This dissertation deals with the development of computational approaches and tools for large-scale comparative analysis of microbial genomes. We present three complementary, yet independent, systems designed for comparative analyses of microbial genomes in general and mycobacterial genomes particularly: LocalCOG, a relational database system built for local use of data from COG (Clusters of Orthologous Groups); BioParser, a program that offers graphical interfaces for parsing and analysing reports of sequence similarity searches; GenoMycDB, a relational database built for comparative analyses of mycobacterial genomes based on their predicted protein content.