Master Art of Translation | The Mass Spectrometry, mzML & Proteomics Evidence Translation System — How m/z, Charge States, Peptide Identifications, FDR and Provenance Keep Their Meaning

How do you translate mass spectrometry results, mzML files, mzIdentML peptide identifications and proteomics data into another language without changing the scientific evidence? Translators searching for LC-MS/MS terminology, m/z meaning, precursor ion translation, mass spectrometry metadata, peptide-spectrum matches, FDR interpretation or proteomics report localization encounter a problem that a bilingual dictionary alone cannot solve. A word such as “peak,” “charge,” “intensity,” “match,” “confidence” or “protein group” can be rendered naturally in a target language while the resulting sentence describes a different quantity, a different inference or a different experimental object. Reliable translation starts by identifying what was physically measured, what software inferred, which processing step intervened and which uncertainty still belongs to the claim.

Accurate mass spectrometry and proteomics translation must distinguish the raw instrument record from exported spectra, the observed mass-to-charge coordinate from neutral mass, an ion’s charge state from a chemical oxidation state, a chromatographic retention time from a biological time point, and a peptide-spectrum match from a confirmed biological protein. The HUPO Proteomics Standards Initiative supplies interoperable formats such as mzML for spectrum-level data and mzIdentML for peptide and protein identification results; its controlled vocabulary helps machines identify meanings that ordinary English labels cannot safely guarantee. The translator should not localize identifiers, edit binary spectra, invent an interpretation of a score or quietly turn a statistical threshold into proof. The task is to preserve the evidence chain while making its human-facing explanation usable.

This world-facing guide teaches how to translate proteomics reports, mzML metadata and tandem mass spectrometry terminology through a repeatable method: source intent, the hidden unit-and-identity problem, the measurement-to-inference mechanism, translation decisions, failure modes, worked diagnostics, and transfer to repositories, publications, instruments and multilingual research collaboration. It is for scientific translators, data curators, editors, bioinformaticians, lab teams, students and reviewers working across languages. Worked samples below are invented teaching records, not measurements or deposited discoveries. They show how to check units, scan references, collision energies, modifications, search databases, false-discovery-rate language, quality-control flags and data provenance before a translated claim is released.

Architecture: EDKSG-TRANS-MASTER-WORLD-800. The canonical translation system remains Master Art of Translation | The Complete System for Moving Meaning Between Languages. This child owns cross-language preservation of mass-spectrometric and proteomics evidence, not the science in isolation, not general vocabulary learning and not the internal mechanics of English. Existing mzML-specific and broader research-translation pages remain independent specialist owners.

Start here: choose the translation question, not the file extension

• “What is mzML and can I translate it?” Translate explanatory labels and documentation; keep structured identifiers, accession codes and numeric arrays intact unless a documented transformation is separately authorized.
• “What does m/z mean?” Read it as mass-to-charge ratio, a measured ion coordinate, not automatically molecular weight or neutral peptide mass.
• “How do I localize a peptide identification?” Preserve the spectrum reference, sequence, modifications, charge, search protocol, score name and threshold context.
• “Is 1% FDR a 99% certainty for every peptide?” No. A false discovery rate is a population-level error-control concept under a stated estimation method, not an individual probability claim.
• “What is the difference between mzML and mzIdentML?” One primarily represents instrument output and spectrum metadata; the other represents identification evidence and the links back to spectra.
• “Can I change decimals or units?” Only as an explicitly documented presentation conversion that is mathematically and contextually correct; never change raw scientific data to fit a language.
• “How should I report uncertain proteins or sites?” Preserve ambiguity groups, localization evidence, thresholds and stated limitations.
• “What is the release test?” Reconcile every translated conclusion against the original spectrum, processing record, source label and quality qualifier.

The search intent: translating scientific data without accidentally translating the science

Search queries for mass spectrometry translation often bundle four distinct activities. A translator may need to render a methods section about liquid chromatography–tandem mass spectrometry. A data engineer may want to expose translated descriptions beside a machine-readable mzML file. A research editor may need to revise peptide-identification tables for an international audience. A curator may need to translate sample metadata while depositing raw, analysis and identification files. All four involve language, yet only some text is legitimately editable. A stable workflow begins by sorting fields into explanatory prose, controlled terminology, immutable identifiers, numerical observations and derived conclusions.

The word “translation” is itself dangerous in a laboratory context. It also denotes a biological process in which ribosomes synthesize proteins from messenger RNA. This page concerns translation between human languages, not the biochemical process of protein synthesis. A search for “protein translation” may mean either subject. In published text, make the intended sense explicit immediately. When the source discusses gene expression, the sentence must distinguish biological translation from translation of a proteomics document. Conflating the two produces search confusion and, more importantly, inaccurate scientific communication.

The core editorial decision is an authorization boundary. A localized graph caption may legitimately explain “relative intensity” in the target language. The underlying intensity array, however, remains a measurement or processed data product, not a sentence to be paraphrased. A target-language user interface may adapt decimal presentation; an accession number such as a controlled vocabulary term or repository dataset identifier must remain exact. A reporter can simplify the explanation of a threshold, but cannot rewrite what the threshold measures. Every translation brief should say which layer can change, how it can change and what record proves that change was safe.

