Researchers say a closer look inside cells could be used by physicians to detect diseases earlier and better match patients to therapies.
Researchers at McGill University have developed an artificial intelligence tool that can detect disease markers within individual cells that conventional analysis methods miss.
The tool, called DOLPHIN, analyzes how genes are spliced together from smaller components called exons, providing a more detailed view of cellular states than traditional gene-level analysis. The research was published in Nature Communications.
“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,” says Dr Jun Ding, assistant professor in McGill’s Department of Medicine and a junior scientist at the Research Institute of the McGill University Health Centre, in a release.
Current gene-level analysis methods combine disease markers into single counts per gene, which can mask critical variations in RNA expression that indicate disease presence, severity, or treatment response potential.
Moving Beyond Gene-Level Analysis
DOLPHIN examines gene splicing patterns at the exon level rather than treating genes as single units.
“Genes are not just one block; they’re like Lego sets made of many smaller pieces,” says Kailu Song, a PhD student in McGill’s Quantitative Life Sciences program and first author of the study, in a release. “By looking at how those pieces are connected, our tool reveals important disease markers that have long been overlooked.”
When researchers tested DOLPHIN on single-cell data from pancreatic cancer patients, the tool identified more than 800 disease markers that conventional analysis methods missed. The tool successfully distinguished between patients with high-risk, aggressive cancers and those with less severe cases.
Applications for Drug Development
The technology generates detailed single-cell profiles that could enable virtual simulations of cellular behavior and drug responses before laboratory or clinical testing begins.
The researchers plan to expand DOLPHIN’s capabilities from analyzing small datasets to processing millions of cells, which could support the development of more accurate virtual cell models.
The research was supported by the Meakins-Christie Chair in Respiratory Research, the Canadian Institutes of Health Research, the Natural Sciences and Engineering Research Council of Canada, and the Fonds de recherche du Québec.
ID 88421052 © Mkarco | Dreamstime.com
