An AI model has been developed to swiftly uncover cancer clues through sugar analyses, surpassing the current semi-manual method in speed and accuracy.
Glycans, the structures of sugar molecules within our cells, are detectable via mass spectrometry and can indicate various types of cancer.
However, interpreting the data from mass spectrometry manually to determine glycan structures is a time-consuming process requiring expertise and may take hours to days per sample.
Streamlining the Analysis Process
To address this bottleneck, researchers at the University of Gothenburg created an AI model named Candycrunch, which can swiftly analyze glycan structures. Through training on a vast database, the AI model can accurately deduce sugar structures in 90% of cases within just a few seconds per test.
Uncovering New Biomarkers
The AI model’s speed and precision rival that of sequencing biological sequences like DNA, RNA, or proteins. This efficiency can aid in the discovery of glycan-based biomarkers for cancer diagnosis and prognosis.
Candycrunch is adept at identifying structures often overlooked by human analysis, potentially leading to the discovery of new glycan-based biomarkers for research and clinical applications.
