Deciphering the mechanisms of molecular reprogramming in plants upon interaction with biotic agents using multi-omics integration and artificial intelligence.

En présentiel ou via Zoom :

https://inrae-fr.zoom.us/j/96884317966

Jury members : 

Président/te of the jury :

• Prof. Elisabeth Pecou Université Cote D’Azur, France : Présidente and Examinater

Rapporteurs :

• Prof. Kristina Gruden University of Ljubljana, Slovenia
• Dr. Pierre Larmande University of Montpellier, UMR DIADE, IRD, France
• Prof. Nunzio D’Agostino University of Naples “Federico II”, Italy

Examinaters :

• Dr. Gabriel Krouk Institut for plant science of Montpellier, France
• Dr. Sylvain Raffaele University of Toulouse, LIPME France

GUEST :

• Dr. Daphne Ezer University of York, England

Abstract :

Plants live in a constantly changing environment that happens to be unfavourable or even hostile. Therefore, they developed high phenotypic plasticity that includes rapid responses to aggressive environmental factors and adaptations to changing environments. In the case of bio-aggressors, interactions are based on a molecular dialogue between the pathogen and its host. Indeed, plants and pests are engaged in an arms race where plants have evolved stratagems to detect the bio-aggressor while the latter have developed mechanisms to bypass the host's immune system. Pathogens modulate the immune response, cell signalling, metabolism or even the growth of the plant by acting directly on its transcriptomic regulation. This modulation of gene expression relies on the secretion of specific pathogenic proteins by the parasite; these are called effector proteins. During the infection, pathogen proteins play a crucial role in the re-wiring of multiple biochemical processes occurring in the host, which ultimately allow for the progression of the infection. Recent work has shown that evolutionarily distant pest effectors interact with the same targets in the host plant. While specific effector targets have long been sought, looking at non-specific conserved mechanisms may hold promise for a broader view of the interaction. As a counterpart, hosts make use of their protein machinery to trigger defence mechanisms against the pathogen. Therefore, it is of primary importance to characterize the plant-pathogen global network of interactions to gain better insight into the biological processes that regulate plants’ infection. To study these interactions, most of the available literature either focuses on plant mechanisms of response and does not consider the pest, or contrariwise based on the pathogens but neglects plant physiology. When considering both, those studies are limited to inspecting one or very few genes or proteins. Therefore, a global overview of the impacted biological processes is still missing. Despite being very informative, taken individually these interactions give only a partial and incomplete picture of the system. In my research project, I propose to bridge this gap by reconstructing integrated and comprehensive models of plant-pathogen interactions in a system biology perspective. We first focused on A. thaliana and tomato (S. lycopersicum) two well-studied plants at the Institut Sophia Agrobiotech; and relatives of peanut of the Arachis genus, leveraging the collaboration with EMBRAPA in Brazil. Several omics data are available of these plants upon infection with multiple pathogens. We have developed a novel model to first integrate multi-transcriptomics data called HIVE and then reconstruct the global network of interactions with a newly developed knowledge graph model called TomTom. The application of those models to three phytopathosystems allowed to characterize specific and/or shared molecular mechanisms which are engaged by plants and pathogens during different attacks. Leveraging POMOdORO, an extensive collection of more than 2000 omics data from the literature that we are developing, we will expand those original frameworks providing four novel models based on variational autoencoders and physically-informed neural network models, in collaboration with Dr. Regis Duvigneau at Inria Sophia-Antipolis and Dr. Lorenzo Sala at INRAE Jouy-en-Josas. Those models will be used to analyse multi-omics collections of multiple plants upon interaction with biotic agents to extract the most informative molecular signatures. To investigate the molecular relationship, we will use two network-based approaches: a hybrid knowledge-data driven in collaboration with Dr. Dugourd at the University of Heidelberg and a novel purely data driven strategy. The analysis of the inferred networks will be performed by adapting and implementing novel theoretical frameworks in the context of topological data analysis in collaboration with Dr. Carriere at Inria Sophia-Antipolis. We will identify potential hubs and pathways imbalanced during the infection that are pathogen-specific and/or shared mechanisms of response. Altogether this project we will provide the needed molecular toolboxes to characterize and predict the plants susceptibility or resistance to diseases. This project will provide a full understanding of the complexity of plant immunity mechanisms and establish the modelling framework that can be applied to other pathosystems. Finally, this project will provide the basis for the development of new, multi-pathogen, more sustainable and environmentally friendly control strategies.

Keywords: Deciphering the mechanisms of molecular reprogramming in plants upon interaction with biotic agents using multi-omics integration and artificial intelligence.