Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 23.
Published before final editing as: Glycobiology. 2026 Aug 22;36(11):cwag068. doi: 10.1093/glycob/cwag068

Ten Years of Integrative Carbohydrate Structure Validation with Privateer

Lucy Schofield 1, Jordan Dialpuri 2, Haroldas Bagdonas 1, Lou Holland 1, Phuong Thao Pham 1, Shuai Wang 1, William Cook 1, Jon Agirre 1,*
PMCID: PMC7619613  EMSID: EMS218765  PMID: 42630059

Abstract

Understanding glycoprotein structures is critical for understanding their biological roles, yet despite the moderate size of the glycans, determining the structure of these atomic models presents unique challenges. Originally designed to help mine the Protein Data Bank for monosaccharides and their context, the Privateer carbohydrate validation software has evolved to become ubiquitous in carbohydrate model building workflows. Privateer detects issues affecting sugar stereochemistry, regiochemistry, pyranoside and furanoside ring conformation, as well as overall glycan conformation and composition. Supporting both X-ray crystallography and cryo-electron microscopy data, Privateer can communicate with refinement software to help restrain sugar rings to their minimal-energy conformations. In this work, we reflect on a decade of Privateer, documenting its core processes and best practices to serve as the primary citation for the software.

Keywords: cryo-EM, crystallography, glycans, glycoproteins, validation

Introduction

Carbohydrates are one of the most structurally diverse classes of macromolecules, owing to their extensive branching patterns and stereochemical variation (Schachter et al. 1982; Werz et al. 2007). They play a central role in many molecular mechanisms, including cellular recognition, immune evasion, and viral pathogenesis (Bangaru et al. 2020; Lenza et al. 2023). Consequently, determining the structure of glycoproteins has long been important for understanding the molecular basis of disease and developing therapeutics, as illustrated by the design of the SARS-CoV-2 mRNA vaccines (Bangaru et al. 2020; Seeberger et al. 2022; Lenza et al. 2023). Structural studies of glycoproteins demonstrate that glycans are rarely passive post-translational modifications; instead, they actively influence protein folding and alter binding interfaces through protein-carbohydrate interactions (Shental-Bechor and Levy 2008; Cummings et al. 2022). Capturing these glycans is therefore essential for observing macromolecular structures in their native, functional states. In the Protein Data Bank (PDB), carbohydrate-containing structures consistently represent ~ 10% of the total annual depositions, making them the most common post-translational modification in the repository (see Figure 1) (Berman et al. 2000; Agirre 2017; Dialpuri, Bagdonas, Schofield, Pham, Holland, and Agirre 2024; Schofield et al. 2024).

Figure 1. Distribution of carbohydrate-containing structure deposition in the Protein Data Bank over time.

Figure 1

Open in a new tab

Grey bars (right axis) show the total number of structures deposited annually. Lines track structure deposition counts for specific carbohydrate categories (left axis): N-glycans, O-glycans, C-glycans, S-glycans and ligands. The analysis indicates around 10% of yearly deposits are sugar-containing structures. The apparent decline in deposited structures for 2025 likely reflects structures that are still undergoing deposition review or on hold awaiting publication. Data were extracted from the Privateer Database (Dialpuri, Bagdonas, Schofield, Pham, Holland, and Agirre 2024).

Figure 1 alt text: Dual-axis chart showing the annual deposition of structures in the PDB over time, tracking total depositions alongside specific glycan categories.

Despite their biological importance and consistent deposition, the structural determination of sugars is often challenging. Their high flexibility, combined with microheterogeneity and complex branching networks, often results in difficult-to-interpret density in both X-ray crystallography and cryo-electron microscopy (cryo-EM) maps, particularly with low-resolution data (Agirre et al. 2017). Structural complexities, including anomeric stereochemistry, regiochemical branching, and ring conformations, make carbohydrate modelling uniquely difficult. As a result of this, the quality of these structures has historically lagged behind that of proteins and nucleic acids (Agirre, Davies, et al. 2015). This discrepancy was largely driven by a deficit in software that provided automatic pipelines to enforce correct sugar chemistry, as standard macromolecular building and refinement tools were optimised for proteins. Analysis of the PDB revealed widespread modelling errors, including structures containing high-energy ring-conformations, incorrect anomeric configurations, or biologically impossible linkages that violate fundamental stereochemical rules (Agirre, Davies, et al. 2015). In fact, a 2004 study identified nomenclature errors in approximately 30% of carbohydrate deposition (Lütteke et al. 2004). Further, a 2015 study found that 30% of deposited structures featured sugars modelled in high-energy conformations, which is surprising given that such distortion typically only occurs within enzyme active sites (Agirre, Davies, et al. 2015).

To rectify some of these issues, PDB-CARE was developed to automatically check the connectivity and naming of oligosaccharides, while the CARP server allowed users to compare torsional data against existing structures (Lütteke and von der Lieth 2004; Lütteke et al. 2005). While valuable for correcting nomenclature and linkages, the field still lacked a unified approach capable of assessing the complex conformations that sugar rings can adopt, as well as evaluating how well a sugar fits the experimentally-derived density. Assessing the fit is key, as density serves as a three-dimensional map of atomic positions derived from experimental measurements into which the atomic model must be built and validated. Ten years ago, to address these systematic issues in the PDB, the software package Privateer was developed (Agirre, Iglesias-Fernández, et al. 2015). It has evolved from a conformational analysis program into a multi-functional carbohydrate structure validation tool that is easily integrated into structural biology pipelines. Today, it is widely accessible to the scientific community, available in both CCP4i2 and CCP4Cloud of the CCP4 Software Suite, CCP-EM, as a plug-in for ChimeraX, via the Privateer web application (https://privateer.york.ac.uk), via Moorhen (https://moorhen.org/), as a stand-alone command line program, or a python API (Emsley et al. 2010; Agirre et al. 2023; Burnley et al. 2023; Meng et al. 2023; Dialpuri, Bagdonas, Schofield, Pham, Holland, Bond, et al. 2024; Holland et al. 2026). The main functionality of Privateer involves the detection and validation of carbohydrates by calculating Cremer-Pople puckering parameters (Cremer and Pople 1975), as well as validating a sugar’s fit-to-density by calculating real-space correlation coefficients (RSCC) in electron density maps, a standard metric used in crystallography to assess how well an atomic model fits the local density. Privateer can be used to validate N-, O-, C- and S-linked glycans across both linear and branched topologies. Complex polysaccharides such as glycosaminoglycans and glycolipids are also supported by Privateer. It can also be applied to protein-sugar complexes, including structures of antibodies, enzymes, lectins and receptors (Pinto et al. 2020; Akkermans et al. 2022; Morais et al. 2023; Trenker et al. 2024). Crucially, Privateer does not only flag errors, it also generates restraint dictionaries that can be fed directly back into refinement engines like REFMAC5, Servalcat, BUSTER and PHENIX.refine, enabling corrections of distorted rings during structure refinement (Murshudov et al. 2011; Afonine et al. 2012; Bricogne et al. 2017; Yamashita et al. 2021; Atanasova et al. 2022; Yamashita et al. 2023). An overview of Privateer’s main architecture and functionality can be seen in Figure 2.

