GPT-5 Pro helps immunologist solve 3-year-old T cell mystery

What happened

Immunologist Derya Unutmaz used GPT-5 Pro, released in late 2025, to analyze results from a 2022 experiment on T cells (immune cells that fight cancer and other illnesses). The model suggested that a glucose-like molecule called deoxyglucose interfered with the construction of a protein called IL-2, which can prevent T cells from becoming inflammatory-response cells known as Th17. This insight explained why T cells exposed to deoxyglucose overwhelmingly became Th17 cells, while those in a low-glucose environment did not.

Why it matters

Unutmaz, a professor at The Jackson Laboratory and the University of Connecticut, says AI has become so central to his work that he can't imagine doing science without it, comparing it to "taking both of your hands away, or half of your brain away." He also asked GPT-5 Pro to simulate an experiment on CD8+ T cells targeting lymphoma; the model correctly predicted their enhanced killing ability, even though the results weren't published. This suggests AI can function like a collaborator that streamlines literature reviews and helps narrow down which experiments are worth repeating, potentially cutting out "weeks to months, even years, of work."

What to watch

The test is whether AI-generated insights remain reliable when the user lacks the subject matter expertise to evaluate their significance. Unutmaz notes that someone without his expertise wouldn't have been able to tell if the mechanistic insight was important. He is now experimenting with advanced AI tools like Codex and GPT-5.2 Deep Research to compile large-scale cancer mutation datasets and generate research materials, including a T-cell-focused draft textbook.