# Protein Data Bank (PDB) in Proteomics: How to Use Structural Information to Validate and Interpret Your Protein Identifications Proteomics experiments routinely return lists of identified proteins, but a protein identifier alone does not tell you whether the identification is biologically meaningful or whether the protein has a known three-dimensional structure that can explain its function. The Protein Data Bank (PDB) provides experimentally determined structures for many proteins, and the AlphaFold Database extends structural coverage to over 214 million predicted protein sequences. This article explains how to query these structural resources, interpret domains and active sites, and integrate structural evidence into your proteomics workflow for validation and functional interpretation. ## Scope and Reader Context This guide is written for biology students, researchers, laboratory professionals, and life-science practitioners who perform mass spectrometry-based proteomics and need to make sense of their protein identification lists. You will learn how to determine whether a protein you identified has a known structure, how to retrieve that structure, what structural features to examine, and how to use structural information to support or question your identifications. The practical outcome is a workflow you can apply to any proteomics dataset, from a single protein of interest to a full quantitative comparison across conditions. The focus is on the Protein Data Bank as the primary repository of experimentally determined structures, with attention to the AlphaFold Database as a complementary resource for proteins without experimental structures. You will also learn how structural information connects to functional annotation resources such as NCBI and EMBL-EBI, and how to use structural features to interpret the biological meaning of your proteomics results. ## At a Glance The table below summarizes the key structural resources, their content, and how they fit into a proteomics validation workflow. | Resource | Content Type | Coverage | Primary Use in Proteomics | | --- | --- | --- | --- | | Protein Data Bank (PDB) | Experimentally determined structures from X-ray crystallography, NMR, and cryo-electron microscopy | Proteins with solved experimental structures | Validate identifications against high-confidence structural data, examine active sites and binding interfaces | | AlphaFold Database | Predicted structures from the AlphaFold2 AI system | Over 214 million protein sequences, nearly complete UniProt coverage | Obtain structural models for proteins lacking experimental structures, generate hypotheses for functional interpretation | | NCBI Structure and Related Resources | Integrated sequence and structure search systems, cross-links to PDB | Proteins with sequence records and associated structures | Identify which of your identified proteins have structural entries, retrieve accession mappings | | EMBL-EBI Training and Data Resources | Training materials and integrated databases including PDB and AlphaFold DB | Curated learning pathways and cross-referenced databases | Learn query syntax, understand structure quality metrics, access integrated annotations | ## Understanding the Protein Data Bank and Its Role in Proteomics The Protein Data Bank is the central repository for experimentally determined three-dimensional structures of biological macromolecules. Structures deposited in the PDB come from techniques including X-ray crystallography, nuclear magnetic resonance spectroscopy, and cryo-electron microscopy. Each entry has a unique four-character PDB code that allows you to retrieve the atomic coordinates, experimental details, and associated functional annotations. For proteomics researchers, the PDB serves several distinct purposes. First, it provides a way to check whether a protein you identified by mass spectrometry has a high-confidence experimental structure. If a structure exists, you can examine whether the peptides you identified map to structured regions, active sites, or interaction interfaces. Second, the PDB allows you to compare your identified protein against structural homologs to infer function when direct annotation is sparse. Third, structural information can help you distinguish between true identifications and artifacts by checking whether your detected peptides are consistent with the known three-dimensional organization of the protein. The National Center for Biotechnology Information (NCBI) provides integrated search systems that connect sequence data to structural records. You can use [NCBI resources](https://www.ncbi.nlm.nih.gov/) to search for a protein by accession or name, retrieve its sequence record, and follow links to any associated PDB structures. This integration is valuable because your proteomics results will typically contain UniProt accessions or gene symbols, and you need a reliable path from those identifiers to structural data. ## The AlphaFold Database as a Structural Complement Experimental structure determination is time-consuming and has historically covered only a fraction of known proteins. The AlphaFold Database addresses this gap by providing predicted structures for over 214 