Figure 2.

Figure 2

Open in a new tab

The architecture of Privateer and its links to other community resources. The core of Privateer is written in C++, which relies on a number of data sources, including the PDB Chemical Component Dictionary, CCP4 and CCP-EM Monomer Library, a PDB-derived database of linkage torsion angles, as well as offline mirrors of GlyConnect and GlyToucan (Westbrook et al. 2015; Tiemeyer et al. 2017; Alocci et al. 2019; Agirre et al. 2023). Restraints can be outputted for use in a number of refinement engines. Privateer is available via the CCP4 or CCP-EM Software Suites, as well as a plugin for ChimeraX (Agirre et al. 2023; Meng et al. 2023; Holland et al. 2026). It is also available in Moorhen, a Coot-based model building web application, as well as via the Privateer web app (Emsley et al. 2010; Debreczeni and Emsley 2012).

Figure 2 alt text: Flowchart showing the architecture of Privateer, its data sources, and outputs.

Going forward, as structural biology moves towards automation and machine-learning driven structure solution, the structural validation of carbohydrate structures deposited into the PDB is extremely important. As structures in the PDB are continually used to inform drug design and train machine learning algorithms, carbohydrate validation is essential to prevent misinterpretation of glycoprotein and protein-glycan complex structures. In this paper, we will reflect on a decade of Privateer, including its recent extensions, scientific impact, and ongoing role in advancing structural glycobiology.

Algorithm details

Cremer-Pople parameters flag high-energy conformations

One key aspect of carbohydrate chemistry is the principle that monosaccharides exist in a dynamic equilibrium between open-chain and cyclic structures, adopting six-membered pyranose or five-membered furanose rings. Unlike planar aromatic rings, carbohydrates exhibit ring puckering to minimise ring strain and steric clashes, by favouring staggered rather than eclipsed substituent orientations. Consequently, these rings have strong preferences for specific low-energy conformations. For pyranoses, these low-energy states are 4C1 chair conformation for d-pyranoses, and 1C4 for l-pyranoses. However, exceptions exist, notably sialic acid, a sugar with high biological importance due to its role in immune response and viral entry. It is an L-pyranose that adopts a 2C5 chair conformation due to its unique C2 anomeric numbering and extended carbon backbone. While genuine transitions into high-energy conformations are biologically relevant, they are rare and typically restricted to the transition states that occur in enzyme active sites. In the absence of this enzymatic stabilisation, the modelling of high-energy conformers such as envelopes (3E), boats (2.5B) and skew-boats (1S5) is likely a modelling or refinement error, with high-energy conformations fit to noisy or fragmented electron density (Agirre, Davies, et al. 2015).

To assess whether a modelled sugar ring is in the expected low-energy conformation, Privateer relies on the Cremer-Pople puckering coordinates (see Figure 3) (Cremer and Pople 1975). In 1975, Cremer and Pople introduced a general definition of ring puckering coordinates, providing a quantitative method for describing the puckering of non-planar ring systems. For both pyranose and furanose rings, the Cremer-Pople parameters can distinguish between different ring conformations, whilst quantifying the degree of distortion from idealised geometries.

Figure 3.

Figure 3

Open in a new tab

Conformational landscape of furanose and pyranose rings. A) Schematic representation of furanose ring puckering shapes, represented by a series of different envelopes and twists. B) The Cremer-Pople Sphere, representing the mathematical mapping of six-membered pyranose puckering coordinates (Q, θ, ϕ) onto a spherical surface. The low-energy 4C1 and 1C4 chairs sit at the North Pole (θ = 0˚) and South Pole (θ = 180˚) respectively. Along the equatorial plane (θ = 90˚), there is flexible pseudorotation between the boat (B) and skew-boats (S). Ring inversion from the North to the South pole requires navigating through high-energy half chair (H) or envelope (E) transition states. C) Structures indicating the planes and atom labels for furanose ring conformations. D) Structures indicating the planes and atom labels for pyranose ring conformations. Figure adapted from (Agirre 2017).

Figure 3 alt text: Four-panel diagram illustrating carbohydrate ring conformations, featuring furanose puckering shapes, the Cremer-Pople sphere mapping pyranose coordinates, and chemical structures with standard atom labeling for both ring types.

For pyranose rings, the puckering coordinates are mapped onto a three-dimensional spherical coordinate system defined by three parameters: total puckering amplitude, Q, which quantifies the magnitude of displacement from planarity, and the angular coordinates θ and ϕ, which specify the directional orientation of the puckering (seeFigure 3). On this conformational sphere, the most thermodynamically stable, lowest-energy conformations occupy the poles. The 4C1 and 1C4 chair conformations reside at the North (theta = 0˚) and South (θ = 180˚) poles, respectively, characterised by opposite ends of the ring puckering in opposing directions. Conversely, the equator (θ = 90˚) represents highly strained, high-energy intermediates, such as boat (B) and skew-boat (S) conformations, where the ring ends are displaced in the same direction. Because transitioning between the stable poles and the strained equator requires a significant energy input larger than 15 kcal/mol, pyranose rings favour the stable 4C1 chair (Biarnés et al. 2007). Consequently, sugars that are modelled as high-energy conformations are typically the result of modelling errors, unless stabilised by an external factor, such as the catalytic environment of an enzyme active site.

