Research by Drs. Jun Ding, Gregory Fonseca, and David Eidelman has developed a powerful new artificial intelligence (AI) tool that can pinpoint the specific cells within a tumor that are most likely to drive aggressive disease.
Cancer is not made up of identical cells. Tumors are complex ecosystems where some cells are far more dangerous than others. Identifying these high-risk cells has been a major challenge, because existing technologies either capture detailed information from a small number of cells or broader data from many patients, but not both.
The new AI tool, called SIDISH, bridges this gap by combining these two types of data. It learns from large patient datasets while also zooming in on individual cells, allowing researchers to identify which cells are linked to poor outcomes and disease progression.
Using this approach, the team was able to detect small populations of high-risk cancer cells across multiple cancer types, link these cells to worse patient survival, identify key genes and pathways driving aggressive disease, and simulate potential treatments in silico (virtually) to find targets that could reduce these harmful cells
Importantly, the tool doesn’t just describe the disease, but it can help predict how a patient’s cancer might behave and suggest more personalized treatment strategies.
By revealing the hidden drivers of cancer at the cellular level, this research moves us closer to precision medicine, where therapies are tailored not just to the patient, but to the most dangerous cells within their tumor.
Read the Article
SIDISH integrates single-cell and bulk transcriptomics to identify high-risk cells and guide precision therapeutics through in silico perturbation. Jolasun Y, Song K, Zheng Y, Wang J, Fonseca GJ, Eidelman DH, Ding J.
Nat Commun. 2025 Dec 10;16(1):11271.
Read More
AI tool pinpoints cells driving aggressive cancers. New approach opens door to better-targeted treatments and faster drug discovery for complex diseases. McGill Newsroom. April 15, 2026.