The analysis and interpretation of glycan profiling data is both an art and science. The investigator should be able to correlate the glycan structures observed with their biological role. Other difficulties include the complexity of carbohydrates, interpreting multiple streams of analytical data and correctly placing glycans into their biological context so that erroneous interpretations are avoided. Glycan profiling data can be used, with the proper interpretation, to map out disease pathways, enhance the quality of therapeutic proteins, and discover clinically relevant biomarkers. Glycan profiling data is represented as chromatograms and mass spectral peaks that can be converted into biological knowledge.
Complexity of structure, biosynthetic heterogeneity and non-template driven assembly of glycan modifications create difficulties in interpretation of glycan profiling data. While difficulty exists in interpreting data derived from glycan modifications, one must still interpret this data due to glycosylation's relevance to protein function, disease pathology and drug efficacy. Structural glycocomponent data without biological context has limited usefulness. Biological interpretation can allow differentiation between important biological changes and experimental noise and relate structure to function.
Fig. 1 Major sources of complexity in the structural analysis of glycans.1,5
Interpretation of glycan structural data is made challenging because glycans frequently exist as isomers with the same elemental composition but different linkages, anomeric configurations, and branching patterns. These issues can lead to large numbers of microvariants that make peak identification and quantitation difficult since many glycan structures may share the same mass but have different biological functions. Another challenge for data interpretation is occupancy information at specific sites, which may further complicate associating glycan structures to protein species. For example, correlation with other analytical factors such as retention time indexes, fragmentation patterns, and biosynthetic processing can assist in fully interpreting glycan structures.
Table 1 Sources of Complexity in Glycan Data Interpretation
| Complexity Source | Analytical Manifestation | Interpretative Challenge |
| Linkage isomerism | Identical mass, different connectivity | Ambiguous structural assignment |
| Anomeric diversity | α versus β configurations | Biological activity uncertainty |
| Site occupancy variation | Multiple attachment positions | Protein-specific function correlation |
Interpretation of glycans based on a single property or characteristic (degree of polarization, retention time, etc.) cannot give enough structural information to allow for both structural confidence and biological insight. Mass spectra will not distinguish isobaric species, nor linkages without some form of separation or fragmentation information being introduced. Retention time will allow comparison between glycans but will not confirm structure without orthogonal information. Binding assays such as lectins can quickly screen glycans for the presence of specific structures but lack detail to differentiate structures further or determine how many are present. Comparison between properties must be made. Evidence from the mass spectra should confirm any assignments made based on retention time. Functional information such as affinity measurements should accompany structural assignments whenever possible.
Limitations of interpretation involve glycan profiling data often being over-interpreted. When interpreting glycomic data it should be kept in mind that statistical analysis can find differences in glycan abundance between samples that may not be biologically significant, or clinically relevant. A difference could simply be due to normal physiological differences. Also, glycans identified using analytical glycan profiling may not represent glycans as they exist on proteins in vivo. For instance, during in vitro sample preparation, glycan structures may be changed or separated from the protein. Ensure that glycan analysis data is reflective of glycan function by pairing glycan analysis with functional assays where possible, and correlating with biological pathway information and clinical metadata.
Outputs are typically chromatograms, mass spectra or binding information. Glycan profiling data output interpretation involves correlating between chromatographic retention data, mass spectrometry data and binding affinity data outputs. Each glycan profiling technique produces unique outputs that contain structural and quantitative information about glycans. Users need to understand the outputs for each technique in order to properly interpret glycan profiling results.
A chromatogram provides a graphical representation of glycan abundance over time with peaks showing retention based on factors such as molecular size, charge, and polarity. Indexing glycans by the number of glucose residues can allow relative retention times to be compared between instruments and labs for reliable glycan assignment when compared to reference glycans run on the chromatographic system. Integration of these peaks, when detected with appropriate sensitivity, can be used to semi-quantitatively determine the relative amount of glycans present in a sample. Structural assignments should be confirmed with MS, as different structures can coelute due to having the same nominal mass.
