The ecology and evolution of pathogens are key factors in predicting the severity and spread of disease, as well as treatment outcomes. However, the effects of multiple trophic levels that include host, microbial competitors and viruses are typically overlooked. In this thesis I develop our understanding of bacteria-phage coevolution, microbial dispersal and the role of the microbiome in disease. The results of these experiments have direct implications for phage therapy: the use of bacteriophages to treat bacterial infections. Firstly, I explore the risks of phage application in the environment and draw parallels with the misuse of antibiotics in selecting for bacterial resistance. I then demonstrate that the evolution of resistance to phages in a plant pathogenic bacterium is context-dependent. Notably, I find a fitness cost in plant infections that is absent when the bacteria are cultured solely in the laboratory. I then characterize four novel phages and use a simple laboratory based assay to predict their potential as phage therapy agents in an agricultural context. Next I show that reservoir species of plant hosts can affect the evolution of virulence, when bacteria are passaged on both a focal and distant host, but find no evidence of local adaptation. I also show that the evolution of such traits can occur in a parallel manner at the genetic level. I then determine a compositional shift in the microbiota associated with the symptoms of bleeding canker disease in Horse Chestnut trees across the length of the UK. Finally, I find an age-elated decline in bacterial species richness and evidence for niche-assembly theories by investigating bacterial dispersal in UK Oak trees in a single woodland.