A two-year post-doctoral position in computer sciences is now opened at INRA-Dijon (France). The post-doctoral scientist will develop logic-based machine-learning algorithms for the reconstruction of microbial ecological networks from metabarcoding datasets. We envision that these algorithms will be very useful to the biocontrol of plant pathogens in the near future, by allowing us to automatically identify candidate biocontrol agents. Plants are associated with a huge diversity of microorganisms that interact. Deciphering these microbial interactions is crucial since they underpin the regulation of plant diseases. These microbial interactions are usually identified using co-culture experiments. This is a tedious and time-consuming process that cannot easily be extended to the whole microbial community with which a pathogen interacts. A current challenge is to reconstruct microbial interaction networks directly from environmental DNA. The purpose of the project will be (1) to develop generic methods for learning microbial interaction networks from metabarcoding data and (2) to demonstrate that these new methods inform biological control of pathogens because they reveal antagonistic interactions between pathogen species and other microorganisms. This proof-of-concept study will be performed using both simulated datasets and real metabarcoding datasets on some major foliar pathogens of grapevine and wheat. The post-doctoral position is part of the BCMicrobiome project funded by the public-private Biocontrol Consortium*. Regular reporting to the Consortium will be requested and there will be an exclusivity period on the results obtained on the real metabarcoding datasets. The post-doctoral researcher will be based in Dave Bohan's lab at INRA-Dijon. He will work in close collaboration with the project's PI (Corinne Vacher, INRA-Bordeaux) and with other ecologists and plant pathologists involved in the project. A three-month stay in the Department of Computing at the University of Surrey will allow the post-doctoral researcher to gain new skills in machine-learning by working in collaboration with Alireza Tamaddoni-Nezhad. Close collaboration with the group of Stéphane Robin (AgroParisTech, Paris), who develops statistical methods of network inference, is also expected. The candidate should ideally have a background in machine-learning and/or statistics, good communication skills, and some interest in ecology and agriculture. Candidates having some experience in the analysis of ecological networks are also welcome to apply.