Bioinformatics tools for cell modeling and metabolomics using Lactococcus lactis as a model organism

Abstract

This decade’s advances in molecular biology coupled to the need for understanding the cell’s metabolism as a system, established the field of Systems biology. Genome-scale metabolic models joined to high-throughput measurements, such as metabolomics, are largely applied to reconstruct the full picture of the cell’s metabolism and its interaction with the environment. Metabolomics is one of the newest omics technologies, yet it is already considered essential to systems biology studies. Together, systems biology and metabolomics are considered revolutionary fields able to address fundamental questions that enhance our understanding of biological systems. Still, the reconstruction of genome-scale metabolic networks, the processing and interpretation of metabolomics data, and its use for feeding systems biology studies remain limiting factors for these fields achieving their full potential. In this thesis, I present algorithms that I developed to address these limiting factors. To assist the reconstruction of genome-scale metabolic networks, I present a set of algorithms able to merge existing networks and find potential metabolic gaps. To improve the data processing, I present MetaBox. MetaBox deconvolutes chromatogram peaks and identifies metabolites analyzed by gas chromatography–mass spectrometry with at least 20% higher accuracy reported by the most commonly used software for this purpose, namely Automated Mass Spectral Deconvolution and Identification System (AMDIS). In addition, MetaBox automates the normalization steps in, and basic statistical analyses of, metabolomics data. To facilitate the interpretation of metabolomics data, I present Pathway Activity Profiling (PAPi), which generates hypotheses by predicting the activity of metabolic pathways based solely on metabolomics data sets. Finally, I present Metabolite to Network (M2N), which facilitates the usage of metabolomics data for feeding metabolic models by predicting enzymes and reactions related to the organism under study. The usage and the power of these new algorithms are demonstrated using the lactic acid bacteria Lactococcus lactis as a model biological system. Among the significant conclusions related to L. lactis’ metabolism are its potential to undergo respiration by using cyanocobalamin as a heme precursor and a newly reconstructed metabolic network containing previously unseen reactions involving the degradation and biosynthesis of fatty acids in this bacterium. The findings from the metabolism of L. lactis, the algorithms and the computational tools described in this thesis represent novel contributions to the fields of metabolomics and systems biology.