In this thesis I look at two different problems in bacterial genomic analysis. The first involves reconstructing the evolutionary history between a group of closely related bacteria. I addressed whether or not it is possible to separate such genomes into different genera, species and strains. Specifically, I addressed how different approaches such as the use of 16S rRNA phylogenetic trees, phylogenetic supertrees and concatenation of individual genes in order to construct phylogenetic trees compare with one another. What effect will problems associated with resolving shallow-phylogeny have on recovering a tree of life? Ultimately I show that for the group of genomes involved, different methods and data produce different results and that the true tree, if a tree-like structure does indeed exist for these genomes, is unrecoverable using such approaches. In the second part of my thesis I examine the phenomenon of gene clustering in bacterial genomes. I present a software program, GenClust, for the identification, analysis and visualisation of gene clusters. I show how GenClust can be used to recover and analyse clusters of genes involved in amino acid biosynthesis across a large !-proteobacterial dataset. Finally, I examine models of gene cluster and operon formation and test them with real data, using a combined approach of comparing clusters on both structural similarity and the underlying phylogenetic signals of the clustered genes. I provide a hypothesis for the selective forces driving cluster and operon formation in bacterial genomes.