In order to learn which genes are present in which locations, first I had to collect a large number of infected plants. Unfortunately, due to both sampling limited by the COVID19 pandemic and weather conditions not conducive to infection, I was only able to obtain dozens rather than hundreds of individuals, and so refocused work to answer related questions with the collection available. To do so, I first sequenced the genome of H. arabidopsidis and annotated protein-coding genes using the latest high quality genomic methods. This represents the first complete genome sequence in the genus Hyaloperonospora, and one of only a handful of oomycete genomes with experimental evidence-based annotation of proteins. Then, we tested which combinations of over 100 different strains of both H. arabidopsidis and A. thaliana lead to infections, a total of more than 10,000 different tests. From these data, I described the broad patterns across geographic space, including that a pathogen was more likely to infect hosts from nearby where it was collected than further away, suggesting that pathogens are locally adapted to the host immune systems nearby them. I then used statistical techniques to link genetic variation at specific pathogen genes to whether or not a pathogen could infect a given set of hosts, highlighting the genes that are likely being recognised by hosts. I also developed several pieces of software required to analyse these genomic datasets, and released them as open source software free for others to use in their own research. One software package is already published as an open access peer reviewed journal article, and I am finalising another peer-reviewed journal article describing the scientific advances made throughout this project.
