Shengyun Peng Thesis Defense: "Understanding virus-host interactions through single cell and whole genome analysis" Thesis Advisor: Dr. Joshua Weitz, School of Biological Sciences and School of Physics, Georgia Institute of Technology Committee Members: Dr. King Jordan, School of Biological Sciences, Georgia Institute of Technology Dr. Ling Liu, School of Computer Science, Georgia Institute of Technology Dr. Justin Meyer, Division of Biological Sciences, University of California San Diego Dr. Frank Stewart, School of Biological Sciences, Georgia Institute of Technology Abstract Viruses and their microbial hosts are widely distributed in the environment, including in oceans, soils, fresh water, and even in extreme environments such as the deep ocean, hot springs and the upper atmosphere. Given the ubiquity of viruses of microbes, it is critical to understand virus-host interactions and their effects on ecosystem functioning. My work addresses the problem of virus-host interactions through three motivating questions: 1) to what extent do viruses and hosts interact in a given environment and who interacts with whom, 2) how do interactions shape the coevolutionary dynamics of viruses and hosts and 3) what is the genetic basis for determining both who infects whom and the efficiency of viral infections. Here, I report findings stemming from analysis of virus-host interactions in a natural environment (Yellowstone National Park hot springs) and from an experimental study of coevolution in vitro. First, I characterized virus-host interactions in a hot springs environment, combining evidence from single-amplified genomes and metagenomes to characterize a natural virus-host interaction network, finding that the majority of cells were infected by one (or more) viruses. Second, I developed a new approach to infer the genetic basis for both qualitative and quantitative changes in virus-host interactions unfolding during coevolution. In doing so, I leveraged whole genome analysis to identify novel mutational candidates that could drive large-scale changes in infectivity; the approach can also be applied to characterize the genotype-phenotype map in other phage-host systems. Overall, the findings help deepen our understanding of virus-host interactions and the consequences of infection on complex virus and microbe communities.