The hidden problem: a mass spectrum is several kinds of meaning at once

A spectrum has physical coordinates and intensities, but it is also the result of an acquisition method, instrument settings, ion-selection choices and software processing. A chromatographic peak is not the same thing as a fragment-ion peak. A detected ion is not automatically a unique chemical structure. A sequence assigned to a spectrum is not automatically a uniquely identified protein. These distinctions are not pedantic: their collapse changes which hypothesis a reader thinks the experiment supports.

An evidence-preserving translator therefore follows a chain: biological or chemical sample; preparation and digestion where applicable; chromatographic separation; ionization; precursor selection; fragmentation; mass analysis; raw signals; centroiding or feature detection; spectrum interpretation; database or library search; error assessment; peptide or molecule identification; protein inference; quantification; normalization; and the published claim. Not every experiment uses every link. A data-independent acquisition workflow has different evidence relationships from a traditional data-dependent experiment. A small-molecule experiment requires different identification confidence language from bottom-up proteomics. The chain tells the translator where to ask a scientist rather than invent a missing link.

Some terms are superficially ordinary English but technically controlled. “Scan” can refer to a recorded acquisition event rather than a visual scanning action. “Isolation” refers to choosing ions within a mass-to-charge interval, not purifying a single chemical identity. “Significant” may be a statistical descriptor, not an emphatic synonym for important. “Identification” can mean an assignment under computational criteria, not direct visual proof. These are exactly the situations where a glossary should retain domain meanings, accession references and counterexamples instead of storing only source and target words.

High-authority references and scope of the evidence

The source hierarchy for this guide is the HUPO Proteomics Standards Initiative’s mzML repository and specification materials, its mzIdentML 1.3.0 specification, the EMBL-EBI PRIDE dataset guidance, the PRIDE SDRF validator, and ISO 5060:2024 for human translation-output evaluation. Published specifications should be read with their stated version and context, and standards or validators checked for revisions at the time of deployment. A scientific glossary is not a substitute for a schema validator, and a schema validator cannot decide whether a translated claim overstates evidence.

HUPO-PSI’s mzML materials describe an XML format built to retain mass spectra and the associated acquisition and processing context, including representations such as profile and centroid data. mzIdentML v1.3.0, dated June 2024 in its specification, represents identification results and the references that connect search evidence to spectra. PRIDE describes datasets that may contain mass spectrometer output, processed identifications, quantification and experimental metadata. Its public submission guidance distinguishes original raw and analysis outputs from standards-based results. These materials support the terminology here; they do not certify any hypothetical example or authorize alteration of research records.

ISO 5060:2024 concerns evaluation of human, post-edited machine and unedited machine translation output. Its error-based approach can inform language review but should not be represented as a proteomics data validation standard. A translation can satisfy a readability evaluation and still corrupt an accession number. Conversely, a perfectly valid XML file can carry a misleading multilingual caption. This article proposes parallel evaluation: language and meaning fidelity, syntactic and structural validity of the data, and scientific evidence consistency. Use all three when a translation accompanies a data release.

The central mechanism: preserve a semantic envelope around each claim

Treat each translated assertion as a record with eight questions: Which source object is being described? Which instrument observation supports it? Which units and reference axes are active? Which acquisition and processing choices produced it? Which database, software or vocabulary version defines its interpretation? Which statistical or quality condition limits the result? Which evidence link allows another reader to verify it? Which parts are human language and which are controlled machine symbols? This envelope is more useful than asking whether every word has an equivalent.

In an ordinary sentence, “the peptide was detected” may look harmless. In a methods report it could conceal at least four different assertions: a precursor feature was observed, a fragmentation spectrum was acquired, a peptide-spectrum match was reported, or a peptide passed a specified confidence threshold. Those statements may be related but they are not interchangeable. A careful translation repeats the narrowest supported claim and carries any qualification along with it. Where the source is ambiguous, the translator records the ambiguity or obtains clarification instead of upgrading the evidence.

An auditable translation package can include a source segment ID, target segment, locked terms, field references, selected translation, alternative rejected rendering, scientific owner decision, reviewer, software version and release date. These are not decorative records. They make it possible to distinguish a linguistic correction from an altered scientific interpretation months later. In multilingual projects, a single terminology note should state whether a symbol such as m/z, a vocabulary accession, an amino-acid sequence, a fixed assay identifier or a file checksum must remain untouched.

The technical mechanism atlas: from ion physics to interoperable records

Each mechanism below owns a distinct translation decision. The four-part treatment states the technical object, the tempting but misleading rendering, the controlled translation choice and an audit question. Use these as a working terminology manual, not as permission to modify machine data. Instrument-specific conventions can vary: when the source includes more precise definitions, they control the translation.