Comparatively, furanose rings occupy a smaller conformational space. Puckering parameters are calculated using a similar framework, where the pseudo-rotation of the ring is along a flat two-dimensional plane rather than a sphere (see Figure 3). Unlike pyranoses, the conformational landscape consists of ten distinct envelope (E) conformations (where four atoms are coplanar and one is out-of-plane) and ten distinct twist (T) conformations (where two adjacent atoms are displaced on opposite sides of the plane).

In order to calculate the ring conformation of a modelled sugar, Privateer is able to identify carbohydrate chains within a structure and references them against a manually curated internal database, which was created by analysing the idealised, low-energy conformers’ coordinates in the PDB’s Chemical Component Dictionary (CCD) (Westbrook et al. 2015). For every supported monosaccharide, a set of ideal, energy-minimised Cremer-Pople coordinates are stored for comparison. Privateer extracts the three-dimensional coordinates of the ring atoms and calculates their Cremer-Pople parameters. By comparing the calculated coordinates (Q, θ, ϕ), the ring conformation is detected. Following this, Privateer calculates the RSCC for each sugar to measure its fit to the electron density. A sugar with an RSCC below 0.7 is flagged as a potential density-fit outlier because it is poorly supported by the local density. This provides an important complement to conformational validation, as an unusual ring conformation may represent either a modelling error or a genuine distortion supported by the density. A key example of this is for iduronic acid, where the 1C4 diagnostic check does not fully reflect the complexity of the glycan, which exists in a dynamic equilibrium between multiple ring conformations. In this case, high-energy ring conformations may be flagged by Privateer, but the calculated RSCC score can be used to identify whether the conformation fits the density. An example of a high energy ring conformation modelled in the PDB can be seen in Figure 4, showing a mannose pyranoside as a high-energy skew-boat conformation. Privateer flags this as a conformational outlier, and calculates an RSCC of 0.41, indicating to the user that this high-energy conformation is not supported by the local density.

Figure 4.

Figure 4

Open in a new tab

Fixing and extending glycan structures with Privateer. Analysis of 3QVP (Kommoju et al. 2011), an otherwise excellent model, using Privateer indicates a high energy skew-boat conformation, poor fit to density (RSCC of 0.41) and lack of GlyConnect ID for the N-glycan where it is found. Upon removal of a mismodelled water molecule and refinement of the N-glycan chain in Coot, the previously high-energy conformation turned into a 4C1 chair. Further inspection of the new maps revealed additional monosaccharides downstream, allowing the glycan chain to be extended.

Figure 4 alt text: Two-panel figure showing a PDB-deposited glycan structure compared to the same structure once rebuilt according to Privateer’s validation output. The deposited structure shows a high-energy skew-boat conformation, which does not have a GlyConnect ID, whilst the re-built structure has all low-energy 4C1 chair conformation, and an extended glycan chain that has a GlyConnect ID.

Updated restraints prevent the modelling of distorted sugars

Whilst Privateer is able to flag these high-energy conformations after structure refinement, it is crucial that refinement programs use correct geometry to avoid the misrefinement of sugar rings. In macromolecular structure solution, refinement programs typically rely on restraint dictionaries that contain prior chemical knowledge, including information about bond lengths, angles, torsions, chiralities and planarities, to interpret the chemistry of compounds in macromolecular structures. These are typically stored in Crystallographic Information Files (CIF) files within libraries like the CCP4 Monomer Library (Agirre et al. 2023). However, these past carbohydrate restraint dictionaries were outdated, containing incorrect geometries (Atanasova et al. 2022). These dictionaries also frequently lacked explicit ring puckering restraints, resulting in refined structures which had skewed, high-energy ring conformations that fit noisy density rather than chemical reality (Atanasova et al. 2022).

To solve this systemic issue, new restraint dictionaries were regenerated for pyranose rings within the CCP4 suite, which were shown to help address conformational issues with a reduction in the gap between Rwork and Rfree (Atanasova et al. 2022). Privateer provided a list of the most frequently modelled sugar rings in the PDB, which were fed into the AceDRG restraint generation software in the form of CIF files. AceDRG sampled a vast ensemble of potential conformers (Long et al. 2017; Atanasova et al. 2022); the lowest-energy conformers were fed into RDKIT and REFMAC5 to optimise the coordinates, and output a revised CIF dictionary (Murshudov et al. 2011; RDKit: Open-source cheminformatics. https://www.rdkit.org). Target values for bond length, angles, and torsions were updated, and outputted as a CIF restraint dictionary stored in the CCP4 Monomer Library. These restraint dictionaries and their generated conformers were then passed to Privateer for geometric validation. A key implementation resulting from this was the introduction of unimodal torsion restraints, which reflected the ideal, lowest energy ring conformation and, due to their unique naming, could be restrained separately from the rest of the molecule (Atanasova et al. 2022). These explicit, tight constraints were shown to effectively prevent refinement programs from driving sugar rings into high-energy conformations, ensuring they remained in their most chemically feasible states (Atanasova et al. 2022). It is important to note that these torsion restraints should only be activated in specific cases, where the user is certain the refinement program is mistakenly skewing the sugar due to poor density. For most cases, refinement should be performed without these restraints, as skewed sugars will most often be a sign of a modelling error elsewhere in the polysaccharide or its environment.

In addition to sugar ring conformational preferences, glycosidic bonds also have distinct conformational preferences. Just as the Ramachandran plot validates peptide backbone torsions, Privateer provides a similar tool for glycan chains (including N-, O- and C- glycans) by calculating torsion linkage preferences based on a validated set of glycosylated structures from the PDB (Ramachandran et al. 1963; Dialpuri et al. 2023). Branched glycan chains (including N- and O-glycans) contain a variety of highly flexible glycosidic bonds defined by specific torsion angles: ϕ, and ψ which are variable, and ω which is typically 180˚ (Dialpuri et al. 2023). The torsion angle of the linkage between a sugar and an amino acid residue is also of interest. To establish a reliable reference baseline, the PDB was mined for well-fitting (RSCC > 0.80) sugar rings in a 4C1 chair conformation (Dialpuri et al. 2023). By filtering for only these validated structures, a database of validated glycosidic and asparagine-sugar linkages was compiled, resulting in relatively clean clusters of preferred torsion angles that serve as a benchmark for each link. For a sugar linkage, Privateer calculates a Z-score which allows users to compare their modelled linkage torsion to the validated observations.

