Abstract
The diversity, composition and assemblages of the soil and plant microbiome are partly determined by the environment and biotic interactions. The AlUla oasis, located in the Saudi Arabian desert, is characterized by strong abiotic constraints, linked to a hyper-alkaline pH and low availability of water and nutrients. The activity and growth of organisms is therefore subject to this punctual and spatially heterogeneous availability of resources. This leads to a spatially discontinuous distribution of plants, known as islands of fertility, which influence soil composition and microbial communities. Plants also form symbiotic relationships with microorganisms, which influence their health, resistance to drought and the acquisition of mineral and water resources, and are particularly critical in this ecosystem. Finally, these ecosystems, already weakened by climate change, are also subject to significant agricultural pressures, leading to soil degradation and associated biodiversity loss. Little is known at present about the biodiversity of hot, arid ecosystems, particularly in Saudi Arabia, a country whose borders have long remained closed.The goal of this thesis is to characterize the bacterial and fungal microbiome of soil and plants in relation to its environment, and its response to different land uses through amplicon sequencing. In order to extend the analytical framework of the study of community diversity and composition, I have sought to use co-occurrence network metrics and explore new methodologies for their study. In a first chapter, the mutual influence of plants, soil and microorganisms in a micro-environment, the fertility islands, is characterized. Then, a field sampling campaign over two seasons enabled me to carry out an extensive analysis of the plant microbiome using the co-occurrence network approach. Focusing on the intra- and inter-kingdom relationships of symbiotic taxa, this work demonstrated the redundant assortativity of mycorrhizal fungi, and the integration of nitrogen-fixing bacteria into the extended plant microbiome. The microbiome's response to a simulated precipitation event in the field was also used to characterize the microbiome's taxonomic response to water availability in the soil. In the third chapter, the microbiome's response to historical contingencies of an anthropogenic or natural nature, describes how cycles of desiccation and flooding affect contemporary microbial communities. By studying the stability of their interactions, it shows how past agricultural activities has had a lasting impact on the structure of the microbiome. Finally, the quantification of community assembly processes has made it possible to determine the effect of past disturbances on bacterial and fungal selection processes.Collectively, the results of this thesis improve our understanding of the assembly and structure of soil and plant microbiota in a little-known desert ecosystem. In addition, co-occurrence analyses have proven to be a valuable tool in the formulation of new fundamental hypotheses on the founding role of symbioses, and the response of the microbiota to disturbance. Continued study of the complex structure of networks, complemented by the exploration of microbial functions and reductionist approaches to be able to couple covariance relationships to ecological processes, promises major advances in microbial ecology in the future.