01. The mass-to-charge coordinate (m/z)

What the record actually means. m/z is the ion mass divided by its charge number in the mass-spectrometric convention, not a synonym for neutral molecular weight. A reader may see two species with different charges but overlapping coordinates. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A translation that says molecular mass when it means m/z falsely converts an instrument axis into a molecular property. It can also obscure isotopic charge spacing. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep the symbol m/z and explain it near first occurrence. When a calculation is discussed, specify charge state, ion/adduct composition and the chosen mass convention. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Ask a reviewer whether a multiply charged peptide could generate the displayed coordinate. If the target wording rules that out without evidence, revise. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

02. Neutral mass and adducted ion mass

What the record actually means. Neutral molecular mass is a property of the uncharged species; the mass-to-charge value also depends on added or lost carriers such as protons or metal adducts. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A caption calling a protonated molecular ion the neutral molecule’s exact mass hides a chemical transformation and may mislead formula comparison. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Distinguish neutral formula, ion formula and theoretical or observed m/z. Preserve the ion notation, charge and any stated monoisotopic convention. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Reconstruct the relationship between neutral mass and the measured species before approving a translated molecular-weight sentence. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

03. Charge state, not oxidation state

What the record actually means. A mass spectrometric charge state describes the net charge on the ion being analysed. An oxidation state is a formal electron-accounting concept used in chemistry. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Using a target term ordinarily reserved for oxidation number can distort multiply charged peptide data, particularly in teaching material. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Use the established target-language mass spectrometry term and show z or a charge marker when useful. Retain superscripts where possible. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Read the spectrum’s isotope spacing and charge assignment in the source. The translation should refer to the same ion. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

04. Positive and negative ion modes

What the record actually means. Ion polarity indicates whether the instrument is detecting positive or negative ions. It is an acquisition condition, not a moral, analytical or biological evaluation. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A report may list the same analyte under different ion forms or polarity settings. A translation that omits polarity can make peaks look directly comparable. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Carry the polarity field into methods and captions, identify the ion species and avoid guessing from the sign of a plotted intensity. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Confirm the translated mode matches the raw acquisition record and the reported precursor/adduct labels. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

05. Monoisotopic mass and isotope envelopes

What the record actually means. A monoisotopic mass uses a defined isotope composition, while an isotope envelope represents related peaks from naturally occurring isotopes. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A translator may confuse the first visible isotope peak with the monoisotopic assignment or erase the charge information encoded in spacing. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Describe whether the algorithm assigned a monoisotopic peak, whether an envelope was deconvoluted and what uncertainty accompanies the result. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Check whether the reported mass refers to a selected isotope, an average mass or a deconvoluted neutral mass before changing the label. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

06. Peak centroid versus profile point

What the record actually means. Profile-mode data retain sampled peak shapes; centroiding reduces them to representative peak positions and intensities under an algorithm. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Calling centroid data raw profiles or an exported peak list the original detector trace conceals processing that affects resolution and noise representation. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep acquisition representation and later signal processing distinct. If software generated centroids, retain its name, version and parameters when supplied. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Inspect whether the file declares centroid or profile spectrum terms; compare the translated prose against that declaration. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

07. Spectral intensity and relative abundance

What the record actually means. A peak intensity may be an instrument response count or another signal measure. A normalized relative intensity is a transformation with a reference. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Writing “concentration” where the source reports intensity overclaims what has been measured; response factors and matrices affect quantification. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep absolute response, relative spectral abundance, chromatographic area and calibrated concentration as separate concepts. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Ask what calibration or normalization justifies each displayed label. No calibration should be invented in translation. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

08. Mass resolution and resolving power

What the record actually means. Mass resolution concerns the ability to distinguish nearby signals; resolving power often uses a ratio tied to a specified peak width definition. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Generic target words such as sharpness or accuracy can conflate resolution with mass error and instrument precision. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Preserve the measurement definition, the reference m/z and any full-width convention used. Use the instrument’s specified terminology. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. A reported resolving power must not turn into a mass accuracy claim. Cross-check the methods table. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

09. Mass accuracy and mass error

What the record actually means. Mass accuracy compares observed and theoretical masses; errors can be expressed in daltons, millidaltons or parts per million. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A numerical mass deviation without units or sign can be misleading. The term accuracy also differs from reproducibility and precision. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep ppm, Da and mDa units legible, and preserve whether the source uses absolute or signed error. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. If 5 ppm is rendered as 5 Da, quarantine the translation. If the source lacks a formula, do not invent one. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

10. Precursor ions and product ions

What the record actually means. A precursor ion is selected or isolated for a fragmentation event; product ions arise from that event and support structural interpretation. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. “Parent” and “daughter” may occur historically, but indiscriminate substitutions can break current instrument metadata and confuse isotope relationships. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Use precursor/product consistently in modern scientific prose. Where the source deliberately uses historical language, preserve intended referents. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Match each precursor field and product-ion assignment to the same MS stage or scan in the source. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

11. MS1 and MS2 levels

What the record actually means. An MS1 survey or precursor spectrum and an MS2 fragmentation spectrum answer different questions. Higher-order MSn experiments add more stages. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Calling a fragment spectrum an original survey scan destroys the evidentiary link between selected precursor and observed fragments. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep the MS level associated with every spectrum. Explain that MS2 commonly concerns tandem-mass-spectrometric fragment information without implying all workflows are identical. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Check that the native spectrum identifier in translated annotations resolves to the intended level. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