Privateer ensures the modelling of biologically feasible glycans

Along with identifying issues regarding sugar chemistry, Privateer is also able to detect issues with their biology through the analysis of glycan composition. Due to both advancements in mass spectrometry and the elucidation of underlying biosynthetic pathways, glycomics databases contain an extensive repository of biologically feasible, naturally occurring glycans (Tiemeyer et al. 2017; Alocci et al. 2019; Bagdonas et al. 2020). Privateer leverages these data by extracting the atomic coordinates of a modelled glycan chain and translating its topology into the standardised WURCS (Web 3.0 Unique Representation of Carbohydrate Structures) notation (Matsubara et al. 2017; Bagdonas et al. 2020). By generating standard WURCS strings, Privateer significantly enhances the FAIRness (Findability, Accessibility, Interoperability and Reusability) of glycan structural data, bridging 3D macromolecular models with the broader glycoinformatics landscape. The WURCS code is cross-referenced against two key resources: GlyTouCan, an international glycan structure repository, and GlyConnect, a glycoproteomics databank (see Figure 2)(Tiemeyer et al. 2017; Alocci et al. 2019). This integration allows researchers to move beyond basic structural validation and enrich macromolecular models with contextual biological metadata, such as tissue specificity, cell types, and species of origin. GlyTouCan can also act as a bridge between PDB and GlyConnect; if a GlyConnect ID is not identified, it signals that this specific glycan topology has never been deposited – almost invariably revealing a modelling error (see Figure 4) (Tiemeyer et al. 2017; Alocci et al. 2019; Bagdonas et al. 2020). Furthermore, it was identified that some N-glycan structures showed incorrect α-linkages between N-acetyl-d-glucosamine and asparagine residues, resulting in sugars being built as 1C4 chairs, as opposed to the more favourable 4C1 chair for d-pyranoses (Agirre, Davies, et al. 2015). These unlikely links can be detected via cross-referencing against both GlyToucan and Glyconnect databases (Tiemeyer et al. 2017; Alocci et al. 2019; Bagdonas et al. 2020). Furthermore, establishing this interoperability provides a vital interface with functional glycomics workflows, such as connecting to GlyGen for glycan microarray binding profiles and GlyConnect for site-specific glycoproteomics (York et al. 2020). Cross-referencing against these functional and structural repositories ensures that modelled glycans are not only compliant with known biosynthetic pathways, but are also biologically meaningful and experimentally plausible (Bagdonas et al. 2020).

Another identified issue in carbohydrate-containing structures was the mismodelling of sugar anomers, leading to unfeasible or incorrectly modelled glycans (Lütteke 2009; Joosten and Lütteke 2017). Sugar anomers are a result of linear sugar cyclisation, upon which the carbonyl carbon (typically labelled ‘C1’) becomes a new chiral centre. This generates two distinct stereoisomers, also known as anomers: α and β. In the PDB, different anomers of the same sugar represent distinct CCD entries, such as beta-d-mannose (BMA) named alpha-d-mannose (MAN). This detail increased the likelihood of erroneously modelled anomers, where the α-anomer has been modelled where a β-anomer should be (Lütteke 2009; Joosten and Lütteke 2017). In order to address this issue, Privateer checks the detected sugar’s anomer against the three letter CCD code, to ensure the modelled anomer matches the CCD residue name. This validation allows Privateer to correct mislabelled or mismodelled sugar anomers, but also helps identify biologically feasible links, as selection of the wrong sugar anomer has resulted in erroneous links in some structures (Joosten and Lütteke 2017).

The exception to the rule: validating C-glycans

Whilst tryptophan C-mannosylation (Trp-Man) is a relatively rare type of glycosylation, analysis of the PDB revealed structures containing this modification are frequently mismodelled, or entirely omitted (Minakata et al. 2021; Holland et al. 2025). This is likely a result of C-glycans favourably adopting the more uncommon 1C4 inverse chair conformation, rather than the more common 4C1 chair conformation adopted by most pyranoses in N- and O-glycans (Holland et al. 2025). This conformational flip minimises steric clashes between the hydroxymethyl group of the mannose and the indole ring of the tryptophan residue (Holland et al. 2025). When the incorrect, default 4C1 chair conformation is initially modelled, subsequent refinement can distort the sugar ring into high-energy conformations to force a fit to the density, often arriving at conformations other than the expected 1C4 inverted chair at low resolution, due to weak gradients in the electron density (Holland et al. 2025). To resolve this, validation of Trp-Man linkages has been added to Privateer, where the sugar linkage is checked to be alpha, and the conformation is checked as 1C4. The dictionary for the Trp-Man connection in the CCP4 Monomer Library has also been updated to contain unimodal torsion restraints, which need to be activated in order to reliably flip the 4C1 starting conformation into the correct 1C4 inverted chair (Holland et al. 2025).

Implementation and impact

Privateer has the ability to assess the chemical and crystallographic validation of carbohydrates and displays these results in an accessible form to the user (see Figure 5). Privateer presents its results in a series of tables and diagrams, translating complex validation into clear visual flags (see Figure 5). The validation results are clearly depicted onto standardised Symbol Nomenclature For Glycans (SNFG) diagrams, where orange highlights indicate a validation issue, flagging to the user to inspect this sugar (Varki et al. 2015). This allows model builders to instantly pinpoint sugar outliers, whether it be due to chemical feasibility or poor density fit. Further, the Privateer web app provides access to the Privateer Database, which enables researchers to analyse carbohydrate structures by providing pre-calculated validation reports for any PDB entry (Dialpuri, Bagdonas, Schofield, Pham, Holland, and Agirre 2024; Dialpuri, Bagdonas, Schofield, Pham, Holland, Bond, et al. 2024). Users can also query the entire repository of carbohydrate-containing PDB structures based on glycosylation type or linkage type (Dialpuri, Bagdonas, Schofield, Pham, Holland, and Agirre 2024; Dialpuri, Bagdonas, Schofield, Pham, Holland, Bond, et al. 2024). This allows glycobiologists interested in protein structures to easily access all deposited carbohydrate-containing structures.