million protein sequences, expanding from an initial release of approximately 300,000 structures in 2021. These predictions are generated by the AlphaFold2 artificial intelligence system and have been integrated into primary data resources including PDB, UniProt, Ensembl, InterPro, and MobiDB, as described in the [AlphaFold Database 2024 update](https://pubmed.ncbi.nlm.nih.gov/37933859). For proteomics researchers, the AlphaFold Database is particularly useful when you identify a protein that has no experimental structure in the PDB. You can retrieve a predicted structure, examine the confidence metrics, and use the model to generate hypotheses about domain organization, active site location, and potential binding interfaces. The database provides access through direct file downloads via FTP, queries using Google Cloud Public Datasets, and programmatic access endpoints. The integration of AlphaFold predictions into UniProt and other primary resources means that your standard proteomics annotation pipeline may already include structural information. When you annotate your protein list with UniProt entries, you may see links to AlphaFold models alongside any experimental PDB structures. Understanding how to interpret these predicted structures, including their confidence scores and limitations, is essential for using them responsibly in your analyses. ## Core Principles for Using Structural Information in Proteomics ### Structural Coverage Determines Your Validation Options Before you can use structural information to validate a protein identification, you need to know whether a structure exists for that protein. The first step in any structural validation workflow is therefore a coverage check. Query the PDB for your protein of interest using its UniProt accession or gene name. If an experimental structure exists, you can proceed with high-confidence structural analysis. If not, check the AlphaFold Database for a predicted model. The distinction between experimental and predicted structures matters for the strength of conclusions you can draw. Experimental structures from the PDB are derived from actual measurements and carry detailed information about resolution, refinement statistics, and experimental conditions. Predicted structures from AlphaFold are computational models with associated confidence metrics that indicate how reliable the prediction is likely to be. Both types of structure can inform your proteomics interpretation, but they warrant different levels of confidence. ### Peptide Mapping to Structural Regions Once you have a structure for your identified protein, you can map your detected peptides onto the three-dimensional model. This mapping serves several validation purposes. Peptides that map to structured regions of the protein are consistent with the protein being present in its native folded state. Peptides that map to disordered regions or to regions not resolved in the experimental structure require more careful interpretation, because these areas may be flexible or absent from the structural model. Mapping peptides to active sites and binding interfaces provides functional insight. If you identify a protein and your detected peptides include residues known to be catalytically important, this supports the interpretation that the protein is present in a functional form. Conversely, if your peptides cover a region that is buried in the protein core, this may indicate that the protein was digested under denaturing conditions, which is typical in bottom-up proteomics workflows. ### Structural Features as Functional Evidence Structural information can reveal functional features that are not obvious from sequence alone. Domains, active sites, post-translational modification sites, and interaction interfaces are all visible in three-dimensional structures. For example, the human bromodomain family consists of 61 protein interaction modules that recognize acetylated lysine motifs. Structural analysis of this family has identified conserved and family-specific features necessary for acetylation-dependent substrate recognition, and has shown that bromodomains recognize combinations of post-translational modifications instead of singly acetylated sequences, as reported in the [bromodomain family structural study](https://pubmed.ncbi.nlm.nih.gov/22464331). When your proteomics experiment identifies a protein with known structural features, you can use those features to interpret the biological meaning of your identification. A protein identified in a pull-down experiment that contains a known interaction domain supports the hypothesis that the protein participates in the expected interaction network. A protein identified in a differential expression experiment that contains a known catalytic site supports the hypothesis that changes in its abundance have functional consequences. ## Practical Workflow for Structural Validation of Proteomics Identifications ### Step 1: Compile Your Protein Identification List Begin with your processed proteomics results. Your list should contain protein accessions, preferably UniProt accessions, along with quantitative information such as fold changes, p-values, and numbers of peptides identified. For