ReCodLiver0.9: Overcoming challenges in genome-scale metabolic reconstruction of a non-model species

Abstract

The availability of genome sequences, annotations and knowledge of the biochemistry underlying metabolic transformations has led to the generation of metabolic network reconstructions for a wide range of organisms in bacteria, archaea, and eukaryotes. When modeled using mathematical representations, a reconstruction can simulate underlying genotype-phenotype relationships. Accordingly, genome-scale models (GEMs) can be used to predict the response of organisms to genetic and environmental variations. A bottom-up reconstruction procedure typically starts by generating a draft model from existing annotation data on a target organism. For model species, this part of the process can be straightforward, due to the abundant organism-specific biochemical data. However, the process becomes complicated for non-model less-annotated species. In this paper, we present a draft liver reconstruction, ReCodLiver0.9, of Atlantic cod (Gadus morhua), a non-model teleost fish, as a practicable guide for cases with comparably few resources. Although the reconstruction is considered a draft version, we show that it already has utility in elucidating metabolic response mechanisms to environmental toxicants by mapping gene expression data of exposure experiments to the resulting model. Author summaryGenome-scale metabolic models (GEMs) are constructed based upon reconstructed networks that are carried out by an organism. The underlying biochemical knowledge in such networks can be transformed into mathematical models that could serve as a platform to answer biological questions. The availability of high-throughput biological data, including genomics, proteomics, and metabolomics data, supports the generation of such models for a large number of organisms. Nevertheless, challenges arise for non-model species which are typically less annotated. In this paper, we discuss these June 23, 2020 1/14 challenges and possible solutions in the context of generation of a draft liver reconstruction of Atlantic cod (Gadus morhua). We also show how experimental data, here gene expression data, can be mapped to the resulting model to understand the metabolic response of cod liver to environmental toxicants. 2 high-throughput genomic, proteomic, and metabolomic data, genome-scale metabolic 3 models (GEMs) have become fundamental tools in the systems biology of metabolism. 4 To develop a GEM, metabolic genes and their product enzymes are assembled into a 5 network of metabolic reactions carried out by an organism [42]. Such networks can be 6 used as a platform to answer relevant biological questions, by transforming biochemical 7 knowledge into a mathematical format and computing appropriate physiological 8 states [41, 46]. Using Boolean logic representation, gene-protein-reaction (GPR) 9 associations depict direct connections between genotype and metabolic capability [39]. 10 GEMs have a wide scope of applications, among which are the contextualization of 11 omics data, hypothesis-driven discovery, support for metabolic engineering, modeling 12 interact...

Metabolic network reconstruction of Euglena gracilis: Current state, challenges, and applications

“…Several automated tools for generating GEM were developed based on the available databases of the model species, and GEM modelling are particularly well-developed in prokaryotes. However, the metabolic network reconstruction processes are not as straightforward for eukaryotic and non-model species (Yan and Fong, 2017;Hanna et al, 2020), especially for ones with complex evolution causing the diverse and cryptic metabolisms, unique subcellular localisation of pathways and organelles, and without an available completed set of genome data, like Euglena. Allegedly, reconstructing the GEM of E. gracilis comes with several challenges that need to be overcome to enable the generation of a descriptive GEM.…”

ReCodLiver0.9: Overcoming Challenges in Genome-Scale Metabolic Reconstruction of a Non-model Species

“…This manuscript has been released as a pre-print (Hanna et al, 2020). We would like to thank Prof. Bernhard Palsson's group and the RAVEN group for joining the workshops and their helpful discussions.…”