A deep learning model developed by Daniel Bojar and his team is now able to analyze exact glycan structures with high precision and in just a few seconds per sample. The results are now published in a scientific article in Nature Methods. Glycans can among other things indicate cancer in the cells.
Scientific article in Nature Methods: Predicting glycan structure from tandem mass spectrometry via deep learning
"With this method, we could soon reach the same levels of accuracy as the sequencing of other biological sequences, such as DNA, RNA or proteins," Daniel said.
Daniel is Associate Senior Lecturer in Bioinformatics at the Department of Chemistry and Molecular Biology, and at the Wallenberg Centre for Molecular and Translational Medicine.
The glycan structures can be measured by mass spectrometry. Glycans can among other things indicate cancer in the cells. Analyzing the data would manually take hours or even days of work per sample.
Translational benefit
With the help of AI and deep learning models, the analysis can now be done at lightning speed without requiring expert knowledge in the methods. This has for Daniel been an important goal.
"Making system biology and glycomics more accessible to life scientist in general will create benefits for the clinical and translational research. With easier access to these techniques, researchers who do not have specialized knowledge on glycomics can now include glycans in their research. This is important because glycans often play a significant role in nearly everything that happens in the human body "
