Beyond Microbial Variability: Disclosing the Functional Redundancy of the Core Gut Microbiota of Farmed Gilthead Sea Bream from a Bayesian Network Perspective.

The significant microbiota variability represents a key feature that makes the full comprehension of the functional interaction between microbiota and the host an ongoing challenge. To overcome this limitation, in this study, fish intestinal microbiota was analyzed through a meta-analysis, identifying the core microbiota and constructing stochastic Bayesian network (BN) models with SAMBA. We combined three experiments performed with gilthead sea bream juveniles of the same hatchery batch, reared at the same season/location, and fed with diets enriched on processed animal proteins (PAP) and other alternative ingredients (NOPAP-PP, NOPAP-SCP). Microbiota data analysis disclosed a high individual taxonomic variability, a high functional homogeneity within trials and highlighted the importance of the core microbiota, clustering PAP and NOPAP fish microbiota composition. For both NOPAP and PAP BNs, >99% of the microbiota population were modelled, with a significant proportion of bacteria (55-69%) directly connected with the diet variable. Functional enrichment identified 11 relevant pathways expressed by different taxa across the different BNs, confirming the high metabolic plasticity and taxonomic heterogeneity. Altogether, these results reinforce the comprehension of the functional bacteria-host interactions and in the near future, allow the use of microbiota as a species-specific growth and welfare benchmark of livestock animals, and farmed fish in particular.