Some Computational Issues Of Metabolic Networks

Organism uses many inter-reactions to convert from nutrition to the ultimate metabolites. Metabolism is complex that many pathways interact together. In the abstract level, the metabolism can be taken as the complex networks that connect many metabolites by reactions and the functions of the organism are implemented by the complex metabolic network. It is one of the most important networks in biological system which bridges the gap between genotype and phenotype. To understand and analyze the metabolic network is one of the most important works in bioinformatics and system biology. We should investigate the biology system not only by single metabolite but also by many interacting biologic data. This dissertation focuses on a few computation issues in metabolic networks. The main results of this study are summarized as follows:1. Understanding the flux distribution of the metabolic network, especially for the null mutant, helps to strain design. However, the flux distribution computation is an open question that there is still no omnipotent model to accurately calculate it. We present a novel computation model to calculate the flux distribution. This model is practical for metabolic network analysis.2. The metabolic networks have many redundancy reactions that not all of these reactions are important for the organism. Though many methods (including flux balance analysis) have been put forward to predict the essential genes, they only focus on the model organism with many restrictive qualifications. As we all known that the gene essentiality can be speculated by reaction essentiality when the relationship of "gene-enzyme-reaction" is given. We should predict the essential genes from essential reactions by following three aspects: (1) simulate the minimal metabolic network for growth to get the most essential reactions; (2) calculate the impact degree for each reaction to whole metabolic network and estimate the reaction essentiality; (3) collect the reaction distribution among various organisms and estimate the reaction essentiality. Since the in vivo experiment to find the essential gene is time-consuming and laborious, these in silico methods should reduce the difficulty and complexity.3. We analyze the impact degree of each reaction to others, which is defined as the number of cascading failures of following and/or forward reactions when an initial reaction is deleted. Alternative metabolic pathways compensate null mutations, which represents that average impact degrees for all organisms are small. Interestingly, average impact degrees of archaea organisms are less than other two categories of organisms, eukayote and bacteria, indicating that archaea organisms have strong robustness to resist the various perturbations during the evolution process. The results show that scale-free feature and reaction reversibility contribute to the robustness in metabolic networks. The optimal growth temperature of organism also relates with the robust structure of metabolic network.4. A probable model presented here to illustrate the evolution process of the current metabolic networks from a few pre-biotic compounds. This evolution model is based on two aspects: reaction weight and metabolite connection. It is discovered that the reaction with large widespread extent among all organisms is essential for the organism survival. Because the essential reaction exists earlier than the nonessential one for most reactions, the reaction with larger widespread extent should emerge earlier. The evolution model is also determined by the strong connection that the substrates of the subsequently evolutive reaction should be existed. Given the reasonable primitive seeds, the evolutive steps of all current reactions in KEGG database are presented. Based on this evolution model, we further analyze the evolution of metabolic networks: (1) investigate when and how the oxygen comes forth; (2) investigate how metabolic network of single organism evolves; (3) evaluate whether reactions for optimal growth emerge in the early period of evolution process.