Evaluating Sphingolipid Biochemistry in the Consensus Reconstruction of Yeast Metabolism

Abstract

Reconstructed metabolic networks are the basis for genome-scale models of cellular metabolism. Such models have demonstrated good predictive accuracy when analyzed with the optimization-based approach of flux balance analysis, so are of particular interest for informing the rational design of metabolic engineering strategies for industrial biotechnology applications. As a preliminary step towards applying a metabolic model of the industrially important yeast Saccharomyces cerevisiae, we validated the underlying reconstruction of yeast metabolism in the Yeast Consensus Reconstruction by examining the model’s representation of yeast sphingolipid metabolism. Sphingolipids, which comprise 30% of the plasma membrane lipids in yeast, incorporate very long chain fatty acids of lengths up to C-26. Such long-chain hydrocarbons are of particular interest for the production of liquid transportation fuels. We began with a literature review and consultation with experts in yeast sphingolipid metabolism to compile a list of 243 reactions associated with yeast sphingolipid metabolism. We found that the Yeast Reconstruction contains only 41 reactions in this pathway, and that the reconstruction includes inaccuracies. This documentation of the incomplete reconstruction of sphingolipid metabolism in the Yeast Consensus Metabolic Reconstruction demonstrates an opportunity for improving the computational reconstruction of established biochemical knowledge. The extent of information missing from the Yeast Reconstruction suggests that incorporating knowledge from the primary literature is an underemphasized tool for refining genome scale metabolic models. Our expert-knowledge driven and pathway-centric approach to reconstruction curation provides a rational approach for community participation in the ongoing effort to reconstruct the yeast biochemical reaction network.