Leptospirosis, caused by the bacterial spirochete Leptospira, is a tropical neglected disease infecting many mammalian hosts and leading to a large amount of morbidity and mortality in humans and livestock around the world. Despite the caused large social and economic damages, the molecular mechanisms underlying Leptospira’s pathogenicity are still not fully understood. With the advancement of Next Generation Sequencing technologies, modern Leptospira research has an unprecedented opportunity to further our understanding of this complex organism. In this thesis, I tested and developed computational methods to facilitate the understanding of Leptospira genome evolution and structural compositions across biological scales using multi-omics approaches. First, I used metagenomics to evaluate some of the most widely used direct read shotgun metagenomics taxonomic profiling software and databases to determine their accuracy and sensitivity in detecting Leptospira from biological samples. I showed that the discrepancies from different software and databases’ profiling results could lead to significant variations in the distinct microbial taxa classified and, thus, cause false positives or false negatives in the detection of Leptospira, especially at the species level resolution. Second, I used transcriptomics to characterize the transcriptomes of different Leptospira serovars using the long-read platform of Oxford Nanopore Technologies. I determined novel RNA molecules and compared the transcriptomes of pathogenic and non-pathogenic Leptospira to identify signals of Leptospira pathogenicity. I also provided evidence for the existence of posttranscriptional polyadenylation for Leptospira RNA expression regulation and the use of ONT sequencing without polyadenylation as a tool to improve our understanding of prokaryotic RNA polyadenylation. Third, I characterized a whole genome sequence dataset of Leptospira collected from public databases to determine associations between genetic variations identified from different Leptospira genome sequences and their pathogenicity. I designed and implemented two automatic workflows, BactASM and BactPrep, for the cleaning, organization, annotation, and characterization of large whole-genome sequence datasets. Using these workflows, I identified genetic variations between and within different Leptospira species genomes. These variations reflect the history of Leptospira’s non-clonal evolutionary mechanisms related pathogenesis. Overall, this thesis provides a comprehensive omics framework that furthers our understanding of Leptospira detection and evolution at multiple biological scales.