McGill University researchers have developed an artificial intelligence tool that can identify small groups of cells most responsible for driving aggressive cancers.

The tool, called SIDISH, offers scientists a clearer path to designing targeted therapies by showing which cells inside a tumour are most strongly linked with poor patient outcomes, rather than treating all cancer cells as if they behave the same way.

In a preclinical study published in Nature Communications, SIDISH successfully identified “high risk” cells across pancreatic, breast and lung cancers using tumour samples collected from patients and analyzed in the lab.

How the tool works

SIDISH’s key innovation is that it connects what happens inside individual cells with patient outcomes, a long-standing challenge in cancer research.

“Single cell data is very detailed, but it usually comes from only a few patients and rarely includes how those patients actually fared. Patient data, on the other hand, often at the bulk level, includes survival information but averages signals from millions of cells, hiding the rare but dangerous ones that drive disease,” said first author Yasmin Jolasun, a PhD student in McGill’s Department of Medicine.
Existing computational tools have struggled to meaningfully bring these two types of data together.

“Our tool builds a bridge between both worlds. It can identify which cells are most strongly associated with faster disease progression and patient survival outcomes,” said Jolasun.

While SIDISH was tested first in cancers, the same approach could be applied to other complex diseases where cell to cell differences play a major role, she added.

SIDISH stands for semi-supervised iterative deep learning for identifying single-cell high-risk populations.

Predicting drug targets before lab testing

Beyond identifying the problem, SIDISH can also simulate how high-risk cells respond when specific genes are turned on or off, helping predict which genes might be promising drug targets.

“This could ease a major bottleneck in drug development, where finding the right targets often requires years of trial and error testing,” said senior author Jun Ding, assistant professor in McGill’s Department of Medicine and a junior scientist at the Research Institute of the McGill University Health Centre.

For example, he said, a patient’s tumour could be analyzed with single cell sequencing, SIDISH would identify the cells driving that tumour and simulate how they respond to different drugs, generating a short list of treatments most likely to be effective.

“In the short term, SIDISH could help us repurpose existing FDA approved drugs using public datasets. In the long term, it has the potential to fundamentally change how new drugs are discovered,” said Ding.

The work remains in development and is not yet used in clinical care. The research team is now applying SIDISH to additional diseases and collaborating with industry partners to further refine the approach