12. Isolation windows and coisolated species

What the record actually means. An isolation window spans a region of m/z values. More than one ion can be present and fragmented together. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Translating “isolated” as chemically pure or uniquely selected can overstate the specificity of a mixed fragmentation spectrum. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Report the nominal isolation target and width when given; distinguish isolation physics from downstream identification confidence. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Where coisolation is possible, do not call the fragmentation evidence exclusively attributable to one peptide without further analysis. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

13. Collision energy and fragmentation technique

What the record actually means. Collision energy is an acquisition parameter, and its notation may involve laboratory-frame energy, normalized settings or method-specific scales. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Changing units, deleting the phrase normalized, or treating different dissociation methods as synonyms can undermine replication. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep the source’s fragmentation label and energy convention, including whether an instrument uses normalized settings. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Check the exact acquisition method and parameter units against source metadata rather than interpreting an ambiguous number alone. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

14. Chromatographic retention time

What the record actually means. Retention time describes an analyte’s position along a chromatographic separation, usually measured relative to a run start. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A spreadsheet may display minutes while machine metadata carry seconds. Translating only the number can shift a peak by a factor of sixty. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Preserve time units, chromatographic context and whether the value denotes peak apex, scan start or integrated interval. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Test a translated example by converting seconds and minutes with an explicit conversion, never by silent substitution. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

15. Extracted ion chromatograms

What the record actually means. An extracted ion chromatogram follows ion signal across chromatographic time for a specified m/z window or related extraction rule. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. It is not equivalent to a mass spectrum, a total ion chromatogram or a measured analyte concentration time series. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. State the extraction window, chosen ion, retention-time axis and whether reported peak area follows integration. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Compare source and target labels for both axes; a chromatogram should not acquire a mass-spectrum x-axis. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

16. Total ion current and base peak chromatograms

What the record actually means. Total ion current typically summarizes signal across ions in a scan, whereas base peak intensity follows the most intense peak under a stated definition. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A prose translation that treats these chromatograms as interchangeable can misdescribe changes in mixture complexity and acquisition coverage. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Retain TIC and BPC acronyms where standard and expand their meanings precisely for target readers. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Confirm whether the plot was made from total signal or a maximal intensity, and whether filtering altered it. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

17. Data-dependent and data-independent acquisition

What the record actually means. DDA selects precursors according to rules during acquisition; DIA collects fragmentation information over planned isolation ranges, often with multiple precursors contributing. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A single translated phrase “targeted MS/MS” cannot stand in for both designs; their evidence and identification logic differ. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Preserve DDA or DIA and its source-specific definition, then explain the strategy without pretending all instruments implement it identically. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. For DIA, verify that the translation does not assert one fragmentation spectrum equals one uniquely isolated peptide. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

18. Selected reaction monitoring

What the record actually means. SRM and MRM are used for targeted measurement of predefined transitions, often linking precursor and product m/z values. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A target-language expression such as general spectrum scanning may conceal the planned transition design and calibration strategy. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep transition identifiers and their precursor/product coordinates paired, and distinguish quantifier from qualifier transitions where the method does. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Ask whether the translated instructions would allow the same transition to be located in the instrument method. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

19. Ion mobility and collisional cross section

What the record actually means. Ion mobility adds a separation dimension related to ion motion through a gas under specified conditions; collision cross section is a derived physical descriptor under calibration or theory. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Treating drift time as retention time or equating a mobility coordinate directly with CCS removes method dependence and calibration. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Preserve the mobility type, unit, calibration, gas and operating assumptions where present; distinguish measured mobility from reported CCS. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Test whether a reader can tell which numeric field is measured and which was calculated. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

20. Spectrum identifiers and scan numbers

What the record actually means. A native spectrum identifier can refer to a specific scan or spectrum under a file’s native-ID format. Its bytes matter for joining results. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. An editor may translate a word embedded in a native ID, trim punctuation or renumber scans during formatting. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep native IDs exact and use separate localized descriptions. Maintain a crosswalk only when transformation rules and losslessness are documented. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Resolve the target annotation against the underlying spectrum record. A human-readable caption is not a substitute for a stable identifier. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

21. Binary arrays and numerical precision

What the record actually means. mzML can encode arrays of m/z and intensity information using binary data representations and associated metadata. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Machine-readable arrays are not ordinary textual content. Replacing decimal characters or re-encoding values without a validated pipeline corrupts evidence. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Leave binary arrays untouched during language localization; validate any separate scientific format conversion with appropriate tools. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Compare counts, numerical tolerances and checksums as applicable. A fluent target paragraph cannot compensate for altered peak data. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

22. Controlled vocabulary accessions

What the record actually means. PSI-MS terms include stable accession identifiers that define technical concepts; a displayed English name is only one human-facing label. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A translated label might be a reasonable synonym yet refer to another CV concept or lose the meaning of a required unit. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep CV accession strings exact, retain versioned vocabulary context and localize descriptions alongside—not in place of—machine keys. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Check that the accession resolves to the original term and that any applicable unit or value constraint is maintained. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

