Innovative AI Tool Uncovers Hidden Disease Indicators Through Cellular Genetic Analysis

Researchers at McGill University have developed an advanced artificial intelligence technology capable of detecting subtle and previously invisible markers of disease within individual cells. Published in Nature Communications, this innovative tool, named DOLPHIN, explores the intricate details of gene expression by analyzing how genes are spliced from smaller components called exons—a process that offers a far more detailed view than traditional gene-level analysis.

Conventional methods tend to simplify genetic data by counting the total expression of each gene, which often masks critical variations that could signal the onset or progression of diseases. In contrast, DOLPHIN zooms into the exon level, revealing the complex architecture of genes and uncovering key biomarkers that are commonly overlooked.

In practical applications, DOLPHIN analyzed single-cell data from pancreatic cancer patients and identified over 800 previously undetectable disease markers. It successfully distinguished patients with aggressive, high-risk cancers from those with less severe forms, offering vital insights to guide personalized treatment strategies.

This breakthrough not only enhances early detection capabilities but also plays a crucial role in the effort to develop digital models of human cells. Such models—virtual representations of cell behavior—could revolutionize drug testing and treatment planning, reducing the need for extensive laboratory experiments and clinical trials.

Moving forward, the research team plans to expand DOLPHIN’s capabilities to analyze millions of cells, paving the way for more accurate virtual cell models and advancing precision medicine. This technological leap signifies a major step toward understanding cellular complexity and tailoring healthcare to each individual's genetic makeup.

For more detailed information, visit the original study in Nature Communications, authored by Kailu Song et al., titled "DOLPHIN advances single-cell transcriptomics beyond gene level" (2025).