GNNGLY: Graph Neural Networks for Glycan Classification

• Published in: IEEE access Vol. 11; p. 1
• Main Authors: Alkuhlani, Alhasan, Gad, Walaa, Roushdy, Mohamed, Salem, Abdel-Badeeh M.
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Artificial neural networks; Biological activity; Biological system modeling; Classification; Complexity; Convolutional neural networks; Deep learning; Fingerprint recognition; Glycan; glycoinformatics; graph convolutional networks; Graph neural networks; Graph theory; Graphical representations; Immune system; Machine learning; Neural networks; Proteins
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2023.3280123

Glycans are important biological molecules that can be found on their own or attached to other molecules. They have complex, branching structures that do not follow the linear structure. Glycans are crucial for many biological processes and they are involved in the development of several important diseases. Due to the complexity and the branched structure of glycans, most of the current studies have mainly focused on the other attached molecules instead of glycans themselves. This paper proposes, GNNGLY, a graph neural networks model for glycans classification. Firstly, Glycans are represented as molecular graphs, where atoms are represented as nodes and bonds are represented as edges. Graph convolutional networks (GCNs) are then used to make predictions on eight taxonomic classification levels and for the level of immunogenicity property. The performance results indicate that GNNGLY outperforms traditional machine learning methods and when compared to other existing tools for glycan classification, GNNGLY showed considerable performance results. GNNGLY could have a significant impact on the field of glycoinformatics and related research areas.