Large-scale changes are common in genomes, and are often associated with pathological disorders. In the work presented in this dissertation, I provide insights into how inverted repeat sequences in budding yeast fuse during replication. Fusion leads to the formation of dicentric chromosomes, a translocation, and other chromosomal rearrangements.Using extensive genetics and some molecular analyses, I demonstrate that dicentric chromosomes are key intermediates in genome instability of a specific chromosome in budding yeast. I provide three pieces of evidence that is consistent with this conclusion. First, I detect a recombination fusion junction that is diagnostic of a dicentric chromosome (using a PCR technique). Second, I show a strong correlation between the amount of the dicentric fragment and the frequency of instability of the entire chromosome. Third, I demonstrate that a mutant known to stabilize dicentric chromosomes suppress instability. Based on these observations, I conclude that dicentric chromosomes are intermediates in causing genome instability in this system.Next, we demonstrate that fusion of inverted repeats is general. Both endogenous and synthetic nearby inverted repeats can fuse. Using genetics, I also show that many DNA repair and checkpoint pathways suppress fusion of nearby inverted repeats and genome instability. Based on our analysis, we propose a novel mechanism for fusion of inverted repeats that we term `faulty template switching.'Lastly, I discuss two genes that are necessary for fusion of nearby inverted repeats. I identified a mutant of the Exonuclease 1 (Exo1) and a mutant of anaphase inhibitor securin (Pds1) that suppress nearby inverted repeat fusion and genome instability. Studies of Exo1 and Pds1 provide us with insights into the molecular mechanisms of fusion.Our finding that nearby inverted repeats can fuse to form dicentric chromosomes that lead to genome instability may have great implications. The generality of this fusion reaction raises the possibility that dicentric chromosomes formed by inverted repeats can lead to genome instability in mammalian cells, and thereby contribute to a cancer phenotype.