Figure 5.

Figure 5

Open in a new tab

Privateer validation outputs for PDB ID 3QVP available via the Privateer web application (https://privateer.york.ac.uk/). (Top left) Privateer validation of PDB ID 3QVP (Kommoju et al. 2011). The results include cross-references to GlyToucan and GlyConnect databases (Tiemeyer et al. 2017; Alocci et al. 2019). Conformational, anomeric, torsion, puckering and chirality issues are also reported. (Top right) Symbol nomenclature for glycan (SNFG) diagram is drawn for the modelled glycan chain. Hovering over each sugar gives detailed information about the residue number, conformation, mean B-factor, RSCC, detected sugar type, and any detected issues by Privateer. Orange highlights over sugars and linkages on the SNFG indicated a detected outlier by Privateer. (Bottom left) An interactive panel for Moorhen, a Coot-based tool commonly used for model building and molecular graphics, which displays the model and density if supplied. Clicking on different sugars in the SNFG will navigate to that sugar in the Moorhen window. (Bottom right) Linkage torsion plots based on a validated set of structures from the PDB. Calculated linkage torsions from the sugar under investigation are plotted on the graph in blue. This results page is available for all glycan chains or ligands in a structure.

Figure 5 alt text: Four-panel screenshot of the Privateer web application interface, showing glycan validation results, an interactive SNFG diagram, a Moorhen 3D molecular viewer, and linkage torsion plots.

Building on current integrations, several key developments will expand Privateer’s scope across model-building, cryo-EM, open-chain sugar support, and machine-learning predictions. Deeper integration into interactive model-building interfaces such as Moorhen will directly link residue-level outliers to local visual inspection and real-time correction, streamlining the iterative cycle between model building, refinement and validation within a unified interface. Furthermore, Privateer’s integration into highly accessible applications, like the Privateer web app and Moorhen, ensures that inspecting and validating glycoprotein structures is easier than ever, requiring nothing more than a web browser. Beyond classic pyranose and furanose rings, Privateer will extend its validation to open-chain monosaccharides, where the traditional Cremer-Pople ring puckering parameters are not applicable. Unlike X-ray crystallography, which requires rigid, highly ordered packing of macromolecules into crystal lattices, cryo-EM captures proteins in a near-native state (Atanasova et al. 2020). Because cryo-EM does not require this symmetry, it is a powerful approach for visualising highly flexible, heterogenous carbohydrate-containing structures. By embedding Privateer into widely used suites such as CCP4 and CCP-EM, carbohydrate validation is now a seamless part of the macromolecular structure solution pipeline for both X-ray and cryo-EM structures (Agirre et al. 2023; Burnley et al. 2023). This integration allows users to model, refine and validate sugars in a single workflow, using Privateer’s validation to guide their next step (see Figure 7). Future additions will incorporate Q-score analysis to assess density fit, analogous to RSCC calculation for X-ray crystallography structures (Pintilie et al. 2020; Burnley et al. 2023; Meng et al. 2023). Finally, a key role of Privateer going forward is for it to serve as a vital tool for evaluating AI-predicted models. Although recent frameworks such as AlphaFold3 and RoseTTAFold All-Atom extend prediction to glycans, all models trained on historic PDB data risk propagating pre-existing errors (Abramson et al. 2024; Krishna et al. 2024). Privateer provides an essential independent benchmark, ensuring that AI-generated glycan models adhere to fundamental carbohydrate chemistry.

Figure 7.

Figure 7

Open in a new tab

The role of Privateer as a validation tool for carbohydrate-containing structures. The generation of valid sugar-containing protein structures requires the iterative cycling of protein model building, where the atomic model is built into the density; refinement, in which the atomic coordinates are improved by balancing fit to electron density and geometric restraints; and validation using Privateer. Model building using Coot is the first step, followed by refinement using refinement engines such as Servalcat (Emsley et al. 2010; Debreczeni and Emsley 2012; Yamashita et al. 2021; Yamashita et al. 2023). Structure refinement can be performed using external restraints generated by Privateer, but these should be used in specific use-cases, not as default. Validation is performed by Privateer, highlighting issues with the sugar chemistry or fit-to-density. The user can continue the loop if Privateer has flagged any issues.

Figure 7 alt text: Circular flowchart showing the iterative cycle of carbohydrate model building with Coot, refinement with Servalcat, validation with Privateer, and final database deposition.

Historically, carbohydrate modelling has been prone to errors, including distorted ring conformations, mislabelled anomers, unexpected linkages, and poor density fit (Lütteke 2009; Agirre, Davies, et al. 2015; Agirre et al. 2017; Joosten and Lütteke 2017). Today, Privateer has streamlined this process, making it straightforward to generate chemically accurate, well-modelled, carbohydrate-containing structures. Analysis of incorrectly modelled sugars (as detected by Privateer) showed that the number of outliers has reduced from ~ 20% to less than 5% (see Figure 6). More specifically, the number of high energy conformations has reduced from ~ 30% in the 1990s to less than 10% today. Over the past decade, Privateer has evolved from a geometric validation tool into a key multi-functional validation tool for carbohydrates in structural biology. By tackling the underlying complex chemistry, combined with validation of density fit, the software has elevated the baseline quality and reliability of glycoprotein structures deposited in the PDB. Privateer’s contribution to the improvement of carbohydrate validation methods has allowed structural biologists who may not be experts in sugar chemistry or glycobiology to build complete, biochemically feasible carbohydrate structures. These structures further the understanding of glycobiology as a whole. Ultimately, by lowering the barrier to entry, Privateer ensures that high-quality carbohydrate models are no longer the exception in structural biology, but the standard.

Figure 6.

Figure 6

Open in a new tab

Annual trends of carbohydrate anomalies detected in the PDB. Grey bars (right-axis) represent the total number of sugar residues deposited per year. Lines track the relative percentage of annual sugar residues flagged as outliers by Privateer (left axis). Red lines denote any high-energy ring conformation by Privateer, while navy lines illustrate conformational outliers deviating from a chair conformation. The apparent decline in deposited structures for 2025 likely reflects structures that are still undergoing deposition review or on hold awaiting publication. Data was extracted from the Privateer Database (Dialpuri, Bagdonas, Schofield, Pham, Holland, and Agirre 2024).

