How do you translate a PDBx/mmCIF protein structure file, localize macromolecular descriptions and explain atomic coordinates without moving a single atom? PDBx/mmCIF is a dictionary-governed format used for macromolecular structural data. It combines readable metadata with structured identifiers, atom positions, chemical components, sequence information and experimental context. A professional localization process must distinguish descriptions intended for human readers from the categories and values that software uses to reconstruct a molecular structure.
Searches for PDBx/mmCIF translation, _atom_site labels, auth_asym_id versus label_asym_id, residue numbering, chain identifiers, CIF loops and chemical component codes often reveal a difficult boundary. A word like ‘chain’ may appear in prose, in a label or in an actual structural identifier. Replacing the wrong occurrence can break a molecular viewer or make a citation point to the wrong residue. Translating meaningful explanations requires the dictionary, the source structure and a clear rule about which fields are writable.
This international longform guide is designed for structural biologists, bioinformatics developers, chemistry educators, data curators, technical translators and scientific publishers. It explains how atom-site records, author identifiers, polymer sequences, ligands, coordinate values and metadata fit together; it also supplies practical testing scenarios. It does not claim that translated text can validate a biological model or establish the function, safety or medical relevance of a protein. Those questions require scientific evidence beyond the language of the file.
The sixty-second answer
Translate approved explanatory language about structural biology while keeping PDBx/mmCIF dictionary items, atom coordinates, chemical components, chain identifiers, residue numbers, sequence references and model relationships stable. Distinguish author numbering from standardized labels; do not turn a molecular model into a different dataset merely to make its names readable. Use dictionary-aware parsing and selected atom comparisons to validate the scientific object.
1. What PDBx/mmCIF does
A macromolecular structure dataset must describe atoms, chemical identities, chains, sequence relationships and the evidence from which a model was derived. PDBx/mmCIF provides a dictionary-based representation for these interconnected facts. A molecular viewer uses much more than the title of the structure to build the scene. Translating the title may improve discovery, while changing atom identifiers can make the model incoherent. Decide whether the project targets scientific publication, database metadata, visualization labels, training materials or an actual structured-file migration.
2. The mmCIF syntax layer
The CIF family uses data blocks, named items and loop structures to associate values with dictionary-defined categories. The syntax is not natural-language prose even when the file can be opened in an editor. An underscore-prefixed item name has a machine-readable identity. Replacing an item name with a translated phrase does not create an equivalent file. Preserve quoting, whitespace conventions, special value markers and loop grouping. Validate with a parser aware of the relevant dictionary, not merely a text viewer.
3. Data blocks and identity
A data block groups associated information within the file. Its identifier may serve as a reference in tooling or archives. Translators often receive a data block whose ID looks like a short English word, but that is not evidence that it should be rewritten. A localized catalog can attach a descriptive title in another language while retaining the stable source identifier. Test whether a viewer, repository or external script still opens the same block and resolves the same structural entry after localization.
4. Dictionary names are contracts
Names such as _atom_site.Cartn_x or _entity_poly_seq.num are defined data items, not labels to be casually reworded. The dictionary specifies what they mean and how they relate to other categories. A multilingual help page can explain these terms, but the stored keys must retain their recognized names. A seemingly small spelling change can turn a known item into an unknown one. Maintain a glossary separating canonical item names, local explanations and forbidden transformations. Verify field recognition with a dictionary-aware reader.
5. Why _atom_site matters
The _atom_site category records individual atomic positions and associated labels and properties. A single model may contain many thousands of rows. Each row participates in a coherent coordinate set and identification system. A translation that rearranges columns or associates a value with the wrong row can create a scientifically false structure without changing the set of words. Preserve row associations, declared items and numerical values. Test representative atoms by chain, residue, atom name and coordinates after any export or import.
6. Cartesian x, y and z
_atom_site.Cartn_x, Cartn_y and Cartn_z provide Cartesian coordinate components of atomic positions. They are mathematical values, not a sentence for translation. Reordering them, changing decimal punctuation or applying an unrequested unit conversion will move atoms in the rendered model. A language-specific interface can explain the coordinate axes but should not edit the stored numbers. Compare selected coordinates and interatomic distances across source and output. A visually plausible molecular shape is insufficient proof of geometric integrity.
