Artificial intelligence may soon detect illnesses before any symptoms show up. Scientists at McGill University have developed a new AI tool that can spot subtle signs of disease hidden deep inside single cells.
The tool, called DOLPHIN, could change how doctors diagnose and treat a wide range of conditions. Here, we’ll look at what the researchers built, how it works, and how it could impact early disease detection and personalized medicine. The project was led by Jun Ding, assistant professor in McGill’s Department of Medicine and a scientist at the Research Institute of the McGill University Health Centre.
How the AI tool works for disease detection
The McGill team created DOLPHIN to see what’s really happening inside cells, something most current genetic tests can’t fully do. Instead of looking at genes as single units, DOLPHIN analyzes how genes are pieced together from smaller segments called exons. This detailed view helps detect molecular changes that could signal early signs of disease.
“Genes are not just one block, they’re like Lego sets made of many smaller pieces”, said first author Kailu Song. “By looking at how those pieces are connected, our tool reveals important disease markers that have long been overlooked”.
The researchers tested the tool on data from pancreatic cancer patients and found more than 800 genetic markers that other methods had missed. It could even tell which patients had aggressive, high-risk tumors and which had milder cases.
According to Ding, “this tool has the potential to help doctors match patients with the therapies most likely to work for them, reducing trial-and-error in treatment”. The system relies on AI to analyze massive amounts of single-cell data, identifying small shifts in RNA expression that point to disease before symptoms ever appear.
The study, published in Nature Communications, highlights how machine learning is reshaping cell biology. Instead of waiting for diseases to become visible through scans or symptoms, doctors could one day use AI to detect problems at the cellular level, long before they progress.
What AI technology could mean for medicine
Beyond spotting hidden disease markers, DOLPHIN moves researchers closer to building “virtual cells”—digital models that simulate how human cells behave, mutate, or respond to drugs. By generating more complete single-cell profiles, scientists could test potential treatments in a computer before moving to lab or clinical trials.
This kind of modeling could cut research costs and speed up drug development. It could also help identify which patients are likely to benefit from specific therapies, making personalized medicine more accurate and efficient.
The next step, Ding’s team says, is scaling DOLPHIN to handle millions of cells across multiple diseases. That expansion could make AI-powered cell modeling a routine part of biomedical research and clinical diagnostics.
The implications are huge. Detecting diseases before they surface could give doctors a critical head start—improving outcomes, reducing invasive testing, and personalizing care in ways that weren’t possible before. Tools like DOLPHIN hint at a future where your doctor might not just treat disease early but catch it before it ever starts.