An insight towards food-related microbial sets through metabolic modelling and functional analysis

Abstract: The dietary food digestion depends on the human gastrointestinal tract, where host cells and gut microbes mutually interact. This interplay may also mediate host metabolism, as shown by microbial-derived secondary bile acids, needed for receptor signalling. Microbes are also crucial in the production of fermented foods, such as wine and dairy. Kefir is fermented milk processed by the symbiotic community of bacteria and yeasts. One such species is a yeast Kluyveromyces marxianus . Its thermotolerance is a desired trait in biotechnology since it may reduce the cooling demands during cultivation. The systems biology tools allow analysing various size microbial communities under the different functional scope. For example, the homology prediction tools can give detailed functional insights when working with metagenomics data. The whole-cell metabolic processes can be summarised in genome-scale metabolic models (GEMs), which enable to predict the metabolic capabilities and allow for the integration of omics data. The work shown in this thesis includes i) in silico analysis of food-related microbes; ii) the development of GEMs and RAVEN. With a focus on bile acid metabolism, hundreds of human gut microbes were annotated based on metagenomics data, thereby suggesting the differences in the potential for bile acid processing between healthy and diseased subjects. These findings may be exploitable once aiming to restore the bile acid metabolism for the patients having inflammatory bowel disease. Also, the metabolism of yeast K. marxianus was characterised in genome-scale. Two K. marxianus strains from kefir grains were isolated, sequenced, assembled, and functionally annotated. They were compared with the other ten strains, providing the core and dispensable physiological features for K. marxianus . Furthermore, the first GEM for K. marxianus , namely iSM996, was reconstructed. It was integrated with transcriptomics data to predict its metabolic capabilities in rich medium and high-temperature conditions. The results might be useful to optimise strain-specific medium for high-temperature applications. The final paper comprises the efforts to improve the usability for RAVEN, a toolbox for GEM reconstruction and analysis. Altogether the outcomes of this thesis suggest the potential applications for medicine and industrial biotechnology, which may be facilitated by the newly upgraded RAVEN toolbox.