The African swine fever modelling challenge: Model comparison and lessons learnt.

Robust epidemiological knowledge and predictive modelling tools are needed to address challenging objectives, such as: understanding epidemic drivers; forecasting epidemics; and prioritising control measures. Often, multiple modelling approaches can be used during an epidemic to support effective decision making in a timely manner. Modelling challenges contribute to understanding the pros and cons of different approaches and to fostering technical dialogue between modellers. In this paper, we present the results of the first modelling challenge in animal health - the ASF Challenge - which focused on a synthetic epidemic of African swine fever (ASF) on an island. The modelling approaches proposed by five independent international teams were compared. We assessed their ability to predict temporal and spatial epidemic expansion at the interface between domestic pigs and wild boar, and to prioritise a limited number of alternative interventions. We also compared their qualitative and quantitative spatio-temporal predictions over the first two one-month projection phases of the challenge. Top-performing models in predicting the ASF epidemic differed according to the challenge phase, host species, and in predicting spatial or temporal dynamics. Ensemble models built using all team-predictions outperformed any individual model in at least one phase. The ASF Challenge demonstrated that accounting for the interface between livestock and wildlife is key to increasing our effectiveness in controlling emerging animal diseases, and contributed to improving the readiness of the scientific community to face future ASF epidemics. Finally, we discuss the lessons learnt from model comparison to guide decision making.

A combination of probabilistic and mechanistic approaches for predicting the spread of African swine fever on Merry Island.

Over the last decade African swine fever virus, one of the most virulent pathogens known to affect pigs, has devastated pork industries and wild pig populations throughout the world. Despite a growing literature on specific aspects of African swine fever transmission dynamics, it remains unclear which methods and approaches are most effective for controlling the disease during a crisis. As a consequence, an international modelling challenge was organized in which teams analyzed and responded to a stream of data from an in silico outbreak in the fictive country of Merry Island. In response to this outbreak, we developed a modelling approach that aimed to predict the evolution of the epidemic and evaluate the impact of potential control measures. Two independent models were developed: a stochastic mechanistic space-time compartmental model for characterizing the dissemination of the virus among wild boar; and a deterministic probabilistic risk model for quantifying infection probabilities in domestic pig herds. The combined results of these two models provided valuable information for anticipating the main risks of dissemination and maintenance of the virus (speed and direction of African swine fever spread among wild boar populations, pig herds at greatest risk of infection, the size of the epidemic in the short and long terms), for evaluating the impact of different control measures and for providing specific recommendations concerning control interventions.

Correction: Estimation of the dispersal distances of an aphid-borne virus in a patchy landscape.

[This corrects the article DOI: 10.1371/journal.pcbi.1006085.].

Biopesticides improve efficiency of the sterile insect technique for controlling mosquito-driven dengue epidemics.

Various mosquito control methods use factory raised males to suppress vector densities. But the efficiency of these methods is currently insufficient to prevent epidemics of arbovirus diseases such as dengue, chikungunya or Zika. Suggestions that the sterile insect technique (SIT) could be "boosted" by applying biopesticides to sterile males remain unquantified. Here, we assess mathematically the gains to SIT for Aedes control of either: boosting with the pupicide pyriproxifen (BSIT); or, contaminating mosquitoes at auto-dissemination stations. Thresholds in sterile male release rate and competitiveness are identified, above which mosquitoes are eliminated asymptotically. Boosting reduces these thresholds and aids population destabilisation, even at sub-threshold release rates. No equivalent bifurcation exists in the auto-dissemination sub-model. Analysis suggests that BSIT can reduce by over 95% the total release required to circumvent dengue epidemics compared to SIT. We conclude, BSIT provides a powerful new tool for the integrated management of mosquito borne diseases.

A framework for estimating the effects of sequential reproductive barriers: Implementation using Bayesian models with field data from cryptic species.

Determining how reproductive barriers modulate gene flow between populations represents a major step toward understanding the factors shaping the course of speciation. Although many indices quantifying reproductive isolation (RI) have been proposed, they do not permit the quantification of cross-direction-specific RI under varying species frequencies and over arbitrary sequences of barriers. Furthermore, techniques quantifying associated uncertainties are lacking, and statistical methods unrelated to biological process are still preferred for obtaining confidence intervals and P-values. To address these shortcomings, we provide new RI indices that model changes in gene flow for both directions of hybridization, and we implement them in a Bayesian model. We use this model to quantify RI between two species of the psyllid Cacopsylla pruni based on field genotypic data for mating individuals, inseminated spermatophores and progeny. The results showed that preinsemination isolation was strong, mildly asymmetric, and indistinguishably different between study sites despite large differences in species frequencies; that postinsemination isolation strongly affected the more common hybrid type; and that cumulative isolation was close to complete. In the light of these results, we discuss how these developments can strengthen comparative RI studies.

Estimation of the dispersal distances of an aphid-borne virus in a patchy landscape.

Characterising the spatio-temporal dynamics of pathogens in natura is key to ensuring their efficient prevention and control. However, it is notoriously difficult to estimate dispersal parameters at scales that are relevant to real epidemics. Epidemiological surveys can provide informative data, but parameter estimation can be hampered when the timing of the epidemiological events is uncertain, and in the presence of interactions between disease spread, surveillance, and control. Further complications arise from imperfect detection of disease and from the huge number of data on individual hosts arising from landscape-level surveys. Here, we present a Bayesian framework that overcomes these barriers by integrating over associated uncertainties in a model explicitly combining the processes of disease dispersal, surveillance and control. Using a novel computationally efficient approach to account for patch geometry, we demonstrate that disease dispersal distances can be estimated accurately in a patchy (i.e. fragmented) landscape when disease control is ongoing. Applying this model to data for an aphid-borne virus (Plum pox virus) surveyed for 15 years in 605 orchards, we obtain the first estimate of the distribution of flight distances of infectious aphids at the landscape scale. About 50% of aphid flights terminate beyond 90 m, which implies that most infectious aphids leaving a tree land outside the bounds of a 1-ha orchard. Moreover, long-distance flights are not rare-10% of flights exceed 1 km. By their impact on our quantitative understanding of winged aphid dispersal, these results can inform the design of management strategies for plant viruses, which are mainly aphid-borne.

