Marine Biology Seminar: Dr. Katherine Duncan (University of Strathclyde, Glasgow)
Title & Abstract: "Defining Patterns in Chemical and Genomic Space to Prioritize Antibiotic Discovery from Marine Actinobacteria"
Microbial drug discovery in the omics era relies on three key data types: biosynthetic, chemical and biological activity. Yet, integrating and interrogating these large and complex datasets remains a challenge and typically results in low-throughput prioritization of strains. Furthermore, linking metabolites to the BGCs responsible for their production and to observed bioactivity is limited and reliant on existing biosynthetic, chemical and antibiotic knowledge, thus overlooking unidentified parent ions (metabolites) or BGCs of unknown function, which constitute the exact chemical and biosynthetic space that should be prioritized to identify novel antibiotics. Here, we have generated two datasets of Actinobacterial strains to combine bacterial genomes (and their predicted BGCs), their chemical profiles (via strain fermentation and comparative metabolomics by molecular networking) and antibiotic screening (against clinically relevant pathogens). The first dataset consists of 25 Antarctic/Arctic rare-actinomycete genomes with almost 200 BGCs, combined with 100 metabolite profiles. The second dataset consists of 10 Streptomyces genomes and four rare-actinomycete genomes with almost 500 BGCs, combined with 80 metabolite profiles. Results from both datasets will be discussed as to how computational tools we have developed establish patterns across strains and learn relationships between BGCs, spectral features and bioactivity data.