each protein you want to validate structurally, record the accession, gene name, and any functional annotation you already have. Organize your list by biological relevance. Proteins that are central to your research question, that show significant abundance changes, or that are novel in your experimental system deserve priority for structural validation. Proteins with well-established structures and functions may require less structural scrutiny than proteins with limited annotation. ### Step 2: Query the PDB for Experimental Structures Use the PDB search interface or the [NCBI structure resources](https://www.ncbi.nlm.nih.gov/) to check whether your protein of interest has an experimental structure. Search using the UniProt accession, gene name, or protein name. Record the PDB codes for any structures you find, along with the experimental method and resolution. For each PDB entry, note the following details: - Experimental method: X-ray crystallography, NMR, or cryo-electron microscopy - Resolution or quality metrics: Higher resolution generally means more reliable atomic positions - Chain and residue coverage: Which parts of the protein are present in the structure - Ligands and bound molecules: Whether the structure includes cofactors, substrates, or inhibitors - Biological assembly: Whether the structure represents the functional oligomeric state These details affect how you interpret the structure in the context of your proteomics data. A structure determined at high resolution with full-length coverage provides stronger evidence than a low-resolution structure covering only a domain. ### Step 3: Retrieve Predicted Structures from AlphaFold Database For proteins without experimental structures, query the AlphaFold Database. The database provides predicted structures for over 214 million protein sequences, covering nearly the complete UniProt database. Retrieve the predicted model and examine the per-residue confidence scores. The AlphaFold Database provides access through multiple mechanisms, including direct file downloads via FTP, Google Cloud Public Datasets, and programmatic access endpoints. Choose the access method that fits your technical comfort and computational environment. For a single protein, the web interface is sufficient. For large-scale analyses across many proteins, programmatic access or bulk downloads are more appropriate. ### Step 4: Map Your Peptides to the Structure Once you have a structure, whether experimental or predicted, map your identified peptides onto the protein sequence and structure. This mapping requires that you know the exact peptide sequences detected by mass spectrometry and their positions in the protein sequence. Several tools can perform this mapping. The [NCBI resources](https://www.ncbi.nlm.nih.gov/) provide sequence search systems that can align your peptides to protein records. Structural visualization software allows you to highlight peptide regions on the three-dimensional model. The key output is a list of which structural regions your peptides cover and which regions are not covered by your detections. Interpret the mapping in the context of your digestion protocol. In bottom-up proteomics, proteins are digested into peptides, typically with trypsin, and the detected peptides represent the observable portion of the protein. Some regions may not be detected because they lack appropriate cleavage sites, produce peptides outside the detectable mass range, or are difficult to ionize. Absence of peptides in a region does not necessarily mean the protein lacks that region. ### Step 5: Examine Structural Features in Your Identified Regions For the structural regions covered by your peptides, examine the features present. Use the structure to identify: - Domains and structural motifs - Active site residues - Binding interfaces - Post-translational modification sites - Regions of disorder or flexibility Compare these features with your experimental context. If you identified a protein in a specific condition, ask whether the structural features support a functional role in that condition. For example, if you identified a metabolic enzyme in a treatment condition where you expect altered metabolism, the presence of its catalytic residues in your detected peptides supports the interpretation that the enzyme is present and potentially active. ### Step 6: Integrate Structural Evidence with Quantitative Data Structural information becomes most powerful when combined with quantitative proteomics data. For proteins with significant abundance changes, examine whether the structural features explain the functional consequences of the change. A protein that increases in abundance and contains a known interaction domain may be increasing to participate in more interactions. A protein that decreases in abundance and contains a catalytic site may be decreasing to reduce a specific enzymatic activity. Consider also whether your quantitative data reveal changes in specific protein forms. If your proteomics workflow distinguishes protein isoforms or post-translational modifications, structural information can help you interpret which forms are present. For example, if you detect a phosphorylation site that is