Google DeepMind published a technical walkthrough on October 15 showing how biologists used Google Antigravity, its agentic development platform, to search for new antimicrobial peptides.
What was done
Antigravity acts as an agentic science workbench that lets researchers discover, structurally evaluate and prioritize targets in a single session, rather than searching several databases by hand. It helped scan millions of uncharacterized proteins in UniProt using ProtNLM, a model that predicts protein function from amino acid sequences. The team reports one novel candidate with predicted activity against pathogens that cause respiratory tract infections.
Caveats
The result is a computational prediction. DeepMind says the lead is now undergoing wet-lab validation, so it is not yet shown to work as a treatment. The company frames the work against rising antimicrobial resistance, which it says could cause an estimated 10 million deaths a year by 2050.
Why it matters
The post is a worked example of agents handling the tedious search stage of drug discovery, with lab testing still the deciding step.