OpenAI partnered with PerturbAI and Scripps Research to run thousands of agents on biological data from millions of mouse cells. The system used GPT-Rosalind to process a whole-brain CRISPR atlas that maps how specific gene changes affect different cell types. The full analysis completed in thirty-two minutes, whereas human researchers would require several months or a year to achieve the same result.
Each agent examined a single gene perturbation, assessed cellular responses, and searched scientific literature for context. The team identified unexpected differences between disease-related genes and a potential link between cholesterol regulation and neurodegeneration. In blind comparisons, human experts preferred the distributed agents’ reports over those from a single agent seventy-five percent of the time. The resulting Reasoning Atlas is available on GitHub, and the accompanying paper is published on bioRxiv.
* The agents searched the scientific literature for explanations during the process
* Human experts preferred the distributed reports over single agent outputs
* The findings aim to help researchers prioritise lab tests