Mass spectra represent molecules by their mass-to-charge ratio allowing the determination of elemental composition of monosaccharide components of glycans. Exact mass can differentiate glycans that vary by small masses such as substitution of a hexose for a hexosamine. Isotopic distribution gives additional elemental composition data. Relative abundance can give semi quantitative information although ionization efficiency differs for different glycans and must be calibrated. Fragmentation patterns in tandem MS allows further structural characterization however chromatographic separation (e.g. GC-MS or LC-MS) is typically necessary to fully assign structure.
The output from lectin arrays is readouts of binding strength to panels of immobilized lectins, which create patterns of recognition or glycan fingerprints. The fluorescent or chemiluminescent signal corresponds to relative amounts of the terminal epitopes targeted by each lectin probe (example: sialic acid, fucose, etc.). It should be noted that lectin specificities vary greatly between lectins and many lectins will bind a class of glycans instead of one unique structure. Pattern analysis software can help predict glycan structure but additional orthogonal testing is needed for structure assignment.
Relative quantitation methods express the amounts of glycans as fractions or percentages of the total sample. Absolute quantitation involves determining the actual molar concentration of glycans in the sample. Currently, most glycomic analyses report relative abundances. Quantitative glycomics approaches typically employ stable isotope-labeled internal standards, which allow for absolute quantitation of glycans present in the sample. Relative abundance measurements are easier to perform and require no authentic standards. Authentic standards for every analyte of interest are typically not available for targeted analyses, so relative quantitation methods are useful for determining changes in glycoform abundance from healthy to diseased samples. Absolute quantitation provides more rigorous data and is necessary when investigating bioavailability or pharmacokinetics, glycobiologging, comparing biosimilars, or when determining glycan concentrations in a clinical setting.
Structural elucidation of glycans from glycan profiling data involves the cognitive association of the analytical responses with carbohydrate structures based on chromatographic, mass spectral, and affinity data interpretation. Assigning structures to observed signals requires combining evidence from accurate masses, fragmentation spectra, and biological considerations to determine compositions and derive structural attributes.
Glycan compositions from mass spectral information can be determined by calculating all possible combinations of monosaccharides that match the mass spectrum's observed mass-to-charge ratio (m/z). On high-resolution mass spectrometers, isobaric compositions can also be differentiated by their accurate masses and elemental formulas (number of Hex, NHex, Fuc, and Neu5Ac residues) can be assigned. Neutral losses of specific masses during fragmentation of glycans can also help determine the composition of the glycan. It is impossible to determine which monosaccharides belong to a particular composition class using only mass information. Insight into the organism's biosynthetic capabilities as well as glycosylation rules can be used to rule out some compositions.
Branching, fucosylation sites, and sialic acid linkages can also be determined by MS/MS fragmentation spectra and chromatographic retention data. Branching of glycans will cause specific fragmentations based on how a glycan branches. For example, the cleavage of an antennae can determine the number of antennae on a glycan. Distinction between core fucosylation vs. antennary fucosylation can be made based on fragmentation patterns after glycosidic bonds are cleaved. Spectra cannot always distinguish α-2,3 linkages from α-2,6 linkages of sialic acids. If possible, orthogonal proof with sialidase treatments or ion mobility to separate these isomers is helpful. Otherwise, additional proof can be found by examining the fragmentation patterns of ions suspected to be containing a given linkage or by performing chemical treatments that will confirm the presence or absence of these features.
Fig. 2 Sialylation and fucosylation sites of human STs and FUTs.2,5
Separation of isomeric glycans with the same composition (same amount of monosaccharide units), but different linkages or anomeric configurations, remains a problem. This separation can be achieved by orthogonal methods, such as chromatography (separation by retention time), ion mobility spectrometry (separation in the gas phase), and tandem mass spectrometry (fragmentation creates linkagespecific fragment ions). If two structures are still isobaric after all available dimensions have been utilized to their fullest, then the ion may be reported as only having a compositional level of assignment or a derivatization method can be used that alters the mass of the glycan creating a difference in the m/z.
