Genetically modified plants as fish feed ingredients

Commonly used plant ingredients in salmon feed such as soybeans and maize are increasingly grown as genetically modified (GM) varieties. The question arises whether these are equally nutritious and safe for the fish as conventional varieties. The target species of this thesis is Atlantic salmon (Salmo salar), while zebrafish (Danio rerio) is used as a potential model. Extensive evaluations of performance, health, organ function and stress response were carried out in a 7-month salmon feeding trial with GM soy (25%). A smaller study was conducted on zebrafish evaluating both GM soy and GM maize as feed ingredients, including tracing of dietary DNA in fish tissues. Feeding salmon GM soy did not affect growth, body composition, haematology or weights of liver, spleen, head-kidney and proximal intestine, compared to non-GM soy. Distal intestine was larger in the GM group at one sampling point, but not at the end of the trial. Mid intestine was consistently smaller in the GM group. Typical soybean meal (SBM) -induced alterations in the distal intestine were seen in both groups. However, GM fed fish had shorter mucosal fold height at one sampling point, and more mucosal fold fusion overall in the trial, which could indicate slightly more pronounced SBM-induced changes. In liver, glycogen deposits were lower in salmon fed GM soy. Apart from this, no morphological differences were detected in any organs. Of parameters related to health, there were no differences in haematology, proportion of leukocytes types, lysozyme concentrations in spleen and head-kidney, or concentrations of plasma enzymes. Plasma triacylglycerol (TAG) was higher in the GM group overall in the trial, while no diet effects were seen on plasma glucose and protein. Fish in both diet groups responded similarly to seawater transfer; they adapted well and rapidly regulated chloride ion content in plasma back to basal levels. Similarly, there were no differences in stress response between the two diet groups. Proteomic comparisons of livers from fish in the two diet groups exhibited small fold differences, and differentially expressed proteins all had a high false discovery rate. The diet groups were not distinguishable by principal component