Metabolic Modeling of the Gut Microbiome-Host Interactions and Meta’omics Integration

Doctoral thesis, 2015

A large number of microbes with different strain types occupy the human gut. These gut inhabitant microbes have key roles in decomposition of indigestible dietary macronutrients before they are metabolised by the host. The gut microbiome has a clear effect on human health and perturbations in its diversity may lead to the development of disorders through changes in metabolic functions. To date, different studies have shown the association of the gut microbiome with metabolic disorders such as obesity, type2 diabetes and certain cancers. It has also been shown that there is a complex interaction between microbe-microbe, host-microbe and microbe-diet, and elucidation of the mechanisms behind these interactions and associations remains a challenge. Due to the extreme complexity of cellular metabolism, mathematical models may be employed for deciphering the role of its individual elements and may thereby assist in providing an increased understanding of these interactions. The emerging research field of systems biology can integrate different highthroughput data, in this case metagenomics and metatranscriptomics, through the use of mathematical models and thus provide a holistic interpretation for this complex system. In this context, genome-scale metabolic modeling has been applied to gain increased knowledge in important biotechnology applications. This thesis presents approaches to facilitate understanding of the causalities and go beyond the association analysis by considering the interactions between microbiome, host and diet. Using genome-scale metabolic models (GEMs), we investigated the contribution of key species in the overall metabolism of the gut microbiome. We developed methods and generated stand-alone software to apply for different case studies on modeling of gut microbiome and finally addressed relevant biological questions. First, GEMs for three bacteria being representatives of dominant phyla in the human gut microbiome were reconstructed. This modeling approach allowed us to establish effective resources for understanding the microbe interactions in the gut. Increasing the number of relevant GEMs representing all key microbes in the human gut resulted in more complexity and therefore we developed the CASINO toolbox, a comprehensive software platform for the analysis of microbial communities. CASINO was validated based on in-vitro studies and thereafter applied to human studies that showed its capability to predict the phenotype of individuals based on their dietary pattern and gut microbes’ abundances. Finally, the application of CASINO was extended and used for modeling of the interactions between gut microbiota and host metabolism. The overall metabolic differences between germ-free and conventionally raised mice were revealed through the use of CASINO. In conclusion, this thesis provides a new approach to human gut analysis by using valuable resources (GEMs) and novel methods (CASINO). As such it contributes to advancing the role of metabolic modeling in human health and designing new clinical interventions.