Defining confidence levels for assigned glycan structures involves consideration of mass accuracy and whether the fragments match those anticipated based on the mass spec platform being used and the knowledge of glycan biosynthesis. Full confidence would likely require accurate mass, extensive fragment matches, and retention time matching to a standard. Partial confidence may be acceptable in some cases, such as for initial discovery where it may be sufficient to know whether a particular glycan is present or absent. In some software platforms there may be a score indicating the likelihood that the assigned structure is correct based on how well the spectrum matches. Users should always review the evidence that supports structure assignments. Users should report their level of confidence ranging from: compositional (connectivity and branching unknown), structural (connectivity and branching known), to structural with defined stereochemistry.
Combining orthogonal data for glycan analysis is one method of helping to make confident structural assignments. Different types of measurements can be used to confirm assignments in multiple independent ways, because each technique can be limited by itself (separation but no identification, MS can tell composition but not distinguish isomers, fragmentation gives some structural information, etc.). Confidence in glycan structure is obtained by correlating all supporting evidence.
Integration of chromatographic retention characteristics and mass spectrometric analysis enables superior glycan identification than that possible using each approach independently. Glycans differing in structure that influences chromatographic properties such as hydrophilicity and charge will elute in a pattern that can be described by their retention times. Ascertaining the accurate mass of a glycan provides a limited set of possible compositions, which when combined with chromatographic data can be used to predict structure. Confirmation of structural features is then accomplished using tandem mass spectrometry to obtain sequence information.
Deglycosylation using exoglycosidases can provide unequivocal confirmation. Enzymatic removal of terminal sugars by exoglycosidases can lead to expected changes in mass and retention time that can confirm presence of structural elements. Exoglycosidases are specific for the cleavage of sugars from the non-reducing end of glycans based on linkage and anomeric preferences. The stepwise digestion of glycans by exoglycosidases will cause predictable decreases in mass and changes in retention time that confirm the structure interpreted from MS. This can be used to confirm any ambiguity over terminal epitopes or branching.
Reference databases give rise to analytical standards that are used to confirm assignments by comparing experimental measurements against known values. Glycan libraries provide retention indexes, reference masses, and fragmentation patterns used as references for assignments. Databases publicly store structure, biosynthetic pathway and analytical metadata which are used to support computer-assisted assignment programs and visual evaluation and allow for consistent reporting between labs and comparison between studies which is necessary for biomarker validation and clinical translation.
Absolute quantitation of glycan profiling data requires conversion of instrument response into units that reflect biological abundance (e.g. molecule percent) and incorporates expected variability for carbohydrates. Quantitative data must differentiate between real change and background noise to accurately reflect alterations in glycoform abundances. When applied correctly, quantitation allows statistical comparison between sample populations and has applications in biomarker discovery, as well as quality control, where decisions on drug activity or disease progression may rely on accurate quantitation of specific glycans.
Percent values indicate changes in relative abundance compared to other glycans within the same sample. A glycoform may increase or decrease in relative abundance on cells due to up- or down-regulation of biosynthetic enzymes or altered metabolism within the cell. On the other hand, percent values can be misleading. For example, if the overall level of glycosylation is reduced, then the percent relative abundance of the remaining glycan structures will increase even though there is no actual biochemical significance to the increase. The opposite is also true and percent values can decrease when the overall glycosylation increase. Therefore, it is important to interpret relative abundance changes with glycan structures that are altered and what is known about their biological function.
One important issue in quantitative glycan measurements is separating out sources of technical variation versus biological variation. Technical variation refers to artifacts in measurement caused by sample preparation, matrix effects, instrumental drift, etc. that masks any true differences between samples. Biological variation refers to variations between samples that are of interest (e.g. differences between populations). Designing experiments with technical replicates and controls allows us to account for both of these sources of variation so that our statistical tests are only identifying meaningful differences that are not due to technical artifacts, and that have potential biological significance.
