The model shows that a combination of variables, such as domestic duck and chicken densities, can be used to predict in which geographical zone the virus is likely to spread upon introduction. Similar models, which take into account the specificities of the currently fast-spreading H5N8 virus, are being built to support prevention and control efforts.
In the last few weeks, several countries - mainly in Europe - have reported the introduction of avian influenza H5N8 viruses. All these countries have reported infections and mortality in wild birds, while only a few have reported outbreaks in domestic poultry. Understanding in which zones the introductions could lead to long-term persistence and to chains of farm-to-farm transmission would allow to (1) focus the survey efforts on the most probable locations, (2) ensure early detection and (3) avoid massive culling.
Using the vast amount of epidemics data on the HPAI H5N1 virus, which has spread extensively across several continents in the past ten years, an international team of scientists, including SIB’s Swiss-Prot and Vital-IT groups, the Université Libre de Bruxelles and the Food and Agriculture Organization of the United Nations (FAO), showed that the spatial distribution of the virus’ hosts, such as domestic duck and chicken, could be reliably used to map areas that are particularly suitable for its spread in domestic poultry upon introduction. This achievement results from a multidisciplinary approach – i.e. applying ecological modeling techniques to virology – powered by machine learning.
With a mortality rate of about 60% in humans infected by birds with H5N1, according to the World Health Organization the virus remains a threat to human health in countries where it persists. It also continues to impact poultry production in countries where it has become endemic.
This modeling approach was also tentatively used to predict the spread of another group of viruses, including those similar to the currently fast-spreading H5N8. While the model had a lower predictive power – probably due to the recent spread of the virus and the much more limited set of presence data – it showed a strong association with intensively raised chicken densities and anthropogenic factors.
Altogether, this study sets out the first methodological foundations towards a predictive approach of disease spread, and similar models accounting for the specificities of the current H5N8 wave of introduction are already being built.
This study also illustrates how SIB supports research on global health issues (SIB became FAO Reference Centre for Bioinformatics in February 2015) by providing its expertise in virus curation (using Viralzone) and bioinformatics.
Reference: Global mapping of highly pathogenic avian influenza H5N1 and H5Nx clade 18.104.22.168 viruses with spatial cross-validation DOI: http://dx.doi.org/10.7554/eLife.19571
Image: Free-range duck farming, Vietnam. Photo credit © Jody McIntyre CC-BY-SA