Effective control of livestock infectious diseases is a major issue for sustainable animal farming and competitive agri-food chains, as well as for public health. Livestock trade movements is one of the most important pathways for pathogen transmission between holdings. Therefore, it is crucial to understand and be able to predict the temporal evolution of the network of animal movements, and assess the risk related to epidemics unfolding on this network. Here, we focus on the study of cattle trade network in France, based on a fully detailed dataset spanning over more than ten years (2005-2015), extracted from the French database of cattle movements, which records life histories of all French cattle from birth to death. These data, formalized by a time-varying network (with holdings as nodes and commercial transactions as links), were analyzed using tools from graph theory and cluster analysis. Whereas this kind of analysis is often straightforward for static networks, it becomes highly challenging when dealing with time-varying networks, such as animal trade networks, which are, in addition, of high dimension (more than 200,000 active nodes and more than 8,000,000 animal movements per year in our data). Results on two aspects, the analysis of the temporal dynamics of cattle movements and the assessment of the associated epidemiological risk, will be presented. On the first axis, we focus on the characterization of the temporal stability of the main descriptors of cattle trade networks and the fidelity over time of transaction partners. On the second axis, proxies for pathogen spread, such as the reachability ratio (accounting for the average number of nodes to be reached during an outbreak, using time-respecting paths in the network), are provided. This second analysis was performed for various assumptions on the transmission probability associated to each link and epidemiological parameters (such as the duration of the infection in each node), and computed using efficient algorithms of network explorationRead less <