The Optimization Of A Saccharomyces Cerevisiae Genome-scale Metabolic Network Model And Its Applications

As a powerful tool for biological knowledge discovery, genome-scale metabolic network model(GMM) has been wide used in molecular evolution, metabolic engineering, phenotype prediction and so on. GMM is a kind of constraint-based model, which means that its performance in making predictions is largely determined by the constraints posed on the network. In this research, a new strategy for GMM refinement was proposed on a Saccharomyces cerevisiae GMM, which was used in the following biological research.Part 1. The analysis of metabolic cooperation in S. cerevisiae. Metabolic flux data from 96 published reports were systematically reviewed and several strong associations among metabolic flux phenotypes were observed. It was found that there is a significant correlation between glucose uptake and oxygen consumption, which was determined by the metabolic state the S. cerevisiae is in. However, the strong correlation between glucose uptake and ethanol production is not influenced by the metabolic state. In addition, glucose uptake correlates with glycerol secretion in some extent.Part 2. Optimization of S. cerevisiae. The relationship between glucose uptake and oxygen consumption, ethanol production, glycerol secretion were quantified with linear models. Obtained models were integrated into a S. cerevisiae GMM, thus constructing a refined model, which was named P-GMM. Experimental result showed the strong ability of P-GMM in predicting the growth rate of S. cerevisiae, proving the plausibility of our strategy in GMM refinement.Part 3. The application of P-GMM in the analysis of the influence of metabolic network on gene expression. P-GMM was used to estimate each reaction’s metabolic flux level and stochastic fluctuation level that a metabolic network could tolerate while maintaining homeostasis. It was revealed that metabolic flux correlates moderately with gene expression. Moreover, in the enzyme-dosage sensitive reactions, there is a strong correlation between metabolic flux fluctuation and gene expression noise. However, the correlation is largely affected by the reaction’s essentiality. For the non-essential reactions, metabolic network shapes gene expression noise. However, for the essential ones, regulation systems will be the dominant factor that controls gene expression noise level.Part 4. The application of P-GMM in the mechanism analysis of hydrogen sulfide production during wine fermentation. By mining the data generated from wine fermentation, it was revealed that the release of hydrogen sulfide hinders S. cerevisiae biomass accumulation, By designing an experiment and analyzing its result, it was concluded that the loss of sulfur during hydrogen sulfide release lead to the hinder effect. In addition, combining computational simulation on P-GMM and data mining in transcriptomic data, 21 genes were identified as the candidate genes that are most likely to be related to hydrogen sulfide production.