Google DeepMind said on Oct 15, 2026 that biologists used Google Antigravity, its agentic development platform, to uncover a novel antimicrobial peptide candidate with predicted activity against pathogens involved in respiratory tract infections; the lead is now undergoing wet-lab validation. DeepMind says the workbench compresses a search once spread across multiple databases into a single session, and scans millions of uncharacterized proteins in the UniProt database using ProtNLM, a model that predicts biological function directly from raw amino-acid sequences to annotate previously unnamed entries. The post ties the effort to rising antimicrobial resistance, which DeepMind says could cause an estimated 10 million deaths a year by 2050, and to the unannotated "dark matter" of protein databases.
Our takeThis is the best flavor of AI-for-science: an agentic tool narrowing a vast search and handing a concrete candidate to wet-lab validation, where it still has to prove out. The compelling part isn't a model inventing an idea but a workflow that links databases, prediction and human judgment end to end. The caveat is that predicted activity is not activity — until the lab confirms it, the achievement is speed, not a drug. DeepMind is careful to say exactly that.