Fig. 2: Core methodology for transforming glycoprofiles to glyco-motif profiles and visualizing cluster-representative substructures using GlyCompare.

a, b A glycoprofile with structure and relative abundance annotation G is obtained. The glycans are decomposed to a substructure set S, and the presence/absence of each substructure is recorded. Presence/absence vectors are weighted by the glycan abundance, and are summed into a substructure vector P. c Seven example glycoprofiles are transformed to substructure vectors as (a) and (b). d A substructure network is constructed to identify the non-redundant glyco-motifs that change in abundance from their precursor substructures. e The glycoprofiles can then be compared by their glyco-motif vectors M to generate more meaningful clusters. Both glycoprofiles and substructures can be clustered for similarity analysis. f Core structure information can be visualized from a substructure cluster. For example, four substructures with different weights were aligned together, and the monosaccharides with a weight over 51% were preserved. Throughout the manuscript, glycan is referred to complete and secreted monosaccharide polymer; a glycan substructure is referred to a complete or incomplete monosaccharide polymer observable within at least one secreted glycan; a glycan motif (glyco-motif) is referred to an enriched functional glycan substructure for a dataset or biological process. Note that both glycan epitopes (typically terminal glycan substructures recognized by lectins) and glycan cores (biosynthetic glycan substructures common to select types (e.g., N- or O-glycosylation) or modes (e.g., complex or high-mannose) of biosynthesis) are glyco-motifs as they are biologically functional, interpretable and will be enriched in datasets selecting for specific glycan presentation of biosynthesis. Glycompare core methods are explained at length in the “Methods” section.