7. Atomic coordinate units
Macromolecular structure coordinates are conventionally expressed in ångströms under the PDBx/mmCIF data model. A translation may give readers an explanatory conversion to nanometres, but the underlying coordinate fields should retain the standard interpretation. A figure that merely relabels the axis without recalculating values misstates distances. Keep unit conversion separate from linguistic localization. Where an educational diagram uses converted units, document the conversion and ensure that every related measurement is handled consistently.
8. _atom_site.id versus an atom label
An atom-site record can have its own unique site ID alongside an atom name associated with a chemical component. These are different identities. The atom-site ID distinguishes rows; the chemical atom label links to a molecular definition. A translator who changes all occurrences of ‘CA’ because it resembles a country abbreviation can corrupt a chemical atom name or confuse alpha carbon with calcium in different contexts. Identify the relevant field and chemical context before translating any explanation. Leave formal identifiers intact.
9. Element symbols
The _atom_site.type_symbol value identifies the chemical element. Symbols such as C, N, O or S are standardized chemical notation, not arbitrary English abbreviations. The name of the element can be translated in a teaching caption, while its symbol remains canonical. A molecule’s identity cannot be preserved by substituting a localized word for an element symbol in structured data. Test that the atomic composition and chemical component references remain unchanged after the localization process.
10. Chemical components
_chem_comp and related information identify chemical components such as amino acid residues, nucleotides, ligands and solvents. Codes are not freely translatable names. A residue code can look like an English word or short acronym but be an exact database identity. Explain the full scientific name in the target language as separate prose, and retain the component identifier used by software. Review unusual modified residues against the current Chemical Component Dictionary; do not guess their meaning from a familiar three-letter pattern.
11. _atom_site.label_comp_id
label_comp_id identifies the component associated with an atom through the standardized label system. Changing its code may make an atom appear to belong to a different chemical residue. A multilingual visualizer can show a localized residue description in a side panel while keeping the underlying label. When testing, select several atoms from different residue types, including water and a ligand, and confirm their component identities match the original. Fluency of the side-panel text does not validate chemistry.
12. label_atom_id and auth_atom_id
A file can store atom names under a standardized label system as well as author-supplied nomenclature. These fields help preserve both interoperability and the conventions used in an associated publication. They are not duplicates to consolidate automatically. A translator might be tempted to harmonize names to a preferred spelling, but doing so could break the mapping to the deposited record. Keep both namespaces distinct. Document which one a user-facing view selects and verify individual atom references in both systems.
13. What label_asym_id identifies
_atom_site.label_asym_id identifies an instance of a molecular entity in the structure’s labeling system and points to the structural asymmetry information. It can differ from an author-defined chain name. This difference is not a translation inconsistency. A structure may have several instances of the same molecule and additional nonpolymer entities with their own assigned labels. Translating a display label should not cause all instances with familiar names to be merged. Check the entity-instance mapping before simplifying a teaching diagram.
14. auth_asym_id and author chain names
_atom_site.auth_asym_id records an author-provided or PDB-assigned chain identifier. It may differ from label_asym_id even when both seem to refer to the same visual chain. The distinction matters when comparing a structural model with a figure, mutation list or manuscript using author nomenclature. A bilingual manuscript must not silently change a chain reference because a localized viewer displays a different internal chain label. Provide explicit mapping where necessary and keep the original identifiers stable.
15. Label versus author identifiers
PDBx/mmCIF deliberately supports standardized labels and author identifiers in parallel. These are alternate naming systems, not two language translations. A professional data guide should explain the system each reference uses and map between them when readers need that connection. A mutation described as ‘chain A, residue 100’ may not point to the same object under an unexamined label scheme. Before localizing a figure legend, confirm which namespace the original author intended. Structural correctness outranks typographic consistency.
16. _entity and molecular identity
The _entity category describes chemically distinct molecular entities, while other categories can identify multiple instances of those entities. A translated caption that treats an entity ID as a chain instance can confuse one protein type with the several copies present in an assembly. Keep entity, instance and atom identities separate. In a tutorial, illustrate the difference using the same protein sequence appearing twice in a model. Validate that entity references and the number of model instances remain unchanged.
