Presented by: Georgia Doing
Of the trillions of microbial cells associated with a human body at any given point in time, about half are identifiable at the genus level, a quarter are culturable, and only a handful have been isolated and extensively studied in the laboratory over decades. Analogous to how model organisms, such as mice, have been used to study human biology, “transfers” of the great depth of knowledge accumulated for model microbial species, such as the pathogen Staphylococcus aureus, to related but less-studied microbes such as the commensal and opportunistic pathogen Staphylococcus epidermidis, will greatly facilitate understanding microbial diversity and microbial communities. To date, transfer learning has been successful in the fields of image, video and natural language processing and has been applied in genomics to bridge different mammalian cell types and is starting to be used to connect different species, including plant and insect model and non-model organisms. The genetic diversity and transcriptional plasticity of S. epidermidis is sparsely annotated, and we aim to apply transfer learning to genetic, transcriptomic and functional data to integrate the wealth of S. aureus data collected over previous decades and drive hypothesis generation around S. epidermidis pathogenicity.