<p>Orthology identification is central to comparative and evolutionary genomics and is an active area of research. Despite a recent shift towards tree reconciliation and other phylogenetic methods, previous comparisons between different algorithms relied on real datasets where true orthology relationships are unknown and did not conclusively show whether phylogenetic methods truly outperform sequence similarity-based methods. Using simulated datasets generated from programs we developed, we show that tree reconciliation does perform better than similarity-based methods when the true species phylogeny is known. Even slight deviations in the species phylogeny can have adverse effects on the performance of reconciliation algorithms and in those cases similarity-based methods may perform better. Fusion and fission complicate orthology identification and are not explicitly considered in most existing algorithms. Programs designed specifically to investigate fusion and fission events are either unavailable or are not specific enough to identify events affecting orthologous genes. We developed a pipeline of programs called FusionFinder that perform this task, gaining new insights to the contributions of fusion and fission to bacterial protein evolution and uncover an unexpected abundance of fissions in <em>Bacillus anthracis</em> that to our knowledge yet to be reported.</p> / Master of Science (MS)