23. Schema validity and semantic correctness

What the record actually means. An XML file may be schema-valid while human commentary misstates its scientific interpretation; a good sentence may accompany schema-invalid data. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Translation QA often tests grammatical fluency but does not test which fields remained machine-valid. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Use an XML/schema or format-specific validator for structural correctness, and conduct a separate scientific review of claims and units. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Record both test outcomes. Neither a passing grammar review nor a passing parser run alone is a release decision. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

24. Instrument model and acquisition software

What the record actually means. Instrument configuration, acquisition software and firmware or export versions can affect the structure and interpretation of generated records. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. An attractive localized equipment name may obscure the exact model needed for replication or compatibility. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Preserve model designations, software product names and version numbers; translate surrounding descriptions without changing identifiers. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Confirm that a reviewer can reconstruct the analytical platform from the translated methods section. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

25. Processing history and centroiding provenance

What the record actually means. Data processing can include conversion, smoothing, deisotoping, noise filtering, peak picking, recalibration and other software operations. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. A translated phrase like “unprocessed spectra” may be false after exporter actions. Some transformations cannot be reversed. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. List processing operations in chronological order when known; retain tools, versions and parameters rather than attributing everything to the instrument. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Compare processing metadata with the source report and identify the first stage at which a described property arose. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

26. Sample identity versus analytical run

What the record actually means. A biological sample can yield multiple injections, fractions, technical replicates and converted files; a file does not necessarily equal an independent specimen. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. Translating run identifiers as sample identifiers inflates biological replication or makes batch effects disappear. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Keep separate IDs for biospecimen, aliquot, fraction, injection, run and derivative data file; explain which comparisons use which level. Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Trace a reported group difference back to biological replicate counts before approving a translated experiment summary. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

27. Blank, carryover and quality-control injections

What the record actually means. Blank runs and control materials help detect contaminants, carryover and instrument drift; they are not identical to experimental specimens. A translator first decides whether this is an observation, an acquisition parameter, a computed quantity, a label or a scientific conclusion. That classification determines whether the target text may paraphrase the expression, must retain a symbol, or requires an expert clarification.

The hidden failure. If a translator labels a solvent blank as an untreated biological control, the quality narrative and denominators change. This is not a cosmetic difference. It changes what an international reader could infer about the instrument, chemistry, sample or confidence level. It is also the kind of fault that a normal spelling checker will not catch because both sentences may be grammatically correct.

Repeatable handling. Retain control purpose, acquisition order and relation to the surrounding samples; avoid flattening every non-treatment run to “control.” Keep source-term and target-term choices in a terminology ledger with the related record or schema field. If a data field is locked, translate a surrounding explanation, not the field itself. Where the source is inconsistent, flag the discrepancy rather than repairing the science by assumption.

Diagnostic. Ask whether a stated signal is analyte-associated, carryover-associated or a background observation before describing detection. For release, read the target on its own and ask what exact experiment it claims happened. Then compare that reading with source metadata and raw or derived evidence. The useful result is a traceable reconciliation, not a vague reviewer comment that the passage sounds technical.

A practical boundary test before continuing

Choose one sentence in the draft and color its meaning categories conceptually: source identifiers, scientific units, experimental operations, observed quantities, derived values, inferential judgments and ordinary rhetorical phrasing. An internationally usable translation may change the last category freely within the brief, but changes to the preceding categories need increasingly strong evidence and authority. The boundary is not decided by whether a term looks foreign or unfamiliar; it is decided by the role it performs in reconstructing the experiment.

For example, a caption saying that a “high-confidence protein” was “found in all samples” invites three independent questions. Does high confidence refer to peptide-spectrum-match scoring, protein inference or a separate confidence classifier? Does “found” mean one spectrum, one accepted peptide or a quantitative detection? Does “all samples” mean independent biological specimens or every injection, including controls? Resolve those questions before choosing natural target-language verbs. One short caption may contain three scientific claims and three different risk profiles.

For terminology acquisition and deliberate word practice, use the protected Vocabulary Learning Hub. For English sentence mechanics or register as an English-language problem, use How English Works V1.1. This article remains the translation bridge: identifying what those expressions must preserve as scientific evidence crosses languages.

Technical mechanism atlas, continued: identification, quantification and the publication chain

The following mechanisms move from acquisition data into statistical and biological interpretation. Each translation decision must preserve the level of inference. A detected precursor, reported peptide, protein ambiguity group, adjusted abundance and causal biological conclusion are not synonyms. In this part of the workflow, the most dangerous editorial change is often an increase in certainty that nobody explicitly approved.

