Background Drought is global environmental stress that limits crop yields. Plant-associated microbiomes play a crucial role in determining plant fitness in response to drought, yet the fundamental mechanisms for maintaining microbial community stability under drought disturbances in wild rice are poorly understood. We make explicit comparisons of leaf, stem, root and rhizosphere microbiomes from the drought-tolerant wild rice (Oryza longistaminata) in response to drought stress. Results We find that the response of the wild rice microbiome to drought was divided into aboveground–underground patterns. Drought reduced the leaf and stem microbial community diversity and networks stability, but not that of the roots and rhizospheres. Contrary to the aboveground microbial networks, the drought-negative response taxa exhibited much closer interconnections than the drought-positive response taxa and were the dominant network hubs of belowground co-occurrence networks, which may contribute to the stability of the belowground network. Notably, drought induces enrichment of Actinobacteria in belowground compartments, but not the aboveground compartment. Additionally, the rhizosphere microbiome exhibited a higher proportion of generalists and broader habitat niche breadth than the microbiome at other compartments, and drought enhanced the proportion of specialists in all compartments. Null model analysis revealed that both the aboveground and belowground-community were governed primarily by the stochastic assembly process, moreover, drought decreased ‘dispersal limitation’, and enhanced ‘drift’. Conclusions Our results provide new insight into the different strategies and assembly mechanisms of the above and belowground microbial community in response to drought, including enrichment of taxonomic groups, and highlight the important role of the stochastic assembly process in shaping microbial community under drought stress.

The commercial aquatic animal microbiome may markedly affect the successful host's farming in various aquaculture systems. However, very little was known about it. Here, two different aquaculture systems, the rice–fish culture (RFC) and intensive pond culture (IPC) systems, were compared to deconstruct the skin, oral, and gut microbiome, as well as the gut metabolome of juvenile Chinese softshell turtle (Pelodiscus sinensis). Higher alpha‐diversity and functional redundancy of P. sinensis microbial community were found in the RFC than those of the IPC. The aquaculture systems have the strongest influence on the gut microbiome, followed by the skin microbiome, and finally the oral microbiome. Source‐tracking analysis showed that the RFC's microbial community originated from more unknown sources than that of the IPC across all body regions. Strikingly, the RFC's oral and skin microbiome exhibited a significantly higher proportion of generalists and broader habitat niche breadth than those of the IPC, but not the gut. Null model analysis revealed that the RFC's oral and skin microbial community assembly was governed by a significantly greater proportion of deterministic processes than that of the IPC, but not the gut. We further identified the key gene and microbial contribution to five significantly changed gut metabolites, 2‐oxoglutarate, N‐acetyl‐d‐mannosamine, cis‐4‐hydroxy‐d‐proline, nicotinamide, and l‐alanine, which were significantly correlated with important categories of microbe‐mediated processes, including the amino acid metabolism, GABAergic synapse, ABC transporters, biosynthesis of unsaturated fatty acids, as well as citrate cycle. Moreover, different aquaculture systems have a significant impact on the hepatic lipid metabolism and body shape of P. sinensis. Our results provide new insight into the influence of aquaculture systems on the microbial community structure feature and assembly mechanism in an aquatic animal, also highlighting the key microbiome and gene contributions to the metabolite variation in the gut microbiome‐metabolome association.