Batch effects are technical variations that occur due to differences in sample processing or analysis on different days. Batch effects can confound comparisons by introducing artificial time trends. Normalization by total area or probabilistic quotient corrects for technical variation in signal intensity between samples. When disease changes the total amount of glycans present or when one glycoform makes up the majority of structures present, normalization may be skewed. Batch correction factors account for technical variation without removing biological differences, which allows for comparisons between samples processed on different days, such as between centers or over time. Use of standardized operating procedures across centers and inclusion of quality control samples that are run on every batch helps with batch correction.
Some general issues include counting equal signal response for structurally different glycans, ignoring isobaric species causing interference that leads to inaccurate quantitation of individual glycans and mistaking statistical significance for biological significance without proper functional validation. Differences in ionization or fluorescence response between glycoforms result in biased quantitation relative to each other if not corrected by appropriate normalization. Signal overlap of different species within a chromatographic peak or mass spectral feature will yield an apparent increase or decrease that is simply an artifact of quantitation. Neglecting to apply proper multiple hypothesis testing correction when analyzing multiple glycans can inflate false positives.
One of the ultimate goals of glycomics is linking glycan structural changes with activity. Glycan structural changes can be linked to activity in order to provide mechanistic insight into drug activity and disease mechanism. Linking structure to activity involves associating defined carbohydrates with responses such as immune cell activation, serum half-life, and signal transduction.
Antibody Fc N-glycans can have important effects on antibody effector function. For example, the degree of fucosylation can alter affinity for activating Fc receptors expressed on natural killer cells. Decreased fucosylation can increase antibody-dependent cellular cytotoxicity (ADCC). Galactosylation at the termini can also regulate binding affinity to C1q and thus impact complement-dependent cytotoxicity. Because of this, therapeutic antibodies can be engineered to lack these antibody effector functions or have stronger effector function if desired.
Table 2 Influence of Fc Glycan Structures on Effector Functions
| Structural Modification | Functional Consequence | Therapeutic Application |
| Core fucosylation reduction | Enhanced Fc receptor binding | Improved tumor cell killing |
| Increased galactosylation | Augmented complement activation | Enhanced inflammatory response |
| Terminal sialylation | Modulated inflammatory balance | Anti-inflammatory activity |
Terminal sialic acids present on glycoproteins are important for their survival in circulation because they hide terminal galactose residues on glycoproteins from binding to asialoglycoprotein receptors in the liver. Preventing binding to these receptors allows for therapeutic proteins to remain in circulation for longer periods of time and increases their half-lives. Glycoproteins that lack terminal sialic acids expose terminal galactose or N-acetylglucosamine residues that are recognized by hepatic asialoglycoprotein receptors. Binding to these receptors promotes removal of the therapeutic proteins from the bloodstream through endocytosis and trafficking to lysosomes. The number of terminal sialic acids present on each branch affects how well glycoproteins are hidden from these receptors. Because of this many therapeutic glycoproteins have modified sialylation to help increase their half-lives.
Terminal sialic acids attached to glycoproteins play an important role in determining their circulation time. Since terminal galactose sugars are recognized by the liver-specific asialoglycoprotein receptor and removed from circulation, terminal sialic acids serve to prevent the recognition of galactose by acting as a protective cap. Removal of sialic acids can be used as a targeting strategy to shorten serum half-life of therapeutic proteins. Therefore, lack of terminal sialic acids results in a shorter serum half-life of a therapeutic protein. Alteration of glycosylation patterns to expose terminal galactose or N-acetylglucosamine residues leads to receptor-mediated uptake into hepatocytes and targeting to lysosomes. The number of sialic acids per glycan branch can also affect protein half-life. Therefore modification of glycoproteins to change sialylation patterns is a common method for altering pharmacokinetics.
