Computational Methods And Data Analysis For Metabolomics PDF Download

Metabolism provides energy for cells and organisms to grow, function and respond to their environment, so studying metabolism is fundamental to understanding how cells work. Ultimately, the ability to analyze and predict the effects of perturbations to metabolism will have profound impacts on applications ranging from bioengineering of microbes and plants to treating human disease. That understanding will come through the emerging field of metabolomics. The technology for acquiring high-throughput metabolomics data is advancing rapidly, as are efforts to curate comprehensive metabolic reaction networks. The goal of this work is to gain insight into metabolism by analyzing metabolomic data in the context of a metabolic network. In this study, first, we have developed a new metabolic network analysis method for genetic discovery. We have identified a new problem, which is to use steady state metabolomics data to find the underlying genetic cause of a phenotype. The solution is a qualitative analysis method which does not require quantitative reaction parameters and is robust even when the metabolomic data and network are incomplete. It could be used to predict genes that are responsible for metabolite concentration differences and identify drug targets. We validate the method on Saccharomyces Cerevisiae using single gene deletion mutants and drugs that were believed to target specific steps in metabolic pathways. Cells have evolved to maintain their metabolic homeostasis through various kinds of regulation. However, in current curated metabolic networks, the information about regulation is often missing. In second part of the study, we have developed a new computational method to predict regulatory targets in metabolic pathways using only steady state metabolite abundances. This method could be used to discover missing regulatory information in current curated metabolic networks. We demonstrate that the method can predict useful regulatory targets in Saccharomyces Cerevisiae ergosterol pathway. Third, large-scale in vivo measurements of the metabolome could potentially be used to estimate kinetic parameters for many metabolic reactions. However, in vivo measurements have special properties that are not taken into account in existing methods for estimating kinetic parameters. In vivo measurements of metabolite concentrations and reaction rates have relative errors. Also, reactions in metabolic networks often have multiple substrates and products and are sometimes reversible enzymatic reactions. Therefore, new method is needed to estimate kinetic parameters taking into account both factors. A new method, InVEst (In Vivo Estimation), is described for estimating reaction kinetic parameters, which addresses the specific challenges of in vivo data. InVEst uses maximum likelihood estimation based on a model where all measurements have relative errors. It can be applied to a family of reversible reactions with multiple substrates and products with single displacement mechanism. It also estimates the standard errors of parameter estimations using the bootstrap. Finally, we have developed a mass spectrometry data analysis tool for targeted analysis based on a database-driven algorithm. With this tool, compound identification is an automatic process which could help reduce the extensive manual processing required in current mass spectrometry data analysis.