Figure 6 alt text: Dual-axis graph showing historical trends of annual carbohydrate depositions alongside the relative percentages of ring conformation and chair-deviation anomalies over time.

Acknowledgments

Wendy Offen (University of York, UK) was the first Privateer user, and provided many practical ideas present in the software; her attention to detail and enthusiasm helped integrate the software into the model building, refinement and validation workflow. Saioa Urresti and Christian Roth (University of York at that time) further tested the software with many different polysaccharides. Kiyoko Aoki-Kinoshita (GlyToucan) and Frédérique Lisacek (GlyConnect) helped us mine their wonderful glycoinformatics resources, which we did, under the invaluable guidance of Daniel Ungar (University of York, UK). Carme Rovira (Universitat de Barcelona, Spain) and Gideon Davies (University of York, UK) provided much of the motivation behind validating ring conformations. Paul Emsley, Robbie Joosten, Fei Long, Rob Nicholls, Stuart McNicholas, Keitaro Yamashita, Kathryn Cowtan, Keith Wilson and, most importantly, Garib Murshudov – our CCP4 collaborators – helped refine our validation ideas for use in practical scenarios. Elena Seiradake (University of Oxford, UK) gave us the incredible opportunity to work on her structures, which prompted us to write the code to support tryptophan mannosylation. This paper is dedicated to the late Marcin Wojdyr (Global Phasing ltd, UK), who not only helped us integrate Privateer into CCP4 in the early days of the software, but also wrote the Gemmi library that Privateer increasingly relies on; though Marcin was taken from us far too early, he leaves a legacy as vast as the void his absence now creates.

Funding

Lucy Schofield is funded by STFC/CCP4 PhD studentship agreement 4462290 (York) / S2 2024 012 (STFC) awarded to Jon Agirre. Jordan Dialpuri is funded by the Biotechnology and Biological Sciences Research Council (BBSRC; grant No. BB/T0072221). Haroldas Bagdonas was funded by The Royal Society [grant number RGF/R1/181006]. Lou Holland was funded by The Royal Society (URF\R\221006). Phuong Thao Pham was a self-funded PhD student. Shuai Wang is funded by the Chemistry Wild Fund PhD Studentship of the University of York. William Cook is a Master’s student in the Department of Chemistry, University of York. Jon Agirre is a Royal Society University Research Fellow (awards UF160039 and URF\R\221006) and is partly funded by the BBSRC (grant no. BB/Y00888X/1).

Footnotes

Conflict of interest

The authors declare that no conflicts of interest exist.

Data Availability Statement

The Privateer source code can be found at https://github.com/glycojones/privateer.