17. _struct_asym references
_struct_asym connects asymmetry-unit instances to molecular entities. That relation is essential to interpreting how a model’s named chains are organized. A localization pipeline that renames one side of the relationship without updating the other can produce dangling references. A safer translation changes only the explanations surrounding the model. For an intentional data migration, verify every foreign-key-like relationship and use a dictionary-aware validation process. Do not infer that alphabetical labels have a geographic or linguistic meaning.
18. Polymer sequences
_entity_poly_seq describes monomers in polymer entities. Sequence position and chemical identity are structured data that determine how the model relates to the protein or nucleic acid sequence. Translating residue codes as words would destroy rather than explain the sequence. A multilingual educational article may describe a residue in the reader’s language, but the canonical sequence remains in its standard chemical notation. Compare residue order, identities, missing regions and sequence length against the source.
19. label_seq_id versus auth_seq_id
_atom_site.label_seq_id connects to standardized polymer sequence numbering, whereas auth_seq_id can preserve the author’s numbering choice. Author numbering need not be a simple consecutive series beginning at one; it can reflect a larger protein, historical convention or publication. A translator should not ‘correct’ gaps or reorder identifiers for a neater-looking local-language table. Check the mapping when explaining residue substitutions or catalytic sites. Report which numbering convention a figure uses so readers can reproduce the location.
20. Insertion codes
An insertion code can distinguish residues where author numbering does not fit a simple integer sequence. It is not an alphabetic suffix offered for translation. Removing it may collapse distinct positions into one label. A translated description of a residue such as 100A should preserve the code exactly when referring to the author-defined sequence. Validate residue lookup against the original structure and associated publication. A useful glossary distinguishes sequence numbering from typography and explains why apparently unusual residue labels can be legitimate.
21. Alternative conformations
Some atoms can occupy alternative positions represented by alternate-location identifiers. This is not a linguistic translation of the same atom name. The entries can represent modeled conformations, often with occupancies. If a localized export combines or drops alternate positions to simplify a table, the structural evidence changes. Keep alternate-location identifiers and their coordinate rows intact. When creating an educational illustration, describe the alternatives as modelled possibilities rather than presenting one as an absolute fact.
22. Occupancy values
Occupancy represents a model property associated with an atomic position; it is not simply a percentage chance that an atom ‘exists’ in an everyday sense. In a file, the numeric value is scientific data, not a presentation label. A translation should avoid exaggerating its interpretation and must preserve the recorded value. If a graphic rounds occupancies for teaching, identify the rounding. Compare alternate-location groupings and occupancy values against the source when validating a rendered structure.
23. B factors and displacement
_atom_site.B_iso_or_equiv gives an isotropic atomic displacement parameter or equivalent measure. It should not be translated as a generic biological risk score or claim of protein activity. Its interpretation depends on modeling and experimental context. A localized legend may explain what is being visualized, but the property and color scale must remain faithful. Test the actual values and legend mappings. Do not equate a visually warmer color with stronger scientific certainty without reference to the chosen visualization.
24. Missing-value symbols
In CIF notation, special markers such as question marks and periods can communicate distinct kinds of missing or inapplicable information under the format’s rules. They are not punctuation for stylistic editing. Replacing them with translated words in machine fields may break parsing and change semantics. A reader-facing table can show an appropriate explanatory label, but the structured source should remain intact. Test missing, unknown and not-applicable cases separately so an interface does not convert absence into an asserted zero.
25. Loop columns and row integrity
A loop associates multiple data item names with row values. The column ordering and tokenization define which value belongs to which item. Spreadsheet-based localization can accidentally sort only some columns, breaking every row association. A stable parser and writer will preserve the group structure. For educational exports, include column definitions and identifiers so readers can reconstruct relationships. Compare a sample of atoms and residues by unique identifiers, not just record counts; matching counts can still hide permuted rows.
26. Quoting and multiline text
CIF syntax includes conventions for quoted values and text fields. Human-language descriptions can contain whitespace, punctuation and multilingual characters, which require valid serialization. A naive replacement that inserts an unquoted value with spaces can disrupt parsing. Use a CIF-capable writer when changing permitted descriptive fields. Verify that the target string is recovered exactly by a reader, and that neighboring columns remain correctly associated. Successful display in a generic editor is not sufficient evidence that a scientific parser will accept the file.
