Although understood in many vertebrate systems, the natural diversity of host-associated microbiota has been little studied in teleosts. For migratory fishes, successful exploitation of multiple habitats may affect and be affected by the composition of the intestinal microbiome. We collected 96 Salmo salar from across the Atlantic encompassing both freshwater and marine phases. Dramatic differences between environmental and gut bacterial communities were observed. Furthermore, community composition was not significantly impacted by geography. Instead life-cycle stage strongly defined both the diversity and identity of microbial assemblages in the gut, with evidence for community destabilisation in migratory phases. Mycoplasmataceae phylotypes were abundantly recovered in all life-cycle stages. Patterns of Mycoplasmataceae phylotype recruitment to the intestinal microbial community among sites and life-cycle stages support a dual role for deterministic and stochastic processes in defining the composition of the S. salar gut microbiome.

In fisheries management, intensive stocking programs are commonly used to enhance population abundance and maintain stock productivity. However, such practices are increasingly raising concerns as multiple studies documented adverse genetic and evolutionary impacts of stocking on wild populations. Improvement of stocking management relies on a better understanding of the dynamic of introgressive hybridization between wild and domestic population and on assessment of the genetic state of wild populations after stocking cessation. In Québec, Canada, over five million captive‐reared Brook Charr (Salvelinus fontinalis) are stocked every year to support recreational fishing activities. Here, we investigated how variation in stocking history and environmental variables, including water temperature, pH, and dissolved oxygen, may influence the impact of stocking practices on the genetic integrity of wild Brook Charr populations. We collected DNA samples (n = 862, average of 30 individuals per lake) from 29 lakes that underwent different stocking intensity through time and also collected environmental parameters for each sampled lake. An average of 4,580 high‐quality filtered SNPs was obtained for each population using genotyping by sequencing (GBS), which were then used to quantify the mean domestic membership of each sampled population. An exhaustive process of model selection was conducted to obtain a best‐fitted model that explained 56% of the variance observed in mean domestic genetic membership. The number of years since the mean year of stocking was the best explanatory variable to predict variation in mean domestic genetic membership whereas environmental characteristics had little influence on observed patterns of admixture. Our model predictions also revealed that each sampled wild population could potentially return to a wild genetic state (absence of domestic genetic background) after stocking cessation. Overall, our study provides new insights on factors determining level of introgressive hybridization and suggests that stocking impacts could be reversible with time.