References

  1. Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, Ronneberger O, Willmore L, Ballard AJ, Bambrick J, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature. 2024;630(8016):493–500. doi: 10.1038/s41586-024-07487-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Afonine PV, Grosse-Kunstleve RW, Echols N, Headd JJ, Moriarty NW, Mustyakimov M, Terwilliger TC, Urzhumtsev A, Zwart PH, Adams PD. Towards automated crystallographic structure refinement with phenix.refine. Acta Crystallogr D Biol Crystallogr. 2012;68(Pt 4):352–367. doi: 10.1107/S0907444912001308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Agirre J. Strategies for carbohydrate model building, refinement and validation. Acta Crystallogr D Struct Biol. 2017;73(Pt 2):171–186. doi: 10.1107/S2059798316016910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Agirre J, Atanasova M, Bagdonas H, Ballard CB, Baslé A, Beilsten-Edmands J, Borges RJ, Brown DG, Burgos-Mármol JJ, Berrisford JM, et al. The CCP4 suite: integrative software for macromolecular crystallography. Acta Crystallogr D Struct Biol. 2023;79(6):449–461. doi: 10.1107/S2059798323003595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Agirre J, Davies G, Wilson K, Cowtan K. Carbohydrate anomalies in the PDB. Nat Chem Biol. 2015;11(5):303. doi: 10.1038/nchembio.1798. [DOI] [PubMed] [Google Scholar]
  6. Agirre J, Davies GJ, Wilson KS, Cowtan KD. Carbohydrate structure: the rocky road to automation. Curr Opin Struct Biol. 2017;44:39–47. doi: 10.1016/j.sbi.2016.11.011. [DOI] [PubMed] [Google Scholar]
  7. Agirre J, Iglesias-Fernández J, Rovira C, Davies GJ, Wilson KS, Cowtan KD. Privateer: software for the conformational validation of carbohydrate structures. Nat Struct Mol Biol. 2015;22(11):833–834. doi: 10.1038/nsmb.3115. [DOI] [PubMed] [Google Scholar]
  8. Akkermans O, Delloye-Bourgeois C, Peregrina C, Carrasquero-Ordaz M, Kokolaki M, Berbeira-Santana M, Chavent M, Reynaud F, Raj R, Agirre J, et al. GPC3-Unc5 receptor complex structure and role in cell migration. Cell. 2022;185(21):3931–3949.:e26. doi: 10.1016/j.cell.2022.09.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Alocci D, Mariethoz J, Gastaldello A, Gasteiger E, Karlsson NG, Kolarich D, Packer NH, Lisacek F. GlyConnect: Glycoproteomics goes visual, interactive, and analytical. J Proteome Res. 2019;18(2):664–677. doi: 10.1021/acs.jproteome.8b00766. [DOI] [PubMed] [Google Scholar]
  10. Atanasova M, Bagdonas H, Agirre J. Structural glycobiology in the age of electron cryo-microscopy. Curr Opin Struct Biol. 2020;62:70–78. doi: 10.1016/j.sbi.2019.12.003. [DOI] [PubMed] [Google Scholar]
  11. Atanasova M, Nicholls RA, Joosten RP, Agirre J. Updated restraint dictionaries for carbohydrates in the pyranose form. Acta Crystallogr D Struct Biol. 2022;78(Pt 4):455–465. doi: 10.1107/S2059798322001103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Bagdonas H, Ungar D, Agirre J. Leveraging glycomics data in glycoprotein 3D structure validation with Privateer. Beilstein J Org Chem. 2020;16:2523–2533. doi: 10.3762/bjoc.16.204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Bangaru S, Ozorowski G, Turner HL, Antanasijevic A, Huang D, Wang X, Torres JL, Diedrich JK, Tian J-H, Portnoff AD, et al. Structural analysis of full-length SARS-CoV-2 spike protein from an advanced vaccine candidate. Science. 2020;370(6520):1089–1094. doi: 10.1126/science.abe1502. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne PE. The Protein Data Bank. Nucleic Acids Res. 2000;28(1):235–242. doi: 10.1093/nar/28.1.235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Biarnés X, Ardèvol A, Planas A, Rovira C, Laio A, Parrinello M. The conformational free energy landscape of beta-D-glucopyranose. Implications for substrate preactivation in beta-glucoside hydrolases. J Am Chem Soc. 2007;129(35):10686–10693. doi: 10.1021/ja068411o. [DOI] [PubMed] [Google Scholar]
  16. Bricogne G, Blanc E, Brandl M, Flensburg C, Keller P, Paciorek W, Roversi P, Sharff A, Smart OS, Vonrhein C, et al. BUSTER. 2017.
  17. Burnley T, Iadanza M, Joseph A, Palmer C, Winn M. CCP-EM 2.0: software tools for efficient, accurate and reproducible management and automation of cryo-EM processing from images to structures. Acta Crystallogr A Found Adv. 2023;79(a2):C17. [Google Scholar]
  18. Cremer D, Pople JA. General definition of ring puckering coordinates. J Am Chem Soc. 1975;97(6):1354–1358. [Google Scholar]
  19. Cummings RD, Schnaar RL, Esko JD, Woods RJ, Drickamer K, Taylor ME. Principles of glycan recognition. Essentials of Glycobiology Cold Spring Harbor Laboratory Press, Cold Spring Harbor (NY) 2022 [Google Scholar]
  20. Debreczeni JÉ, Emsley P. Handling ligands with Coot. Acta Crystallogr D Biol Crystallogr. 2012;68(4):425–430. doi: 10.1107/S0907444912000200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Dialpuri JS, Bagdonas H, Atanasova M, Schofield LC, Hekkelman ML, Joosten RP, Agirre J. Analysis and validation of overall N-glycan conformation in Privateer. Acta Crystallogr D Struct Biol. 2023;79(Pt 6):462–472. doi: 10.1107/S2059798323003510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Dialpuri JS, Bagdonas H, Schofield LC, Pham PT, Holland L, Agirre J. Monitoring carbohydrate 3D structure quality with the Privateer database. Beilstein J Org Chem. 2024;20(1):931–939. doi: 10.3762/bjoc.20.83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Dialpuri JS, Bagdonas H, Schofield LC, Pham PT, Holland L, Bond PS, Sánchez Rodríguez F, McNicholas SJ, Agirre J. Online carbohydrate 3D structure validation with the Privateer web app. Acta Crystallogr F Struct Biol Commun. 2024;80(Pt 2):30–35. doi: 10.1107/S2053230X24000359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of coot. Acta Crystallogr D Biol Crystallogr. 2010;66(Pt 4):486–501. doi: 10.1107/S0907444910007493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Holland L, Dialpuri JS, Schofield LC, Pham PT, Wang S, Cook W, Croll TI, Agirre J. Privateer for UCSF ChimeraX : Carbohydrate structure validation and visualization. Protein Sci. 2026;35(8):e70716. doi: 10.1002/pro.70716. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Holland L, Pham PT, Bagdonas H, Dialpuri JS, Schofield LC, Agirre J. Methods for detecting, building, and improving tryptophan mannosylation in glycoprotein structures. Protein Sci. 2025;34(2):e70025. doi: 10.1002/pro.70025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Joosten RP, Lütteke T. Carbohydrate 3D structure validation. Curr Opin Struct Biol. 2017;44:9–17. doi: 10.1016/j.sbi.2016.10.010. [DOI] [PubMed] [Google Scholar]
  28. Kommoju P-R, Chen Z-W, Bruckner RC, Mathews FS, Jorns MS. Probing oxygen activation sites in two flavoprotein oxidases using chloride as an oxygen surrogate. Biochemistry. 2011;50(24):5521–5534. doi: 10.1021/bi200388g. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Krishna R, Wang J, Ahern W, Sturmfels P, Venkatesh P, Kalvet I, Lee GR, Morey-Burrows FS, Anishchenko I, Humphreys IR, et al. Generalized biomolecular modeling and design with RoseTTAFold All-Atom. Science. 2024;384(6693):eadl2528. doi: 10.1126/science.adl2528. [DOI] [PubMed] [Google Scholar]
