Finding the molecular switches behind new infectious diseases
DeepMind Blog · 2026-05-16
University of Cambridge Professor Clare Bryant used Google DeepMind's Co-Scientist AI to identify specific amino acid targets for infectious disease research, compressing what would normally require 2–3 years of experimental work into an estimated six months.
Appears in
Extraction
Topics: ai-for-scienceinfectious-diseaseco-scientisthypothesis-generationmolecular-biology
Claims
- Co-Scientist generated novel research hypotheses, including identifying a protein not previously on Bryant's radar, by analyzing her grant proposal.
- Iterative back-and-forth with Co-Scientist sharpened hypotheses from candidate proteins down to specific amino acids suitable for lab experiments.
- Work that would normally take 2–3 years to reach the amino-acid targeting stage is now projected to complete in six months.
- Co-Scientist can ingest unpublished research material and keep it confidential within the tool.
Key quotes
To get to the point of identifying precise amino acids would normally have taken two to three years of experimental work. But her lab is now on track to complete it in six months.
Co-Scientist had prioritised a protein that hadn't been on her radar, connected to several signalling pathways she was already interested in.