known from structural studies to regulate activity, you can use the structure to understand how that modification might affect protein function. ## Options and Tradeoffs in Structural Validation Approaches ### Experimental Structures versus Predicted Models The choice between using experimental PDB structures and AlphaFold predictions depends on availability and your confidence requirements. Experimental structures provide the highest confidence because they are derived from actual measurements. However, they are available for only a fraction of known proteins, and the coverage may be biased toward well-studied proteins, particular domains, or specific organisms. AlphaFold predictions provide near-complete coverage of known protein sequences, making them valuable for proteins that lack experimental structures. The tradeoff is that predicted structures carry uncertainty, and the confidence varies by region. You should examine the per-residue confidence scores and treat low-confidence regions with caution. The AlphaFold Database provides these confidence metrics with each model, and you should incorporate them into your interpretation. ### Single-Structure Analysis versus Large-Scale Annotation For a focused study of a few proteins of interest, you can manually retrieve and examine structures, map peptides, and interpret features. This approach gives you deep insight but is time-consuming and does not scale to entire proteomes. For large-scale analyses, you can use programmatic access to retrieve structures for many proteins and automate the mapping of peptides to structural regions. The AlphaFold Database provides programmatic access endpoints that support such workflows. The tradeoff is that automated analysis may miss context-specific features that require manual examination. A practical approach is to use automated methods for initial screening and manual examination for the most biologically relevant proteins. ### Web Interfaces versus Programmatic Access Web interfaces are accessible and require no programming skills. They are suitable for examining individual proteins and for learning how to interpret structural data. The PDB and AlphaFold Database both provide user-friendly web interfaces with visualization tools. Programmatic access is necessary for large-scale analyses and for integrating structural validation into reproducible workflows. The AlphaFold Database provides programmatic access endpoints, and the [Galaxy Training Network](https://training.galaxyproject.org/) offers accessible workflow training that can help you build reproducible analysis pipelines. The [Carpentries lessons](https://carpentries.org/lessons) provide foundational computing and data skills that are useful for implementing programmatic approaches. ## Observations and Measurements for Structural Validation ### What to Record for Each Protein For each protein you subject to structural validation, record the following information in your laboratory notebook or electronic records: - UniProt accession and gene name - PDB codes for experimental structures, if any - AlphaFold Database identifier and model version, if used - Experimental method and resolution for PDB structures - Per-residue confidence scores for AlphaFold predictions - Peptide coverage of the protein sequence - Structural regions covered by detected peptides - Functional features present in covered regions - Your interpretation and any follow-up experiments planned This record allows you to reproduce your analysis, compare across experiments, and provide evidence for your conclusions in publications or reports. ### Quality Metrics to Examine For experimental structures, examine the resolution and refinement statistics. Higher resolution structures generally provide more reliable atomic positions. For cryo-electron microscopy structures, examine the overall resolution and the local resolution in regions of interest. For NMR structures, examine the number of restraints and the root-mean-square deviation across the ensemble. For AlphaFold predictions, examine the per-residue confidence scores. The AlphaFold Database provides a predicted local distance difference test score for each residue, which indicates the expected accuracy of the prediction at that position. High-confidence regions are suitable for detailed interpretation, while low-confidence regions should be treated as unreliable. The Predicted Aligned Error viewer in the AlphaFold Database allows you to assess the relative positions of domains and the confidence in their arrangement. ### Coverage Assessment Peptide coverage is a critical measurement in structural validation. Calculate the percentage of the protein sequence covered by your detected peptides and examine the distribution of coverage across structural regions. High coverage of structured regions supports the identification, while coverage gaps in structured regions may indicate that the protein is present in a modified or partially degraded form. Compare your coverage with the structural model. If your peptides cover a region that is disordered in the structure, this is consistent with the protein being present in a flexible state. If your peptides cover a region that is buried in