Comparative co-occurrence network analyses of the cichlid fish gut microbiota: community conservation and diet-associated shifts

Co-occurrence networks of bacteria associations are a powerful approach to explore ecologically relevant aspects of the gut microbiota structure, beyond community composition alone. Here we exploit the remarkable diversity of cichlid fishes and their multiple lake assemblages to investigate a) network features and patterns of microbial associations that were robust to major phylogeographical variables, and b) community structure changes along cichlid dietary shifts. We tackled these objectives using the large gut microbiota sequencing dataset available (nine lakes from Africa and America), building geographical and diet-specific networks and performing comparative analyses. Major findings indicated that lake and continental networks were highly resembling in global topology and node taxonomic composition, suggesting important constraints in the cichlid gut community assembling. A small fraction of the observed co-occurrence pairwises was conserved across all lake assemblages; while the origin and ecological relevance of these core associations remains unclear, their persistence suggests a potential functional role in the cichlid gut. Comparison of carnivores and herbivores-specific networks as well as mapping of diet-specific values on the African Lake Tanganyika network revealed a clear community shift as a function of diet, with an increase in complexity and node taxonomic diversity from carnivores-omnivores-plantkivores to herbivores. More importantly, diet-associated nodes in herbivores formed complex modules of positive interactions. By intersecting results from association patterns and experimental trials, future studies will be directed to test the strength of these microbial associations and predict the outcome of community alterations driven by diet. ImportanceThe gut microbiota is a complex community of interacting bacteria. Predicting patterns of co-occurrence among microbes can help understanding key ecological aspects driving community structure, maintenance and dynamics. Here we showed a powerful application of co-occurrence networks to explore gut bacteria interactions in a primary model system to study animal diversification, the cichlid fishes. Taking advantage of the large scale of phylogeographical and ecological diversity of this fish family, we built gut microbiota networks from distinct lake and continental fish assemblages and performed extensive comparative analyses to retrieve conserved and trait-specific patterns of bacteria associations. Our results identified network features that were independent from the fish biogeography and that indicated an important host selection effect on gut community assembling. Focusing on a single lake assemblage, and therefore excluding the major geographical effect, we observed that the gut microbiota structure dramatically shifted from carnivore to herbivore fishes, with a substantial increase in the number and complexity of microbial interactions.