How Co-Scientist Helps Find Molecular Switches of New Infectious Diseases

Professor Clare Bryant at the University of Cambridge is using Google DeepMind's Co-Scientist to hunt for molecular switches that cause severe diseases like sepsis when pathogens leap from animals to humans. She first tested the tool by feeding it a summary of a grant proposal studying flu in birds and humans, and it generated and ranked promising hypotheses, some of which were unfamiliar and thought-provoking. Later, she fed in the full proposal, and Co-Scientist prioritized a protein she hadn't considered, connected to signaling pathways she was interested in. Back in the lab, she added unpublished data, and with each iteration, the hypotheses sharpened from candidate proteins to specific amino acids. Bryant's team is now building cell lines with those mutations to test the hypotheses. Normally, identifying precise amino acids would take two to three years of experimental work, but her lab expects to complete it in six months if Co-Scientist's targets are correct.