28. Peptide-spectrum matches

Definition and evidence level. A peptide-spectrum match, or PSM, links a tandem mass spectrum with a candidate peptide sequence under a search model. It is an algorithmic association supported by mass and fragment evidence, not direct visual recognition of a peptide. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Translating PSM as proven protein detection promotes a match to a higher inference level. The search may report candidates with different scores, and one peptide can map to several proteins. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep spectrum reference, peptide sequence, charge, modifications, search conditions and score together. Name the reporting level: PSM, distinct peptide, protein group or quantified protein. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Select one PSM and trace it to the original scan and candidate sequence. Confirm that the translated conclusion does not claim unique protein identity from an ambiguous peptide. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

29. Database searching and sequence provenance

Definition and evidence level. Many bottom-up identification workflows compare spectra against theoretical peptides derived from a sequence database. The database’s organism scope, accession system and release are part of the analytical model. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. If a translated report simply says “searched the human database,” the reader may miss mixed proteomes, contaminants, custom sequences or a dated protein reference release. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve database name, version or download date, target/decoy construction, contaminants and applicable filtering. Avoid treating an accession’s display label as the full sequence identity. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Can an independent analyst rebuild the search universe described by the translation? If not, request missing provenance instead of proposing a likely database. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

30. Enzymatic digestion and cleavage specificity

Definition and evidence level. Trypsin and other proteases generate peptides according to enzyme specificity, preparation conditions and processing. Search settings may allow missed cleavages or semi-specific peptides. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A sentence that replaces “semi-tryptic” with “trypsin-digested” may conceal a search constraint. Peptide evidence changes when candidate space changes. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Differentiate physical sample digestion from enzyme specificity assumed by the search engine. Preserve missed-cleavage allowance and enzyme or no-enzyme settings. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Compare translated methods with search parameters. If the source reports one protease but the search used another specificity, flag rather than harmonize. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

31. Fixed and variable modifications

Definition and evidence level. A fixed modification is applied under the search model to all eligible sites; a variable modification is considered among candidate possibilities. These are computational assumptions, not always direct observations. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. If fixed carbamidomethylation becomes measured alkylation of every specific residue, the translation can falsely assert molecular verification. Variable oxidation is not necessarily experimentally proven. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep modification name, mass shift, eligible residues and whether it is fixed, variable or inferred. Distinguish sample chemistry from search configuration. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Inspect the search parameter list and a peptide annotation. Can a reader tell which chemical state is assumed and which is supported by fragment evidence? The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

32. Modification sites and localization

Definition and evidence level. An assigned peptide modification may be confidently associated with a sequence while its exact residue position remains uncertain. Site localization uses additional evidence and may be probabilistic. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translation stating “phosphorylated at residue 12” can be stronger than “phosphorylation assigned within this peptide; localization ambiguous.” This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve alternative candidate sites, localization score or probability and the specific algorithm when known. Do not move a modification label to the nearest visually convenient residue. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Ask the biological reviewer whether the evidence uniquely locates the site. If not, retain explicit ambiguity in the target caption. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

33. ProForma and proteoform notation

Definition and evidence level. ProForma offers a structured way to write peptide and protein sequences with modifications. The notation can encode evidence-related distinctions beyond a plain sequence string. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Translating or rearranging brackets, mass tags and accession references as punctuation changes the structure rather than the human explanation. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Lock the machine-readable sequence notation. Explain its syntax in surrounding target language and distinguish known modification identity from mass-only annotations. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Parse or validate the unchanged notation with a compatible implementation where possible. A spelling review cannot verify sequence markup. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

34. PSI-MOD and Unimod references

Definition and evidence level. Protein modification ontologies and curated mass-shift resources support standardized descriptions of modifications, with distinct histories and accession systems. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translator may treat an ontology accession as a generic synonym and replace it with a word that points to a different modification or residue specificity. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep ontology namespaces and accession strings fixed; translate display descriptions only after verifying their technical scope. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Trace each target-language modification description back to the source accession. Preserve unresolved specificity where the source does. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

35. False discovery rate as a set-level estimate

Definition and evidence level. FDR concerns the expected proportion of false discoveries among a set of accepted discoveries under a stated procedure and assumptions. It is commonly estimated using target-decoy strategies. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A 1% FDR threshold does not mean that any named peptide has a literal 99% probability of being correct. That individual interpretation is a category error. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Write explicitly which level is controlled—PSM, peptide, protein or modification site—and the threshold and estimation procedure. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Take a sentence claiming “99% certain protein identification.” If its only source is a 1% PSM FDR, reject it as overconfident. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

36. Local posterior error probability

Definition and evidence level. Posterior error probability, when calibrated and reported, concerns the probability that an individual match is incorrect under a particular model. It differs from a ranked-list q-value or FDR threshold. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Replacing PEP with FDR erases the level at which probability is assigned. A low PEP is not a guarantee of biological truth independent of model assumptions. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep the score’s exact name, scope, calibration method and whether smaller or larger numbers indicate better evidence. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Inspect the scoring software’s documentation. Does the translated probability apply to one PSM or a selected collection? The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

37. Q-values and threshold interpretation

Definition and evidence level. A q-value is commonly used to summarize the minimum FDR level at which a finding would be accepted in a particular multiple-testing setting. Its definition depends on method and analysis level. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Calling a q-value a p-value or a direct confidence score creates statistical errors that can propagate into figure annotations. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Translate q-value as a distinct metric and preserve its definition, direction, and acceptance rule from the analysis. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Compare an accepted and rejected example against the stated threshold. The translated inequality must not flip the decision. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

