ProfileOTMM: Automated Glycan Motif Analysis Through Pattern Recognition of Tree Structures

In recent times, tree-structured data has been rapidly spreading with the advent of the World Wide Web and Bioinformatics. It is employed as a common format for the storage and depiction of data. Glycans, essential substances in all organisms, are biomolecules that play diverse functional roles in the body. They are commonly modeled as tree-structured data and analyzed using machine learning techniques. Nevertheless, more detailed information is imperative to comprehensively understand the biological attributes and functionalities of gly-cans. In this research, we aim to propose a new model called profileOTMM (profile Ordered Tree Markov Model) for the complete automation of glycan analysis. As an initial step, feature engineering was conducted using the OTMM model, which serves as the foundational model for profileOTMM. By improving the accuracy of glycan data input through OTMM, we integrated monosaccharides, the elemental constituents of glycans, along with their linkage types into tree-structured glycan data. This enhanced approach enabled the extraction of biologically meaningful structures with a high degree of precision and accuracy, surpassing the results of previous research efforts. This research will encourage the understanding of glycan structures and their functions.