Reconstruction and analysis of tissue-specific models of Atlantic salmon metabolism

Abstract

The set of all biochemical reactions taking place in a living organism, converting nutrients to energy and growth, is referred to as metabolism. Metabolism is highly interconnected, forming metabolic pathways that can be mathematically formulated as graphs, with nodes (metabolites) connected by edges (reactions). By integrating information from sequenced and annotated genomes, a metabolic network may be extended to a genome-scale model (GEM), computationally representing all known metabolic information of the organism, as well as including all associated genomic attributes. GEMs constitute a predictive framework for simulating and quantifying the flow of biochemical compounds, referred to as metabolic fluxes, by formulating mechanistic genotype-phenotype relationships. Such tools facilitate research where a thorough understanding of metabolic processes is of interest. An example is the aquaculture industry, where GEMs simulate the effect of different feed compositions on the fish. However, not all metabolic pathways are active under all cellular states, making GEMs superfluous. In this study, context-specific metabolic models are extracted from an Atlantic salmon GEM using an experimental transcriptomic data set. Through a novel approach, these models are subsequently combined and generalized to constitute tissue-specific reconstructions of the Atlantic salmon gut, liver, and muscle. The tissue-specific metabolic models agree better with the expression data than the GEM, corroborating the notion that such models better capture the metabolic state of biological systems. The reconstructions are analyzed by utilizing constraint-based modeling (CBM), revealing key characteristics of metabolism in the three tissues. More specifically, the variation in possible flux states is investigated, revealing the metabolic scope and range of the three tissues. Furthermore, the study challenges the very definition of tissue-specific metabolic activity by investigating the agreement between model- and data-driven interpretations of the term. The results indicate that the agreement between the two definitions is low, supporting the utilization of approaches that consider biological constraints when analyzing metabolism. The findings of this study contribute to the current knowledge in Atlantic salmon systems biology research and will be beneficial for tissue-level simulation studies of the Atlantic salmon metabolism.