38. Target-decoy competition

Definition and evidence level. Target-decoy strategies create or use decoy candidates to estimate false matches under an analysis design. Competition rules and database construction influence the reported estimate. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. “Decoy” does not mean an experimentally added negative-control sample. Confusing them changes the provenance of statistical evidence. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep target, decoy, entrapment and control-sample language separate. Report the search design rather than just saying “verified against controls.” Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Ask how false discoveries were estimated and which data generated the decoys. A translated explanation should point to the actual source. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

39. Threshold pass flags

Definition and evidence level. The mzIdentML format can record whether an identification item passes a threshold. A pass flag is a recorded decision within a protocol, not a universal proof token. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translation might change “passed the applied threshold” to “correct” or “validated by experiment,” despite no independent orthogonal confirmation. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Retain pass/fail flags, associated protocol thresholds and any exceptional case where no threshold is specified. Distinguish a reported result from an independently verified result. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Can the reader reconstruct the rule that made a match pass? If the target text does not say, add a qualified explanatory note rather than an absolute label. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

40. Peptide-level and protein-level inference

Definition and evidence level. Peptides are often evidence for proteins, but one sequence may belong to multiple proteins or isoforms. Protein inference groups compatible candidates. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translator who replaces a protein group with a unique gene product can create a discovery absent from the source data. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve grouping terms, accession lists, indistinguishable candidates and shared peptide logic. State whether evidence is peptide-level or protein-level. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Check a shared peptide against at least two parent proteins. If both are possible, the target language must not quietly choose one. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

41. Protein ambiguity groups

Definition and evidence level. A protein ambiguity group can represent multiple proteins compatible with the available peptide evidence. Group representatives are often software conveniences. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A table’s leading accession may be styled as the only identified protein; that overstates uniqueness when the group contains alternative members. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Differentiate group identifier, representative accession, associated members and evidence supporting or distinguishing each member. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Review a row with multiple indistinguishable accessions. The translated heading must keep the ambiguity visible. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

42. Unique peptides and shared peptides

Definition and evidence level. A peptide described as unique usually has a defined relationship to a reference database or protein grouping. Whether it is unique depends on the searched sequence universe. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. The everyday adjective “unique” may imply biological exclusivity even when a peptide is merely unique within a limited database. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve the database-relative meaning of uniqueness and distinguish group-unique from protein-unique definitions where the analysis does. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Ask whether adding an isoform or changing the sequence database would alter the uniqueness claim. If yes, qualify it accordingly. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

43. Razor peptides and parsimony

Definition and evidence level. Some protein-inference algorithms allocate shared peptides to preferred groups according to parsimony or other rules. Such allocation is not proof of exclusive peptide origin. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. If a translated abundance table says every razor peptide physically came from the selected protein, computational assignment has been reified as a biological fact. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep the assignment rule as a method description. Where protein-level intensities use razor peptides, say so explicitly. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Rebuild one protein group from shared and unique peptides; confirm the target text does not erase the assignment procedure. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

44. Protein isoforms and reference sequence accessions

Definition and evidence level. Different isoforms can share most peptide sequences and differ only in small regions. Reference databases can update identifiers or isoform definitions. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Replacing an isoform accession with a familiar gene name may conceal that the evidence does not discriminate isoforms. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Carry the original accession, sequence version when available, and the supported level of isoform resolution. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Find whether a distinguishing peptide is present. Without that evidence, retain isoform ambiguity in the translated result. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

45. Fragment-ion annotations

Definition and evidence level. b, y and other fragment-ion series encode interpretations of peptide backbone or other fragmentation products; notation may incorporate charge, position and neutral loss. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A phrase such as “fragment b3 confirms protein B3” shows how easy it is to confuse an ion-series annotation with a biological identifier. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve ion-series symbols and indices. Translate the explanatory sentence about what matching fragments support, not the ion notation. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Check that a labeled fragment belongs to the stated candidate sequence and acquisition method, including charge where reported. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

46. Neutral losses in MS/MS

Definition and evidence level. A fragment ion can be annotated with a neutral loss such as water or ammonia under a specific interpretation. The annotation may use a mass shift or composition. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Dropping the loss qualifier changes the assigned fragment mass. Calling a neutral loss an entire disappearing peptide exaggerates the event. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Retain the loss species and annotation, including whether it is a proposed or assigned fragmentation pathway. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Confirm the reported fragment’s m/z and notation remain mutually consistent after localization of nearby prose. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

47. Search-engine scores and directions

Definition and evidence level. Identification tools use different scores: some increase with quality; others decrease. Names and scales are method-specific. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Replacing a named score with “confidence percentage” may be wrong when it is an expectation value, transformed significance or uncalibrated ranking statistic. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep the native score name and documented direction, then provide plain-language context without inventing calibration. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Compare the same two PSMs under the source ordering. If the target explanation reverses which one is better, it fails. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

48. Mass tolerance and precursor matching

Definition and evidence level. Search engines apply precursor and fragment mass tolerances that constrain candidate matches; units may be ppm, Da or other specified measures. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Mislabeling fragment tolerance as precursor tolerance or swapping ppm and Da can lead readers to reconstruct a different search. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Record each tolerance with its ion level, units and whether it reflects instrument resolution, search setting or post-search filtering. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Cross-check one theoretical match against the stated window. If the target sentence would include a source-rejected candidate, revisit the term. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

