| Development of sequencing technologies makes it possible to access the microbial diversity on the molecular level. Through sequencing of microbial genomes and metagenomes, insights are gained into the metabolic roles played by the microbes and the formation of complex microbial communities. In order to obtain comprehensive understanding of the microbial diversity, it is essential to know what makes the microbes different from each other, what shapes the microbes and microbiomes to adapt to diverse environments, and how to access the microbial world where a majority of the microbes are yet novel. Despite of successes in literature that identified structural or metabolic differences between various microbes and microbiomes, the description and interpretation of microbial diversity is yet controversial and often dependent on the level of detail one may want to conclude. For example, microbiome structures can be resolved on various taxonomic levels, and microbiome dynamics can be explained with biological processes, pathways or various molecular functions. Consequently, thorough study of microbial diversity is essential and the methods used are supposed to be adaptable to varying demands.;This dissertation aims to improve our current understanding towards microbial diversity in varying microbes and microbiomes, by methods that are generally adaptable in comparative genomics and comparative metagenomics. As the fundamental prerequisite for studying microbial diversity, genomic comparisons are performed. Methods are developed to predict novel 16S rRNA genes and to benchmark alternative marker genes for enhanced inference of phylogenetic relationship between microbes. A genomic comparison database is also constructed to provide public access to comparing all prokaryotic complete genomes. Furthermore, feature selection methods are introduced to identify microbial signatures that are responsible for varying environmental factors in comparative genomics or metagenomics studies. By implementing feature selection, this dissertation proves the possibility to identify protein signatures in human gut microbiomes that are relevant to age, and to identify pathway signatures differentially expressed in a fungal organism that help explain how the organism reacts to alternative nutrient. As a method adaptable for various data types, feature selection can be widely used to explain the microbial diversity in microbes or microbiomes under varying circumstances. Finally, the dissertation aims to provide access to the novel members of microbiome, which represent a more diverse microbial world than in our current knowledgebase. Using a pipeline with comparative genomics and metagenomic approaches combined, a novel microbial genome enriched in a soil microbiome is found and putatively recovered in comparison with its close relatives. With developed understanding of microbial genomes, microbial signatures shaping the environment and access to the novel microbes in microbiomes, the interpretation of microbial diversity is greatly improved. |