the protein core, this indicates that the protein was digested under denaturing conditions, which is standard in bottom-up proteomics. ## Common Failure Patterns in Structural Validation ### Assuming Structure Exists for Every Identified Protein A common mistake is to assume that every protein you identify has a structure in the PDB. In reality, experimental structure coverage is incomplete, and many proteins, particularly those from less-studied organisms or with disordered regions, lack experimental structures. Always check the PDB before assuming a structure exists, and use the AlphaFold Database as a complement for proteins without experimental coverage. ### Overinterpreting Predicted Structures AlphaFold predictions are valuable but have limitations. Low-confidence regions may be incorrectly modeled, and the predictions do not capture all aspects of protein behavior, such as conformational changes upon ligand binding or the effects of post-translational modifications. Do not draw strong conclusions from predicted structures without examining the confidence metrics and considering alternative interpretations. ### Ignoring Coverage Gaps Peptide coverage gaps are common in proteomics and do not necessarily indicate problems with your identification. However, ignoring coverage gaps can lead to incorrect conclusions. If your peptides cover only a small portion of the protein, you have limited evidence about the protein's overall state. Consider whether the coverage is sufficient for your conclusions and whether additional experiments, such as different digestion protocols or targeted mass spectrometry, would provide better coverage. ### Confusing Sequence Identity with Structural Identity Two proteins with similar sequences may have different structures, and proteins with different sequences may have similar structures. When using structural information to validate identifications, be careful not to assume that a structure from a homolog applies directly to your identified protein. Check the sequence identity between your protein and the structure you are using, and consider whether the structural features you are interpreting are conserved. ### Failing to Document Structural Evidence Structural validation is only useful if you document what you did and what you found. Failing to record PDB codes, AlphaFold identifiers, confidence scores, and coverage information makes your analysis irreproducible and weakens your conclusions. Maintain complete records for every protein you validate structurally. ## Limitations of Structural Information in Proteomics ### Experimental Structures Have Context Dependence Experimental structures are determined under specific conditions that may not reflect the cellular environment. Crystallographic structures are obtained from crystals, which may not represent the protein's native conformation. NMR structures are determined in solution but often at high concentrations. Cryo-electron microscopy structures may capture specific conformational states. Consider whether the structural context matches your experimental context before drawing conclusions. ### Predicted Structures Have Uncertain Accuracy AlphaFold predictions are highly accurate for many proteins, but the accuracy varies by protein and by region. The confidence metrics provided with each model indicate the expected accuracy, but they are not guarantees. For proteins with unusual features, such as extensive disorder or complex domain arrangements, predictions may be less reliable. Validate predicted structures against any available experimental data before using them for strong conclusions. ### Structural Information Does Not Capture All Biology Structures provide information about the three-dimensional arrangement of atoms but do not capture all aspects of protein biology. Protein abundance, localization, interactions, post-translational modifications, and degradation are not directly visible in structures. Use structural information as one line of evidence among many, and integrate it with your quantitative proteomics data, functional annotations, and biological knowledge. ### Coverage Bias in the PDB The PDB has historically been biased toward proteins that are amenable to structural determination, such as soluble, abundant, and stable proteins. Membrane proteins, intrinsically disordered proteins, and proteins from less-studied organisms are underrepresented. This bias means that the absence of a PDB structure for your protein does not indicate that the protein is unimportant or that it lacks structure. ## Safety and Regulatory Context for Structural Data Use ### Data Integrity and Reproducibility Using structural data in proteomics research carries responsibilities for data integrity and reproducibility. Record the exact versions of databases and tools you use, including PDB release dates and AlphaFold Database versions. This documentation allows others to reproduce your analysis and ensures that your conclusions are tied to specific data versions. ### Ethical Use of Structural Predictions AlphaFold predictions are generated by artificial intelligence systems and are available for nearly all known protein sequences. Use these predictions responsibly