Bacterial endosymbionts hijack host cell biology by secreting effectors that mimic eukaryotic protein function. These effectors encode domains that are homologous to eukaryotic toxins and tend to cluster inside phage genomes that infect endosymbionts. Phage-associated toxins can have dramatic impacts on host cell biology. For example, expression of phage-associated toxins underlies cytoplasmic incompatibility (CI) between males and females infected with the maternally transmitted endosymbiont Wolbachia in Drosophila and mosquitoes. But, expression of non-phage-associated toxins does not appear to affect CI. It remains unclear, therefore, whether association with the phage genome itself impacts the function and evolution of endosymbiont toxins. Tandy Warnow and Irene Newton address this question by comparing the functional diversity and molecular evolution of phage- versus non-phage-associated toxins from Wolbachia. Over the last ten years, hundreds of diverse endosymbiont and associated phage genomes have been sequenced, but the impact of this diversity on the coevolution of host-symbiont interactions remains unclear. This project aims to capture much of the functional arms race between host, symbiont, and phage by characterizing the diversity and evolution of horizontally transferred endosymbiont toxins.