Software for estimating disease-based changes in glycans developed by Nagoya University

Many proteins carry branched glycans, which help regulate protein stability, cell-cell recognition, receptor activity, and other biological functions. Multiple types of glycans can be attached at the same site on the same protein. However, there has been no standardized method to measure how this structural diversity changes in response to disease or environmental factors.

A research group including Designated Assistant Professor Bingyuan Zhang, Assistant Professor Koichi Himori, and Associate Professor Yusuke Matsui at the Institute for Glyco-core Research, Nagoya University, developed GlycoTraitR, a tool that quantifies and compares the structural diversity of glycans using mass spectrometry data. The tool is expected to help researchers and companies without specialized expertise in glycans organize complex datasets and identify potential biomarker and drug-target candidates. The study was published online in Bioinformatics Advances.

Recent advances in mass spectrometry and glycoproteomics search software have enabled large-scale identification of glycan attachment sites and glycan structures. However, the resulting datasets present several challenges. The vast number of glycan-peptide combinations dramatically increases the number of features to be analyzed, while some glycans may not be detected in every measurement, leading to datasets with many missing values. In addition, structurally similar glycans often share common substructures or compositions, meaning that treating each glycan as an independent feature can disperse related changes across numerous variables and obscure broader biological trends.

GlycoTraitR imports results generated by glycoproteomics search software, including pGlyco3 and Glyco-Decipher, and decomposes each glycan into measurable characteristics such as overall size, monosaccharide composition, branching, glycan class, and the presence of core fucosylation.

To account for differences in measurement intensity, GlycoTraitR converts glycan abundances into relative values within each glycosylation site and protein. It then defines structural diversity as the abundance-weighted variance of individual glycan features. Furthermore, using permutation tests that repeatedly shuffle labels assigned to case and control groups, the software evaluates differences between groups without assuming a specific statistical distribution. Users can also incorporate glycan motifs of particular interest into the analysis.

The researchers reanalyzed publicly available data (ProteomeXchange: PXD032219) from postmortem frontal cortex samples obtained from cognitively normal individuals, individuals with pathological changes but no clinical symptoms, and patients with symptomatic Alzheimer's disease. Applying GlycoTraitR to results generated by two glycoproteomics search programs, the researchers examined both average glycan characteristics and structural diversity at the protein level and at individual glycosylation sites. The analysis revealed a decrease in high-mannose glycans and an increase in complex-type glycans in the neural cell adhesion molecule NCAM2, suggesting a shift toward more mature glycan structures.

The software also detected reductions in glycan-size variability as well as changes in branching patterns at specific glycosylation sites. These findings demonstrate that measures of heterogeneity can reveal biologically relevant information that may be missed when considering average abundance alone.

By summarizing complex glycan datasets into interpretable features, GlycoTraitR enables comparisons between disease states, treatment groups, and other biological conditions. The framework may also help researchers prioritize candidate molecules and generate hypotheses for biomarker discovery and drug-target identification. Because both the software and analysis workflow are publicly available, the approach promotes reproducibility and comparability across studies while lowering barriers to entry for glycomics research.

Going forward, the researchers plan to support standardized glycan-structure formats, intensity-based quantification methods, and data-independent acquisition (DIA) datasets, which typically contain fewer missing values. They also aim to extend the framework to enable comparisons across studies and disease areas. In addition, they plan to incorporate functional annotations and knowledge of glycan biosynthesis to make it easier to identify and interpret candidates linked to disease mechanisms and drug discovery research.

Journal Information
Publication: Bioinformatics Advances
Title: GlycoTraitR: an R package for characterizing structural heterogeneity in N-linked glycoproteomics data
DOI: 10.1093/bioadv/vbag244