Contextualization of glycan analysis data mainly refers to separating statistically significant differences derived from the manufacturing process and clinically relevant differences. This implies setting acceptance ranges of glycan characteristics, determining the comparability of glycans from biosimilar candidates or different batches of the same product and justifying compliance with regulatory standards. The data obtained from glycan profiling needs to be interpreted for the manufacturing process to turn assay results into actionable quality control decisions. This means making sure the therapeutic holds its efficacy and safety promise without unnecessarily limiting the manufacturing process parameters.
Determining what constitutes a meaningful difference in product quality would allow one to define scientifically sound acceptance criteria for distinguishing between allowable batch-to-batch variation in the manufacturing process and changes which trigger investigation or action. Correlation of glycan structure with biofunctional outcomes such as receptor binding, effector function or clearance can be used to distinguish differences that do and do not affect clinical performance. Statistical process control tools can determine when glycans change outside of an established limit and then root cause analysis would be performed to determine if process changes are needed. This determination must take into account both technical aspects of assay sensitivity as well as the biological consequences of variations.
Glycan profiles are used for comparability studies to show consistency between different batches produced under altered manufacturing processes, or biosimilars to confirm biosimilarity to a reference product. Glycoform distributions, site occupancies, and terminal sugar variations must be compared directly (head-to-head) using multiple orthogonal techniques, especially when determining if observed differences are scientifically and clinically insignificant. Data are typically evaluated to understand if differences in glycosylation cause functional differences using assays to correlate analytical differences with clinical findings. Therefore, information is combined to support that a therapeutic product is "highly similar" to the reference molecule despite small glycan differences.
Regulatory agencies expect glycan results to be analysed following validated procedures, with traceable and statistically sound methods which are fit for purpose when submitted as part of a regulatory filing. Agencies recommend characterisation of glycosylation as a quality attribute of the therapeutic protein. Agencies expect glycan interpretation to include how specifications are derived based on clinical and manufacturing considerations. Guidance on data integrity expect fully traceable records of how glycan analyses were performed, including instrument calibrations and investigation of any out-of-specification results. Reporting of limitations and uncertainty should be clear to enable appropriate evaluation by regulators.
Glycan profiling data are subject to several caveats that must be kept in mind when attempting to interpret the results and turn them into useful biological insight. These limitations arise from many sources, including basic characteristics of glycans themselves, biological variability, and mismatches between analytical capabilities and biological relevance. Recognizing and accounting for these caveats can help prevent overinterpretation of glycan information.
Often times structural assignments will remain unknown because isomeric glycans exist with the same mass but different linkage positions, anomeric configurations, or branch orders. Spectral evidence by MS alone cannot confirm these small structural differences without chromatographic resolution or fragmentation evidence. Multi-dimensional analysis sometimes cannot determine linkage positions or anomeric configurations. Structures will then have to be annotated as compositional only.
Glycans functions are often context-specific. Changes in glycan structures do not have one defined interpretation. For example, the same glycan change can have different effects based on the protein on which it is found, where it is expressed, or under what conditions. Glycan changes that are disease-associated could also be naturally occurring changes under different contexts. Associations with metadata and functional studies are often needed to determine the biology behind a glycan alteration.
There is currently a considerable disconnect between analytical resolution afforded by glycan profiling platforms and biological understanding used to mechanistically interpret structural results. Mass spec and chromatography platforms can easily resolve slight differences in mass and/or chromatographic movement, however it may be unknown or only vaguely understood what the biological significance is of small structural modifications like position of linkage changes.
Robust interpretation of glycan measurements require best practice experimental designs, including appropriate experimental design, validation using orthogonal methods, and collaboration between analytical chemists and biologists. Such practices allow for glycan differences to be attributed to biological differences with high confidence allowing downstream application of analytical data to mechanisms and therapeutic targets.
