infect different infrapopulations, farms and/or host species. The accuracy, sensitivity and specificity of the model were >90% for both training and cross-validation series. We also assessed the applicability domain of the model. This type of QSPR model is a potentially powerful tool for epidemiological studies and could be used to manage and prevent the spread of fasciolosis.
Keywords: Codominant genetic markers; Complex networks; Fasciolosis; Graph theory; Linear discriminant analysis.
© 2013 Published by Elsevier Ltd.