27. Model numbers and ensembles
A structure entry may contain multiple models representing an ensemble. _atom_site.pdbx_PDB_model_num identifies relevant model membership. It is not a chapter number that can be renumbered for language-specific presentation. Collapsing an ensemble into one view can lose information about variability. A localized narrative should clarify whether it describes a single model or several. Validate model membership and compare selected coordinates across models, especially when an interface allows switching between them.
28. Crystallographic context
For crystallographic structures, unit-cell and symmetry information may be important to reconstructing the context of observed coordinates. These parameters are scientific constraints, not decorative metadata. Translating a caption about crystal symmetry should not change the symmetry operators or cell dimensions. An explanatory diagram may show a generated assembly, but it should distinguish generated copies from deposited coordinates. Verify the underlying parameters and the viewer’s interpretation rather than relying on the apparent geometry of a single screenshot.
29. Asymmetric unit and assembly
A crystallographic asymmetric unit is not necessarily the same thing as a biologically relevant molecular assembly. A localized summary that calls them identical can give readers an incorrect mental model. Assembly generation may rely on specified operations and relationships. Preserve the corresponding structured metadata and explain the distinction in prose. When testing a viewer, check whether the intended assembly appears with the correct number of chains and transformations. Do not translate ‘assembly’ into a term that implies an experimentally proven physiological complex.
30. Ligands and solvent
A structure can include ligands, ions, cofactors and water molecules as nonpolymer or solvent components. Their chemical identifiers and atomic coordinates remain part of the scientific record. Localized side-panel descriptions may name common substances in the target language but should not substitute a different chemical compound or assume every bound molecule is biologically functional. Verify component IDs with a chemical dictionary and confirm the displayed location and chain association. Handle unfamiliar ligand names by lookup rather than guessing from an acronym.
31. Chemical nomenclature
Scientific chemical names may have authorized systematic forms and common names. The best target-language expression depends on whether the reader needs a lay explanation, an established chemical name or an immutable database identifier. A molecular file’s component code must remain machine-readable. In prose, pair a precise identifier with a clear explanation where needed. Do not translate a residue code merely because it forms a recognizable word in the target language. Domain review should resolve ambiguous compounds.
32. Experimental method
PDBx/mmCIF can describe methods and data-collection context used to determine a structure. Crystallography, cryo-electron microscopy and NMR have different evidence types and limitations. Translating a method name should preserve what was actually done, not upgrade it to an apparently more rigorous technique. A visible atomic model is not by itself proof of experimental precision. Keep methodological metadata, model-versus-experiment distinctions and source references accessible. Check that multilingual summaries do not erase relevant uncertainty.
33. Resolution and its limits
Resolution is an important contextual measurement in some structural methods, but not a universal scalar quality score interpreted identically across techniques. A translated summary that labels a smaller resolution value simply as ‘more accurate protein’ may overstate what is established. Preserve the recorded value, unit and method. In educational text, explain what resolution can and cannot tell a reader. When comparing structures, avoid collapsing differences in data quality, refinement and experimental design into a single promotional claim.
34. Provenance and database references
Structural records can include related publications, accession identifiers and database links. These should remain resolvable after language changes. Translating an article title for discovery should not change its DOI or the structure’s stable accession identity. A multilingual catalog can add translated summaries but should link them to the same authoritative record. Validate references, identifiers and relationships. If a source reference is withdrawn or corrected, document that status rather than quietly paraphrasing the older claim as certain.
35. Software visualization and color
Molecular viewers can color by chain, element, secondary structure or calculated property. A localized legend must state the actual mapping. Colors are presentation choices rather than structural identifiers, and they may vary between software packages. A translator should not assume that blue always identifies one chain or that red means a biological danger. Test the same structure with the same visualization settings under every locale. Provide accessible alternatives that describe the selection criteria in words and codes.
36. Residue selection expressions
Analysis tools often use selection expressions that combine chain IDs, residue numbers and atom names. These look partly like everyday language but are executable queries. Translating them into natural words can break selection or point to the wrong atoms. Keep command syntax and identifiers literal, and translate surrounding instructional prose. Test examples against the exact viewer or programming library documented in the tutorial. A tutorial should explain which numbering namespace is used so readers can reproduce the selection.