Well-planned experiments should include reference standards, blanks, and quality control samples to ensure accurate glycan analysis. Blanks, standards, and quality controls should be randomly interspersed with samples to eliminate batch effects. Additionally, it may be useful to include a pooled sample to monitor run-to-run variations. Negative controls should be included to ensure there is no contamination with glycan release and derivatization reagents. In addition, experimental design should account for sample collection/storage variation so that any observed variation is biological variation.
Orthogonal validation approaches also provide multiple independent methods to validate glycan assignments. Using separate analytical technologies based on different physicochemical properties to validate glycan assignments can decrease technology bias leading to false assignments. Chromatography, mass spectrometry, and lectin capture technologies provide independent sources of structural information that can be used to validate glycan assignments. Mass spectrometry can suggest possible glycan assignments which can be confirmed by chromatography retention time and independent lectin validation techniques.
Sharing information about data interpretation with a colleague skilled in analytical chemistry who can offer insight into the technical aspects of what can and cannot be determined from the analytical methods, as well as with a biologist familiar with glycan biosynthesis, related diseases, or biology of the cells from which the glycans were derived allows one to avoid over-interpreting the glycan data. Having both of these perspectives in discussions about glycan structure determination will enable you to make valid interpretations about your glycan sample that are grounded in technical reality and biological significance.
Table 3 Collaborative Framework for Glycan Data Interpretation
| Expertise Domain | Contribution to Interpretation | Integration Outcome |
| Analytical chemistry | Method selection and validation | Technically robust measurements |
| Biological sciences | Functional context and relevance | Biologically meaningful conclusions |
| Bioinformatics | Data integration and pattern recognition | Comprehensive insight generation |
Data interpretation remains a complex task in glycan profiling and involves integrating data collected by chromatography, mass spectrometry, or affinity techniques in order to understand carbohydrate structure. Orthogonal confirmation, statistical analysis to define noise vs. biological change, and knowledge of biological functions will be critical for interpretation of glycan profiling data. Although instrumentation to help improve separation of glycans is constantly improving, there is still a lack of understanding about how well glycans can be correlated to specific biological effects based on linkage location, branching order, or terminal groups. Effector functions, half-life, and receptor binding are all effected by glycan structure but are often difficult to predict. Following best practice guidelines with regards to proper study design, method confirmation with multiple orthogonal assays, and collaborating between scientists that perform the assays and those that understand the biology will help eliminate confusion during data interpretation. Attention to these details during data analysis will allow glycan profiling to reach its potential in the biopharmaceutical industry and result in data that can be understood from a mechanism-based perspective that will aid in product improvement and biomarker identification.
Interpreting glycan profiling data requires more than assigning peaks or listing glycan compositions. Translating chromatographic profiles and mass spectrometry results into meaningful biological or biopharmaceutical conclusions demands structured analysis, orthogonal validation, and contextual understanding. Our glycan profiling and data interpretation services are designed to bridge the gap between raw analytical output and actionable insight—supporting research studies, drug development programs, and regulatory submissions.
Reliable interpretation begins with structural confidence. We integrate multiple analytical dimensions to strengthen glycan assignments and reduce ambiguity:
Beyond structural annotation, we evaluate quantitative trends and their potential biological implications. This includes assessing:
By combining structural analysis with functional context, we help ensure that glycan profiling data supports sound scientific conclusions rather than overinterpretation.
Data interpretation in glycomics can be complicated by isomeric ambiguity, ionization bias, and variability across batches or platforms. Our expert support includes:
For biopharmaceutical applications, we assist in interpreting glycosylation as a critical quality attribute (CQA), including comparability assessments and batch trend analysis. For disease research, we support hypothesis-driven interpretation linking glycan changes to biological pathways or clinical phenotypes. Our goal is to provide interpretation that is scientifically rigorous, transparent in its assumptions, and aligned with the intended application.
If you need support interpreting glycan profiling data, validating structural assignments, or contextualizing glycosylation changes in research or biopharmaceutical settings, our team offers customized analytical and reporting solutions. Contact us to discuss your dataset, study objectives, and the level of structural and quantitative insight required.
References