by acknowledging their limitations and by not overstating the confidence of conclusions based on predicted structures. When publishing results that depend on predicted structures, disclose that the structures are predictions and provide the confidence metrics. ### Compliance with Database Usage Policies The PDB, AlphaFold Database, NCBI, and EMBL-EBI resources have usage policies that govern how their data can be accessed and used. Review these policies before downloading large datasets or using programmatic access. For large-scale analyses, consider using the provided access mechanisms, such as Google Cloud Public Datasets for AlphaFold data, to avoid overwhelming the primary servers. ## Professional Escalation Criteria ### When to Seek Expert Structural Biology Support If your proteomics analysis depends critically on structural interpretation and you encounter any of the following situations, consider consulting a structural biologist or bioinformatics specialist: - You need to interpret a structure with unusual features or low quality metrics - Your conclusions depend on the precise arrangement of residues in an active site or binding interface - You are uncertain whether a predicted structure is reliable for your purposes - You need to design experiments based on structural information, such as mutagenesis or cross-linking studies - Your analysis involves protein complexes or assemblies that require careful interpretation of biological units ### When to Escalate Data Quality Concerns If you observe inconsistencies between your proteomics data and structural information, investigate before drawing conclusions. Possible explanations include: - Your protein identification is incorrect or the protein is a different isoform - The structure represents a different conformational state than the one in your sample - Post-translational modifications or processing alter the protein's structure - The structural model has errors or is incomplete If you cannot resolve the inconsistency, escalate to a colleague with structural biology expertise or consult the primary literature for the structure you are using. ### When to Seek Training If you are new to structural analysis or need to expand your skills, consider formal training. The [EMBL-EBI Training program](https://www.ebi.ac.uk/training) provides bioinformatics learning pathways and data-resource training that cover structural databases and their use. The [Galaxy Training Network](https://training.galaxyproject.org/) offers accessible workflow training that can help you build reproducible analysis pipelines. The [Carpentries lessons](https://carpentries.org/lessons) provide foundational computing, data, shell, Git, and programming training that is useful for implementing structural validation workflows. ## Integrating Structural Validation into Your Proteomics Pipeline ### Building a Reproducible Workflow Structural validation should be a defined step in your proteomics analysis pipeline, not an ad hoc process. Build a workflow that includes: 1. Input of your protein identification list 2. Query of the PDB for experimental structures 3. Query of the AlphaFold Database for predicted structures 4. Retrieval of structural models and confidence metrics 5. Mapping of detected peptides to structural regions 6. Annotation of structural features in covered regions 7. Output of a structural validation report The [Galaxy Training Network](https://training.galaxyproject.org/) provides training on building accessible and reproducible workflows that you can adapt for structural validation. The [nf-core documentation](https://nf-co.re/docs) describes community pipeline standards and usage that can inform your workflow design, particularly for large-scale analyses. ### Automating Structural Annotation For large proteomics datasets, automate the structural annotation process. Use programmatic access to the PDB and AlphaFold Database to retrieve structures for all proteins in your list. Use sequence alignment tools to map peptides to structures. Generate a table that summarizes, for each protein, whether a structure exists, what type of structure it is, what confidence metrics apply, and which structural regions are covered by your peptides. Automation allows you to apply structural validation consistently across your entire dataset and to identify proteins that warrant manual examination. The [Bioconductor project](https://bioconductor.org/) provides packages and workflows for reproducible genomic and proteomics analysis that can support automated structural annotation. ### Combining Structural Validation with Functional Annotation Structural validation is most powerful when combined with functional annotation from resources such as [NCBI](https://www.ncbi.nlm.nih.gov/) and [EMBL-EBI](https://www.ebi.ac.uk/training). Use these resources to obtain domain annotations, pathway memberships, and functional descriptions for your identified proteins. Combine this information with structural features to build a comprehensive interpretation of each protein's role in your experimental system. For example, if you identify a protein with a known bromodomain, the structural information can tell you which residues are involved