37. Mutations and publication crosswalks
Publications may describe a substitution using author residue numbering while a downloaded structure displays standardized label numbering. A translation that changes the sequence number to match its figure without explaining the mapping can misidentify the mutation site. Keep a crosswalk between auth and label identifiers and confirm the associated chemical component and chain. When a manuscript refers to a residue absent from the observed coordinate model, state that limitation. Avoid inventing coordinates for unresolved regions.
38. Alignment and superposition
Researchers may align or superpose structures to compare their shapes. That is a geometric transformation with algorithmic choices, not a language translation. A bilingual caption can explain what alignment was performed without overwriting deposited coordinates. If a derived file intentionally contains transformed coordinates, label it as a derivative and preserve the original’s provenance. Validate transformation matrices and selected distances. Do not mix transformed and untransformed models within the same description without an explicit coordinate reference.
39. Missing atoms and incomplete models
A deposited model may omit atoms or residues that could not be resolved or were otherwise not modeled. Translation should not fill these gaps with guessed names or positions. A user interface can communicate that a region is missing, but it should distinguish unknown coordinates from a confirmed absence of the underlying biological sequence. Compare _entity_poly_seq information with modeled atom-site records. Accurate communication of missing evidence is as important as correct translation of observations.
40. A reproducible multilingual workflow
Begin with a read-only original structure, documented dictionary version and a scoped list of human-readable metadata. Protect identifiers, atom positions, chemical components and relational keys. Translate approved prose with scientific terminology review, then validate syntax, dictionary constraints and selected atomic records. Test viewer rendering, chain and residue lookup, search, accessible labels and cross-links to publications. Release a record of approved language changes and any unresolved ambiguities. The goal is easier access to the same scientific structure, not a new coordinate model.
Six worked cases
Case: chain A means different things in two views
A structure browser uses standardized label_asym_id values, while an older paper describes author chain identifiers. A translator sees the same letter in both documents and assumes they are interchangeable. A mutation annotation is then placed on the wrong chain instance. The correct response is to inspect the mapping between label and author namespaces and cite the convention in the translated figure. Preserve both identifiers in machine-readable data. Test an atom selection in the viewer and compare its residue chemistry and coordinates to the source publication; agreement of letters alone is not enough.
Case: residue 100 and 100A
A source manuscript refers to two distinct positions using an insertion code. During translation, an editor normalizes the second label to a clean-looking integer because suffixes seem untidy. The resulting table now points to the same apparent residue twice. That is a structural reference error, not a typographic improvement. A professional review preserves author residue numbering, insertion codes and the mapping to standardized sequence positions. When explanatory text uses a simplified description, the full unambiguous residue identifier should remain available to the reader.
Case: missing values become zero
An international data catalog exports scientific metadata into a spreadsheet and replaces a missing-value marker with the number zero for consistency. In the target-language description, the zero is then translated as an observed physical measurement. This introduces a new false fact. The corrective process restores the original marker and explains its applicable meaning, distinguishing absent or unknown values from measured zeros. Test both schema parsing and reader display. A concise translation that preserves uncertainty is more useful than a clean table that invents certainty.
Case: a protein structure becomes visually plausible but wrong
A script reconstructs a CIF loop after translating selected atom labels. During sorting, it orders the identifier column but not the coordinates. The output has the same number of atoms and looks protein-like when rendered, yet many atomic positions are associated with incorrect labels. A spot-check of the molecular image will not necessarily catch it. The remedy is to preserve row integrity with a standards-aware parser and compare selected atom IDs, component codes, coordinates and calculated distances to the source. The safest translation usually does not modify the atom-site loop at all.
Case: an accessible multilingual molecular lesson
An educator wants a protein structure lesson in Arabic, Korean and French. The structural file can remain stable while the learning interface translates explanations of chains, residues and atoms, offers a bilingual glossary and annotates selected model features. A useful interface provides literal identifiers when learners need to reproduce a selection. It explains that colors depend on the chosen visual scheme, distinguishes deposited structures from generated biological assemblies and includes accessible text descriptions. Testing should cover mixed-script formatting and the accuracy of selections, rather than approving only the appearance of translated captions.
