Know more

The Animal Health Department launched the first international animal epidemiology modelling challenge on 28/08/2020. This challenge aims to improve the collective capacity of teams of modellers to predict the spread of a large-scale epizootic and to support public decisions in crisis situations. A virtual epidemic was generated using the example of African swine fever in a realistic European context, at the interface between wild boar and animal husbandry. Since the launch, nine international teams have been using the simulated synthetic data in real time and mobilising statistical and mathematical models to predict the outcome of the epidemic and attempt to identify the most effective control strategies to limit the impact of the disease on the territory. The final predictions are expected in the spring of 2021. At the end of this period, a collective reflection will aim to identify the most promising modelling methods or combinations of methods to act in a reactive and reliable manner in the event of a real health crisis.

Rift Valley Fever (RVF) is transmitted by mosquitoes, mainly to cattle, causing waves of abortions and high mortality in younger animals. It is a zoonosis, the severe form of which can be fatal to humans. RVF is on the WHO's list of priority emerging diseases. A mathematical model has been developed to estimate the epidemic potential of RVF in northern Senegal, a region that has been regularly affected since the late 1980s. It is in September that the introduction of the virus can cause the most secondary cases and allow an epidemic to start. The locations most at risk of becoming an outbreak vary from year to year. In these at-risk locations, increasing cattle immunity would be more effective in reducing virus transmission than increasing small ruminant immunity. On the other hand, mosquito densities are such that reducing their population is not a viable way of reducing risk. This work will be completed by integrating the spatio-temporal transmission of the virus via seasonal animal mobility. Testing climatic scenarios linked to global changes would also make it possible to anticipate the risk at Europe's doorstep.

With several hundred thousand human cases per year, tick-borne diseases are the most important vector-borne diseases in Europe. They involve a wide variety of wild and domestic vertebrate species that are feeding hosts used by ticks for their blood meals but also reservoirs of infectious agents transmitted by these mites. The mosaic of habitats that make up bocage landscapes (forests, hedges, crops, meadows...) influences the spatial distribution of these pathogens. The rate of infection by tick-borne bacteria varies in small rodents sampled at sites with different woodland areas and hedgerow densities. The frequency of anaplasmosis bacteria increased with the proportion of wooded habitat, which in turn correlated with the abundance of woodland mice that are effective reservoirs of this pathogen. For Lyme disease bacteria, the greater the interface between woodland and grassland, the more frequent they are in small rodents. These results illustrate the need to consider community and landscape ecology to better understand the epidemiology of these diseases.

Animal autopsy is a highly technical veterinary procedure, essential for regulatory and therapeutic purposes. IVAN is an innovative decision support tool, in the form of a WEB application, for veterinarians performing autopsies on cattle. IVAN uses artificial intelligence (AI) methods to assist the veterinarian. From general data on the animal, IVAN suggests a list of organs to focus on, looking for lesions that it can help identify. The suggestions at each stage of its reasoning are ranked according to their probability and submitted to the practitioner for validation. This makes the AI-guided process explicit for the user (no "black box" effect) and allows the veterinarian to control the entire diagnostic chain. The system then lists the morphological diagnoses likely to correspond to the lesions and finally proposes the possible diseases, always in order of probability. IVAN can also suggest additional tests to confirm the diagnosis. IVAN is a unique and powerful assistant that provides significant help to the veterinarian in the autopsy process. All the steps of these deductions are based on AI methods (Bayesian networks), trained by the data collected for several years by Autopsy Service at Oniris.

Endemic diseases are constantly circulating and can cause heavy losses in animal husbandry in the medium to long term. Controlling these diseases is essential for sustainable agriculture and competitive agri-food chains. Interdisciplinary scientific collaborations between biologists, economists and modellers have highlighted how mathematical modelling in epidemiology contributes to a better understanding and prediction of the circulation of these diseases, as well as to guiding their control at all scales, from the animal to the territory and the primary production chain. The scientific and methodological challenges that still exist in proposing targeted control options and assessing their impact were identified. The strategic decision-making of farmers needs to be included in order to better understand the trade-off between individual and collective management and to better orient incentives. Integrating the immune response of hosts to the infection would also make it possible to refine interventions, particularly therapeutic and preventive (vaccination). Finally, feeding the models with observable data from animal husbandry would increase their realism and practical usefulness, to support public or private collective policies.

Mechanistic epidemiological models, such as those developed in the BIOEPAR Unit, allow for a detailed understanding and prediction of the spread of pathogens, as well as the evaluation and comparison of control scenarios. Making these models usable independently by animal health managers can support public policies or improve collective health management in animal husbandry. This implies transforming them into decision support tools, which usually requires significant ad-hoc software development. The ATOM prematurity project, financed by INRAE's Partnership and Transfer for Innovation Department, consists of developing a software chain that automatically transforms academic epidemiological models into decision-making tools that can be used independently and adapted to specific needs. This innovative initiative, combining artificial intelligence and software engineering methods, aims to facilitate the transfer of results from academic research in animal health to the field. The use of the tools produced will also make it possible to base decision-making on the scale of the farm, the sector or the territory.

Bovine paratuberculosis is a chronic bacterial infection of the gut that causes significant losses on dairy farms. The Irish dairy system uses grazed grass as the main source of feed for lactating cattle, which means that calf births take place in the spring in the majority of herds. To understand how to control Johne's disease in Irish dairy herds, this seasonal herd demography must be taken into account. The French Johne's disease epidemiological model was adapted to simulate transmission dynamics in a seasonal context and to evaluate different control options. Exposure of calves to environments contaminated by cows remains the main risk to be controlled in this seasonal context, particularly for young calves, which are the most susceptible to infection. Testing and culling of excretory animals is an effective means of control, provided that it is used before the calving period, so as to reduce the number of highly excretory cows present at the time of calf birth.

A partnership with the Institut Sénégalais de Recherches Agricoles (ISRA), Laboratoire National d'Elevage et de Recherches Vétérinaires (National Laboratory for Livestock and Veterinary Research), initially facilitated by CIRAD, is gradually becoming a long-term project.

I am Floor and I was born and raised in the Netherlands. Last year I arrived in France to work here for three years, however, my employer is Irish. An awesome collaboration between three European countries, but also a challenge. Apart from practical difficulties, this is also a great opportunity to learn. Because, even though France, Ireland, and the Netherlands are all Western European countries, the cultures are really different. Every day feels like an adventure. The French with their quirky habits, the Irish with their overly politeness, and me, the Dutchman observing it all. There is something new to learn every day!

With the COVID-19 crisis, the question of viral co-infections has become a major issue. What do respiratory viruses do when they meet and what are the consequences for the animal host? The analysis of the consequences of the encounter of swine influenza viruses (Influenza A virus - IAV) with the porcine respiratory dysgenesis syndrome virus (vSDRP), a virus from a viral family (Arteriviridae), which is fairly close to that of the coronaviruses, is a relevant model for understanding these coinfections. The pig is in fact not only a species of major agronomic interest but also a model for humans. The viral interference is strong between these viruses whatever the cell type considered. Remarkably, the epithelial cell, the preferred target of the influenza virus, will react very differently to the influenza virus if it has been pre-exposed to vSDRP. Indeed, it becomes almost resistant to infection by the IAV at various levels. This observation, valid for both wild and vaccine viruses, is a key step in understanding the course of polyinfections and may have important implications for vaccine strategies.