in acetyl-lysine recognition, and the functional annotation can tell you which biological processes the protein participates in. Together, these lines of evidence support a specific interpretation of your proteomics result. ## Case Examples of Structural Validation in Proteomics ### Validating a Target Identification with Thermal Proteome Profiling Thermal proteome profiling is a technique that identifies drug targets by measuring changes in protein thermal stability upon ligand binding. In one study, researchers used thermal proteome profiling to identify nicotinamide phosphoribosyltransferase (NAMPT) as the target of the phenanthroindolizidine alkaloid PF403 in glioma cells. The identification was confirmed by multiple biophysical methods, and X-ray diffraction revealed that PF403 binds to NAMPT primarily through pi-pi interactions with residue Tyr188. The structure was deposited in the PDB with code 8Y55, as reported in the [thermal proteome profiling study](https://pubmed.ncbi.nlm.nih.gov/40486861). This example illustrates how structural information validates a proteomics-based target identification. The thermal proteome profiling data suggested that PF403 stabilizes NAMPT, and the crystal structure provided atomic-level evidence for the binding interaction. For your own proteomics experiments, this workflow demonstrates the value of combining proteomics measurements with structural validation. ### Using Structural Features to Interpret Cysteine Reactivity Quantitative reactivity profiling is a proteomics method that measures the intrinsic reactivity of cysteine residues in native biological systems. Cysteine is the most intrinsically nucleophilic amino acid in proteins, and its reactivity is tuned to perform diverse biochemical functions. However, there is no consensus sequence that defines functional cysteines, which has hindered their discovery and characterization, as described in the [quantitative reactivity profiling study](https://pubmed.ncbi.nlm.nih.gov/21085121). Structural information can help interpret cysteine reactivity data by revealing which reactive cysteines are located in functional sites. Hyper-reactive cysteines have been found to specify a wide range of activities, including nucleophilic and reductive catalysis and sites of oxidative modification. By mapping reactive cysteines onto protein structures, you can identify which ones are likely to be functionally important and which ones are merely surface-exposed. ### Applying Structural Analysis to Covalent Drug Discovery Live-cell activity-based protein profiling with mass spectrometry enables the proteome-wide quantification of compound reactivity. The CysDig platform was developed as an enrichment-free chemoproteomics platform for targeted covalent drug discovery in live cells. In a screen of 288 cysteine-reactive electrophiles against 300 functionally annotated cysteine sites, researchers identified covalent binders that liganded dozens of sites and found multiple instances of acute compound-induced protein degradation, as reported in the [CysDig platform study](https://pubmed.ncbi.nlm.nih.gov/41251312). Structural information is essential for interpreting such screens because it reveals which cysteine sites are functionally important and accessible to ligands. By combining chemoproteomics data with structural analysis, you can prioritize sites for follow-up studies and understand the structural basis of ligand binding. ## Records and Measurements for Structural Validation ### Maintaining a Structural Validation Log Create a structured log for your structural validation activities. For each protein, record the date of analysis, the databases and versions used, the structures retrieved, the confidence metrics, and your interpretation. This log serves as the basis for reproducible analysis and provides evidence for your conclusions. ### Measuring the Impact of Structural Validation Track how structural validation affects your proteomics interpretations. Record how many proteins in each experiment have experimental structures, how many have predicted structures, and how many lack structural coverage. Measure how often structural information changes your interpretation of a protein's function or your confidence in an identification. These measurements help you assess the value of structural validation and refine your workflow. ### Documenting Coverage and Confidence For each protein you validate structurally, document the peptide coverage and the confidence in the structural model. This documentation is essential for interpreting your results and for communicating your findings to others. Include coverage maps that show which regions of the protein are covered by your peptides and which structural features fall within covered regions. ## Frequently Asked Questions ### How do I find the PDB structure for a protein I identified? Search the PDB using the UniProt accession, gene name, or protein name. The [NCBI structure resources](https://www.ncbi.nlm.nih.gov/) also provide integrated search systems that connect sequence records to structural entries. If you find multiple structures, examine the experimental method, resolution, and coverage to select the most appropriate