Researchers at the University of Gothenburg have developed an AI model that's faster and better at detecting cancer than the current semi-manual method.
Glycans (structures of sugar molecules in cells) can be measured by mass spectrometry to detect cancer. However, the data must be carefully analyzed by one of the few experts in the world—taking up to days per sample.
Detecting cancer in seconds
Researchers at the University of Gothenburg have now developed an AI model named "Candycrunch" that does it in just a few seconds per test, as reported in the journal Nature Methods.
"The AI model was trained using a database of more than 500,000 examples of different fragmentations and associated structures of sugar molecules in 90 percent of cases," said Daniel Bojar, Associate Senior Lecturer in Bioinformatics at the University of Gothenburg in a statement.
Detects biomarkers missed by human analyses
That means the AI model could soon reach the same levels of accuracy as the sequencing of other biological sequences, such as DNA, RNA or proteins. In addition, the AI model is fast and accurate in its answers, so it can accelerate the discovery of glycan-based biomarkers for both diagnosis and prognosis of a cancer.
The Candycrunch model is also able to identify low concentrations of biomarkers, which are often missed by human analyses.
Citation: Urban, J., Jin, C., Thomsson, K. A., Karlsson, N. G., Ives, C. M., Fadda, E., & Bojar, D. (2024). Predicting glycan structure from tandem mass spectrometry via deep learning. Nature Methods, 1-10. https://doi.org/10.1038/s41592-024-02314-6 (open access)