  30. Lenza MP, Egia-Mendikute L, Antoñana-Vildosola A, Soares CO, Coelho H, Corzana F, Bosch A, Manisha P, Quintana JI, Oyenarte I, et al. Structural insights into Siglec-15 reveal glycosylation dependency for its interaction with T cells through integrin CD11b. Nat Commun. 2023;14(1):3496. doi: 10.1038/s41467-023-39119-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Long F, Nicholls RA, Emsley P, Graǽulis S, Merkys A, Vaitkus A, Murshudov GN. AceDRG: a stereochemical description generator for ligands. Acta Crystallogr D Struct Biol. 2017;73(Pt 2):112–122. doi: 10.1107/S2059798317000067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lütteke T. Analysis and validation of carbohydrate three-dimensional structures. Acta Crystallogr D Biol Crystallogr. 2009;65(Pt 2):156–168. doi: 10.1107/S0907444909001905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lütteke T, Frank M, von der Lieth C-W. Data mining the protein data bank: automatic detection and assignment of carbohydrate structures. Carbohydr Res. 2004;339(5):1015–1020. doi: 10.1016/j.carres.2003.09.038. [DOI] [PubMed] [Google Scholar]
  34. Lütteke T, Frank M, von der Lieth C-W. Carbohydrate Structure Suite (CSS): analysis of carbohydrate 3D structures derived from the PDB. Nucleic Acids Res. 2005;33(Database issue):D242–6. doi: 10.1093/nar/gki013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lütteke T, von der Lieth C-W. pdb-care (PDB carbohydrate residue check): a program to support annotation of complex carbohydrate structures in PDB files. BMC Bioinformatics. 2004;5(1):69. doi: 10.1186/1471-2105-5-69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Matsubara M, Aoki-Kinoshita KF, Aoki NP, Yamada I, Narimatsu H. WURCS 2.0 update to encapsulate ambiguous Carbohydrate Structures. J Chem Inf Model. 2017;57(4):632–637. doi: 10.1021/acs.jcim.6b00650. [DOI] [PubMed] [Google Scholar]
  37. Meng EC, Goddard TD, Pettersen EF, Couch GS, Pearson ZJ, Morris JH, Ferrin TE. UCSF ChimeraX: Tools for structure building and analysis. Protein Sci. 2023;32(11):e4792. doi: 10.1002/pro.4792. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Minakata S, Manabe S, Inai Y, Ikezaki M, Nishitsuji K, Ito Y, Ihara Y. Protein C-mannosylation and C-mannosyl tryptophan in chemical biology and medicine. Molecules. 2021;26(17):5258. doi: 10.3390/molecules26175258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Morais MAB, Nin-Hill A, Rovira C. Glycosidase mechanisms: Sugar conformations and reactivity in endo- and exo-acting enzymes. Curr Opin Chem Biol. 2023;74(102282) doi: 10.1016/j.cbpa.2023.102282. 102282. [DOI] [PubMed] [Google Scholar]
  40. Murshudov GN, Skubák P, Lebedev AA, Pannu NS, Steiner RA, Nicholls RA, Winn MD, Long F, Vagin AA. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr D Biol Crystallogr. 2011;67(Pt 4):355–367. doi: 10.1107/S0907444911001314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Pintilie G, Zhang K, Su Z, Li S, Schmid MF, Chiu W. Measurement of atom resolvability in cryo-EM maps with Q-scores. Nat Methods. 2020;17(3):328–334. doi: 10.1038/s41592-020-0731-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Pinto D, Park Y-J, Beltramello M, Walls AC, Tortorici MA, Bianchi S, Jaconi S, Culap K, Zatta F, De Marco A, et al. Cross-neutralization of SARS-CoV-2 by a human monoclonal SARS-CoV antibody. Nature. 2020;583(7815):290–295. doi: 10.1038/s41586-020-2349-y. [DOI] [PubMed] [Google Scholar]
  43. Ramachandran GN, Ramakrishnan C, Sasisekharan V. Stereochemistry of polypeptide chain configurations. J Mol Biol. 1963;7(1):95–99. doi: 10.1016/s0022-2836(63)80023-6. [DOI] [PubMed] [Google Scholar]
  44. RDKit: Open-source cheminformatics. https://www.rdkit.org .
  45. Schachter H, Narasimhan S, Gleeson P, Vella GJ, Brockhausen I. Oligosaccharide branching of glycoproteins: biosynthetic mechanisms and possible biological functions. Philos Trans R Soc Lond B Biol Sci. 1982;300(1099):145–159. doi: 10.1098/rstb.1982.0162. [DOI] [PubMed] [Google Scholar]
  46. Schofield LC, Dialpuri JS, Murshudov GN, Agirre J. Post-translational modifications in the Protein Data Bank. Acta Crystallogr D Struct Biol. 2024;80(Pt 9):647–660. doi: 10.1107/S2059798324007794. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Seeberger PH, Freedberg DI, Cummings RD. Essentials of Glycobiology. Cold Spring Harbor Laboratory Press; Cold Spring Harbor (NY): 2022. Glycans in biotechnology and the pharmaceutical industry. [Google Scholar]
  48. Shental-Bechor D, Levy Y. Effect of glycosylation on protein folding: a close look at thermodynamic stabilization. Proc Natl Acad Sci U S A. 2008;105(24):8256–8261. doi: 10.1073/pnas.0801340105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Tiemeyer M, Aoki K, Paulson J, Cummings RD, York WS, Karlsson NG, Lisacek F, Packer NH, Campbell MP, Aoki NP, et al. GlyTouCan: an accessible glycan structure repository. Glycobiology. 2017;27(10):915–919. doi: 10.1093/glycob/cwx066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Trenker R, Diwanji D, Bingham T, Verba KA, Jura N. Structural dynamics of the active HER4 and HER2/HER4 complexes is finely tuned by different growth factors and glycosylation. Elife. 2024;12(RP92873) doi: 10.7554/eLife.92873. RP92873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Varki A, Cummings RD, Aebi M, Packer NH, Seeberger PH, Esko JD, Stanley P, Hart G, Darvill A, Kinoshita T, et al. Symbol nomenclature for graphical representations of glycans. Glycobiology. 2015;25(12):1323–1324. doi: 10.1093/glycob/cwv091. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Werz DB, Ranzinger R, Herget S, Adibekian A, von der Lieth C-W, Seeberger PH. Exploring the structural diversity of mammalian carbohydrates (“glycospace”) by statistical databank analysis. ACS Chem Biol. 2007;2(10):685–691. doi: 10.1021/cb700178s. [DOI] [PubMed] [Google Scholar]
  53. Westbrook JD, Shao C, Feng Z, Zhuravleva M, Velankar S, Young J. The chemical component dictionary: complete descriptions of constituent molecules in experimentally determined 3D macromolecules in the Protein Data Bank. Bioinformatics. 2015;31(8):1274–1278. doi: 10.1093/bioinformatics/btu789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Yamashita K, Palmer CM, Burnley T, Murshudov GN. Cryo-EM single-particle structure refinement and map calculation using Servalcat. Acta Crystallogr D Struct Biol. 2021;77(Pt 10):1282–1291. doi: 10.1107/S2059798321009475. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Yamashita K, Wojdyr M, Long F, Nicholls RA, Murshudov GN. GEMMI and servalcat restrain REFMAC5. Acta Crystallogr D Struct Biol. 2023;79(Pt 5):368–373. doi: 10.1107/S2059798323002413. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. York WS, Mazumder R, Ranzinger R, Edwards N, Kahsay R, Aoki-Kinoshita KF, Campbell MP, Cummings RD, Feizi T, Martin M, et al. GlyGen: Computational and informatics resources for glycoscience. Glycobiology. 2020;30(2):72–73. doi: 10.1093/glycob/cwz080. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The Privateer source code can be found at https://github.com/glycojones/privateer.

RESOURCES