Case: a drug-binding claim becomes stronger in translation
A structural paper describes an observed ligand position and qualifies its functional interpretation. A translated summary drops the qualification and states that the molecule has proven clinical efficacy. The structural dataset did not justify that claim. A responsible translation preserves the distinction between geometric observation, biochemical hypothesis and medical outcome. It links readers to the underlying entry and publication while retaining uncertainty. Review the title, figure legend, press summary and machine-translated snippets together, because a misleading claim can originate outside the file even when its coordinates are untouched.
Frequently asked questions
Can PDBx/mmCIF dictionary tags be translated?
Their definitions can be explained in other languages, but canonical item names such as _atom_site.Cartn_x are machine identifiers and should be kept intact in the file.
What is the difference between label_asym_id and auth_asym_id?
The first follows a standardized instance-labeling scheme; the second preserves an author-related chain identifier. They can differ and must be mapped explicitly when needed.
Can I translate residue codes?
Explain their chemical names in prose if helpful, but retain the standard codes in structured coordinates and sequence fields.
Does a translated molecular diagram change the atomic model?
It should not. Localization changes captions, labels and descriptions, while the scientific coordinates and identifiers remain unchanged unless a separate transformation is intended.
Why do missing values matter?
Unknown, inapplicable and absent information are not identical to a measured zero. Preserve the source semantics.
What confirms a correct release?
Dictionary-aware parsing, preserved cross-references, selected atom comparisons, chain and residue lookup, visualization tests and scientific terminology review.
Professional release checklist
Establish the source entry, dictionary version, desired languages and exact output format. Classify values as official identifiers, physical measurements, domain-specific terminology or translatable prose. Lock the first two groups, and review the third through authorized definitions and model context. Validate CIF syntax, loop structure, atom-site rows and links among entities, chains, sequences and chemical components. Compare selected coordinates and residue lookups before and after. Finally, check rendered labels, accessible text and publication links in every target language. Record any scientific ambiguity rather than inventing a neat answer.
Why structural translation needs a coordinate invariant
A high-quality translation should define invariants before localization begins. For a PDBx/mmCIF coordinate model, examples include the number of relevant atoms, atom-site identifiers, Cartesian coordinates, model membership, chain relationships and chemical component identity. Some user-facing descriptions may change freely, but these scientific invariants cannot be treated as stylistic choices. When a purported translation changes an invariant, the team should stop and determine whether the output is actually a modified scientific derivative. A comparison report can make this boundary visible to both scientific reviewers and language professionals.
A chain mapping table that readers can trust
When an article refers to an author chain, provide a mapping to standardized labels if the downstream dataset uses a different scheme. The mapping should include the relevant entry, model if necessary, entity or chain instance and numbering convention. Avoid assuming that all instances with the same entity are equivalent locations in an assembly. An explicit mapping makes translated instructions reproducible: the reader can select the intended residues using the named software and verify the same chemical environment. A beautiful bilingual figure without reproducible identifiers is insufficient for technical training.
Provenance in a world-facing science library
A translated structure description may circulate far beyond its original journal context. Future readers need to know the source accession, publication, method and translation version. Link the explanatory language to the record instead of inserting it into machine identifiers. If a scientific correction changes the structural interpretation, the translated summary should also be reviewed. This is not an instruction to translate the numbers in the model. It is a reminder that language can overstate evidence even when the file remains byte-for-byte identical.
Scientific uncertainty in captions
A protein illustration can tempt a writer to describe every curve as established fact. Experimental structure models have uncertainties and limitations. Translation should preserve phrases such as ‘modelled’, ‘observed density’, ‘consistent with’ and ‘proposed’. Omitting qualifiers may convert an interpretation into an assertion. Whenever a caption names a functional site, mutation or ligand interaction, verify the underlying residue identity and original evidence. The same principle applies to accessible descriptions: explain what the figure encodes while distinguishing visual representation from experimentally demonstrated biological effect.
A final paired review
An effective handoff pairs a language reviewer with someone capable of checking the structural model. The language reviewer examines terminology, legibility, ambiguity and reading flow. The structural reviewer checks item names, chemical codes, numbering conventions, atom references and visualized coordinates. Both compare the final product against the intended audience and source claims. Disagreements should be documented rather than resolved by silent guessing. This cooperation makes multilingual scientific resources easier to use without weakening the reproducibility that structural biology depends on.