Combining biomarkers from salmon metabonomics, genomics and proteomics with high-field NMR profiling of fish feed nutrients

Improvements of salmon feed are complex and very expensive, because evaluation of new ingredients and their formulation is by long term feeding trials. The main objective of this program is to develop biomarkers to reduce the time and expense involved wit h feeding trials. Increased speed will allow EWOS Innovation to perform more trials based on early metabolic consequences of the ingredients. Secondly, the aim is to gain new insight into basic mechanisms of fish metabolism. A major challenge is that the most expensive protein ingredient is fish meal which is highly variable in price and quality, and needs to be partially replaced by plant protein sources. The problem with using vegetable protein sources is that the amino acid composition does not suppor t the salmon requirements. Balancing the dietary composition is of utmost importance for synthesis of protein, and proper anabolic signaling from tissue amino acids. Selection of raw materials to accomplish this can be aided by precise and rapid character ization using high-field NMR. The effect of specific ingredients on growth will be aided by detailed knowledge of growth regulation in salmon using genomic, proteomic and metabonomic platforms. Molecular biomarkers provide an early sign of a change in an organism's physiological status, thus permitting detection of early feed performance changes. The objective is a suite of biomarkers to predict the biological consequence of feeding fish a particular diet without having to conduct expensive and time-consu ming feed trials. Group feeding and sampling has been the traditional method which needs to be replaced by individual fish trials that are adapted to -omics technology. The statistical challenge will be to combine these multiple blocks of multivariate dat a to extract essential cross-sectional information related to the single object under study. A main statistical task in this project will be to optimize these and related statistical methods for analysis.