one for your purposes. ### What if my protein has no experimental structure? Check the AlphaFold Database for a predicted structure. The database provides predictions for over 214 million protein sequences, covering nearly the complete UniProt database, as documented in the [AlphaFold Database 2024 update](https://pubmed.ncbi.nlm.nih.gov/37933859). Examine the per-residue confidence scores and use the model with appropriate caution, particularly in low-confidence regions. ### How do I map my peptides to a protein structure? Align your peptide sequences to the protein sequence, then map the sequence positions onto the structure. Structural visualization software can highlight the mapped regions on the three-dimensional model. The [NCBI sequence search systems](https://www.ncbi.nlm.nih.gov/) can help you align peptides to protein records. ### What structural features should I look for in my identified proteins? Examine domains, active sites, binding interfaces, post-translational modification sites, and regions of disorder. The specific features that are relevant depend on your experimental context and research question. Use functional annotation from [NCBI](https://www.ncbi.nlm.nih.gov/) and [EMBL-EBI](https://www.ebi.ac.uk/training) resources to guide your examination. ### How much confidence should I place in AlphaFold predictions? Examine the per-residue confidence scores provided with each model. High-confidence regions are suitable for detailed interpretation, while low-confidence regions should be treated as unreliable. Consider validating predicted structures against any available experimental data before using them for strong conclusions. ### Can structural information distinguish between protein isoforms? Structures can help distinguish isoforms if the isoforms differ in regions that are resolved in the structure. However, many isoforms differ in regions that are disordered or not covered by the structure. Use peptide-level evidence from your mass spectrometry data to distinguish isoforms, and use structural information as a complement. ### How do I report structural validation in my publications? Report the PDB codes for experimental structures and the AlphaFold Database identifiers for predicted structures. Include the confidence metrics for predicted structures and describe how you mapped your peptides to the structures. Provide enough detail that others can reproduce your analysis. ### What training is available for structural analysis in proteomics? The [EMBL-EBI Training program](https://www.ebi.ac.uk/training) provides bioinformatics learning pathways and data-resource training. The [Galaxy Training Network](https://training.galaxyproject.org/) offers accessible workflow training for reproducible analysis. The [Carpentries lessons](https://carpentries.org/lessons) provide foundational computing and data skills. The [Bioconductor project](https://bioconductor.org/) provides packages and documentation for reproducible genomic analysis. ## Related Bioinformatics Guides - [The Protein Data Bank (PDB): Structural Formats, Coordinates, and Archival Validation Standards](/knowledge/bioinformatics/protein-data-bank-formats-archival-validation) - [The Protein Data Bank (PDB): Archival Standards, Structural Validation Metrics, and Bioinformatics Integration Protocols](/knowledge/bioinformatics/protein-data-bank-archival-validation) - [Genomic Data Analysis Tools: A Comparative Guide for Researchers](/knowledge/bioinformatics/genomic-data-analysis-tools-a-comparative-guide-for-researchers) - [Proteomics Analysis Tools: A Comparative Guide for Functional Interpretation](/knowledge/bioinformatics/proteomics-analysis-tools-a-comparative-guide-for-functional-interpretation) - [How to Interpret Gene Set Enrichment Analysis Results](/knowledge/bioinformatics/how-to-interpret-gene-set-enrichment-analysis-results) ## References and Further Reading - [NCBI Data Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. - [EMBL-EBI Training](https://www.ebi.ac.uk/training). European Bioinformatics Institute. - [Bioconductor](https://bioconductor.org/). Bioconductor Project. - [Galaxy Training Network](https://training.galaxyproject.org/). Galaxy Project. - [nf-core Documentation](https://nf-co.re/docs). nf-core. - [The Carpentries Lessons](https://carpentries.org/lessons). The Carpentries. - [Quantitative reactivity profiling predicts functional cysteines in proteomes.](https://pubmed.ncbi.nlm.nih.gov/21085121). Nature, 2010. - [AlphaFold Protein Structure Database in 2024: providing structure coverage for over 214 million protein sequences.](https://pubmed.ncbi.nlm.nih.gov/37933859). Nucleic acids research, 2024. - [Histone recognition and large-scale structural analysis of the human bromodomain family.](https://pubmed.ncbi.nlm.nih.gov/22464331). Cell, 2012. - [Thermal proteome profiling (TPP) reveals NAMPT as the anti-glioma target of phenanthroindolizidine alkaloid PF403.](https://pubmed.ncbi.nlm.nih.gov/40486861). Acta pharmaceutica Sinica. B, 2025. - [Enrichment-Free, Targeted Covalent Drug Discovery in Live Cells.](https://pubmed.ncbi.nlm.nih.gov/41251312). ACS chemical biology, 2025. > This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.