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

Doktorsavhandling, 2020

Nature contains numerous examples of how living forms adapt to their niche environments. One may assume that the organisms exhaustively compete for the same limited food resources, but the species diversity in biochemical level suggests that they might also complement each other and develop somehow mutual interactions. Such interactions may sometimes become essential and even form the basis of the local communities like in soil, food, seawater or the human gut. Although much work has been done to identify the community-driven adaptation mechanisms, there are still many communities worthwhile to be annotated. A better understanding of these communities and their member species may highlight new candidates for applications in biofuel or biopharmaceuticals production, thereby contributing to a sustainable future.
This thesis includes several bioinformatics analyses for food-related microbial species and presents the newly upgraded toolbox for metabolic modelling. The systems biology tools allow annotating microbial communities of various sizes. While metagenomics data is a crucial source for community characterisation, the genome sequencing data is the primary source during the reconstruction of a genome-scale metabolic model (GEM), a tool allowing to predict the cell behaviour in a given condition.
Firstly, hundreds of human gut microbes were annotated with a focus on bile acid metabolism and were then used to compare bile acid biotransformation potential between healthy and diseased patient groups. These findings may be exploitable once aiming to restore the bile acid metabolism for the patients having inflammatory bowel disease.
Secondly, a non-conventional yeast Kluyveromyces marxianus was isolated from kefir grains and compared with its other strains, providing the core and accessory physiological features for this species. The first K. marxianus GEM was also reconstructed and integrated with transcriptomics data to predict its metabolic capabilities in rich growth medium and high-temperature (45°C) conditions. The results might be useful to optimise strain-specific medium for high-temperature industrial applications.
Finally, this thesis presents the efforts to improve the usability for RAVEN, a toolbox for GEM reconstruction and analysis. Overall, the results of this thesis hint the potential applications for healthcare and industrial biotechnology, which may be facilitated by the newly upgraded RAVEN toolbox.