Human-microbiota interactions in health and disease :bioinformatics analyses of gut microbiome datasets

Abstract

The human gut harbours a vast diversity of microbial cells, collectively known as the gut microbiota, that are crucial for human health and dysfunctional in many of the most prevalent chronic diseases. Until recently culture dependent methods limited our ability to study the microbiota in depth including the collective genomes of the microbiota, the microbiome. Advances in culture independent metagenomic sequencing technologies have since provided new insights into the microbiome and lead to a rapid expansion of data rich resources for microbiome research. These high throughput sequencing methods and large datasets provide new opportunities for research with an emphasis on bioinformatics analyses and a novel field for drug discovery through data mining. In this thesis I explore a range of metagenomics analyses to extract insights from metagenomics data and inform drug discovery in the microbiota. Firstly I survey the existing technologies and data sources available for data mining therapeutic targets. Then I analyse 16S metagenomics data combined with metabolite data from mice to investigate the treatment model of a proposed antibiotic treatment targetting the microbiota. Then I investigate the occurence frequency and diversity of proteases in metagenomics data in order to inform understanding of host-microbiota-diet interactions through protein and peptide associated glycan degradation by the gut microbiota. Finally I develop a system to facilitate the process of integrating metagenomics data for gene annotations. One of the main challenges in leveraging the scale of data availability in microbiome research is managing the data resources from microbiome studies. Through a series of analytical studies I used metagenomics data to identify community trends, to demonstrate therapeutic interventions and to do a wide scale screen for proteases that are central to human-microbiota interactions. These studies articulated the requirement for a computational framework to integrate and access metagenomics data in a reproducible way using a scalable data store. The thesis concludes explaining how data integration in microbiome research is needed to provide the insights into metagenomics data that are required for drug discovery.