Using sensitivity analysis to identify key factors for the propagation of a plant epidemic.

Identifying the key factors underlying the spread of a disease is an essential but challenging prerequisite to design management strategies. To tackle this issue, we propose an approach based on sensitivity analyses of a spatiotemporal stochastic model simulating the spread of a plant epidemic. This work is motivated by the spread of sharka, caused by plum pox virus, in a real landscape. We first carried out a broad-range sensitivity analysis, ignoring any prior information on six epidemiological parameters, to assess their intrinsic influence on model behaviour. A second analysis benefited from the available knowledge on sharka epidemiology and was thus restricted to more realistic values. The broad-range analysis revealed that the mean duration of the latent period is the most influential parameter of the model, whereas the sharka-specific analysis uncovered the strong impact of the connectivity of the first infected orchard. In addition to demonstrating the interest of sensitivity analyses for a stochastic model, this study highlights the impact of variation ranges of target parameters on the outcome of a sensitivity analysis. With regard to sharka management, our results suggest that sharka surveillance may benefit from paying closer attention to highly connected patches whose infection could trigger serious epidemics.

A random forest approach for predicting the presence of Echinococcus multilocularis intermediate host Ochotona spp. presence in relation to landscape characteristics in western China.

Understanding distribution patterns of hosts implicated in the transmission of zoonotic disease remains a key goal of parasitology. Here, random forests are employed to model spatial patterns of the presence of the plateau pika (Ochotona spp.) small mammal intermediate host for the parasitic tapeworm Echinococcus multilocularis which is responsible for a significant burden of human zoonoses in western China. Landsat ETM+ satellite imagery and digital elevation model data were utilized to generate quantified measures of environmental characteristics across a study area in Sichuan Province, China. Land cover maps were generated identifying the distribution of specific land cover types, with landscape metrics employed to describe the spatial organisation of land cover patches. Random forests were used to model spatial patterns of Ochotona spp. presence, enabling the relative importance of the environmental characteristics in relation to Ochotona spp. presence to be ranked. An index of habitat aggregation was identified as the most important variable in influencing Ochotona spp. presence, with area of degraded grassland the most important land cover class variable. 71% of the variance in Ochotona spp. presence was explained, with a 90.98% accuracy rate as determined by 'out-of-bag' error assessment. Identification of the environmental characteristics influencing Ochotona spp. presence enables us to better understand distribution patterns of hosts implicated in the transmission of Em. The predictive mapping of this Em host enables the identification of human populations at increased risk of infection, enabling preventative strategies to be adopted.

Mixtures of GAMs for habitat suitability analysis with overdispersed presence / absence data.

A new approach to species distribution modelling based on unsupervised classification via a finite mixture of GAMs incorporating habitat suitability curves is proposed. A tailored EM algorithm is outlined for computing maximum likelihood estimates. Several submodels incorporating various parameter constraints are explored. Simulation studies confirm, that under certain constraints, the habitat suitability curves are recovered with good precision. The method is also applied to a set of real data concerning presence/absence of observable small mammal indices collected on the Tibetan plateau. The resulting classification was found to correspond to species-level differences in habitat preference described in previous ecological work.

Landscape composition and spatial prediction of alveolar echinococcosis in southern Ningxia, China.

BACKGROUND: Alveolar echinococcosis (AE) presents a serious public health challenge within China. Mass screening ultrasound surveys can detect pre-symptomatic AE, but targeting areas identified from hospital records is inefficient regarding AE. Prediction of undetected or emerging hotspots would increase detection rates. Voles and lemmings of the subfamily Arvicolinae are important intermediate hosts in sylvatic transmission systems. Their populations reach high densities in productive grasslands where food and cover are abundant. Habitat availability is thought to affect arvicoline population dynamic patterns and definitive host-intermediate host interactions. Arvicoline habitat correlates with AE prevalence in Western Europe and southern Gansu Province, China. METHODS AND FINDINGS: Xiji County, Ningxia Hui Autonomous Region, borders southern Gansu. The aims of this study were to map AE prevalence across Xiji and test arvicoline habitat as a predictor. Land cover was mapped using remotely sensed (Landsat) imagery. Infection status of 3,205 individuals screened in 2002-2003 was related, using generalised additive mixed models, to covariates: gender; farming; ethnicity; dog ownership; water source; and areal cover of mountain pasture and lowland pasture. A Markov random field modelled additional spatial variation and uncertainty. Mountain pasture and lowland pasture were associated with below and above average AE prevalence, respectively. CONCLUSIONS: Low values of the normalised difference vegetation index indicated sub-optimality of lowland pasture for grassland arvicolines. Unlike other known endemic areas, grassland arvicolines probably did not provide the principal reservoir for Echinococcus multilocularis in Xiji. This result is consistent with recent small mammal surveys reporting low arvicoline densities and high densities of hamsters, pikas and jerboas, all suitable intermediate hosts for E. multilocularis, in reforested lowland pasture. The risk of re-emergence is discussed. We recommend extending monitoring to: southern Haiyuan County, where predicted prevalence was high; southern Xiji County, where prediction uncertainty was high; and monitoring small mammal community dynamics and the infection status of dogs.