Antibiotic resistance is a tough problem—one some biologists worry cannot be solved with a business-as-usual approach to drug development. So researchers led by Gerard D. Wright at McMaster University tried out a new way of searching for antibiotic compounds. The team built a phylogenetic tree of gene clusters that code for glycopeptide antibiotics and then looked for self-resistance genes, which help bacteria protect themselves from the antibiotics that they make. The researchers found resistance genes for glycopeptide antibiotics on one large branch of the tree but not another, suggesting to the researchers that the antibiotics from the branch without the resistance genes might be part of a new functional family. So they searched that branch for antimicrobial compounds and purified two of them, including a newly found metabolite they named corbomycin. Corbomycin treats methicillin-resistant Staphylococcus aureus infections in mice, and it works in a novel way. The team tested the compounds and found that they prevent the breakdown of the peptidoglycan matrix in bacterial cell walls, effectively “caging” them so they cannot grow or divide (Nature 2020, DOI: 10.1038/s41586-020-1990-9).