49. Open searches and unexpected modifications

Definition and evidence level. Open modification searches permit a broad range of mass differences, helping detect unexpected changes but increasing the need for careful interpretation. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translated “identified modification” might merely be an unexplained mass shift consistent with several chemical processes. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Distinguish observed mass shifts, algorithmic hypotheses and chemically assigned modifications. Preserve localization and validation limits. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Ask what orthogonal evidence, if any, distinguishes competing explanations for the same mass delta. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

50. Spectral libraries and reference matches

Definition and evidence level. Spectral libraries store reference spectra and metadata for searching or validating observed signals. Entries can be empirical, predicted or curated according to the library. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A library match does not always mean the analyte was independently confirmed under the local experimental conditions. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Report library identity, version, entry type, similarity metric and acquisition compatibility when available. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Inspect whether the library entry was observed experimentally or predicted. The target should not make a predicted reference sound like a physical standard. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

51. Library-free DIA analysis

Definition and evidence level. Some DIA workflows infer evidence using predicted spectra or other models without an experimental library built in advance. Others use empirical or hybrid reference resources. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Calling every DIA result “library-confirmed” erases model provenance and the difference between predicted and observed reference information. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Name the analysis approach, software and reference source. Avoid a blanket assertion of direct validation. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Trace one peptide claim through extraction, scoring and error control. Translate the evidence path actually used. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

52. Precursor-level and peptide-level quantification

Definition and evidence level. A quantified precursor feature refers to a particular peptide ion form, often with charge and modifications. A peptide-level value may summarize several features. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A target-language report may refer to the precursor’s area as total protein abundance without aggregation rules or normalization. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve the level of quantification, charge, modification and normalization status. Describe any aggregation separately. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Inspect whether a number is raw area, normalized precursor quantity, peptide rollup or protein-level estimate. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

53. Label-free quantification

Definition and evidence level. LFQ combines signals across runs through methods such as feature matching, normalization and protein-level summarization, depending on the software. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. The label-free adjective does not mean no calibration, no normalization or no missing values. It simply distinguishes the broad labeling strategy. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Translate the method and exact output scale. When the software’s term is a branded algorithm, preserve its identity. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Compare replicate structure and normalization notes. Do not convert LFQ intensity into concentration unless justified. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

54. Reporter-ion quantification

Definition and evidence level. Isobaric tagging experiments may infer relative sample abundance from reporter-ion signals released upon fragmentation under a specified method. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translation that ignores channel mapping or correction factors can attribute a signal to the wrong experimental condition. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve reagent design, channel IDs, sample mapping, interference correction and ratio reference. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Trace two channels through the design table into the output. A translated legend should reconstruct the same biological comparison. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

55. Stable isotope labeling

Definition and evidence level. Stable isotope labels can distinguish samples or chemical states through predictable mass differences. Not every isotope-labeled signal is an absolute quantitative standard. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Words like heavy or light refer to isotope labeling conventions, not necessarily sample weight or physical density. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Use established isotope and labeling terminology with atom notation and labeling design; retain internal channel names. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Check whether the translated comparison distinguishes label identity from condition identity, especially after label swaps. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

56. Absolute quantification and calibration

Definition and evidence level. Absolute concentration claims need calibrated relationships to amount, response and often internal standards, depending on method. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A normalized abundance unit or arbitrary intensity cannot be rendered as ng/mL simply because the audience prefers concentration units. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve calibration model, unit, matrix, recovery assumptions and the analytical level being quantified. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Ask whether a traceable standard and unit conversion support the claimed amount. If not, retain relative language. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

57. Limits of detection and quantification

Definition and evidence level. LOD and LOQ describe distinct analytical performance criteria under a method, with potentially different statistical definitions. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Replacing below quantification with absent or zero creates false negative claims and distorts comparisons. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep below detection, detected below LOQ, nondetect and missing values distinct; retain any stated threshold and units. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Use a hypothetical low-signal specimen to confirm the translation does not convert a censored observation into a measured zero. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

58. Missing values and imputation

Definition and evidence level. A missing value in a quantitative matrix can arise from non-detection, failure, filtering or alignment. Imputation adds modeled values under assumptions. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translated table that labels imputed zeros as observed absence or treats missing as zero changes effect sizes. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve missingness code, imputation rule, affected features and the version of the derived table. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Review which cells were measured and which inferred. The target caption must disclose the same distinction. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

59. Batch effects and normalization

Definition and evidence level. Measured differences can reflect acquisition order, instrument drift, sample preparation and batch structure as well as biology. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translation that calls corrected intensities raw data suppresses a transformation. A note saying “batch removed” may overstate correction. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Describe normalization, batch correction and its assumptions in operational terms. Keep batch assignments and biological grouping distinct. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Inspect an uncorrected and corrected summary to ensure the translated claim identifies the proper analysis stage. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

60. Biological replicates versus technical replicates

Definition and evidence level. Biological replicates sample independent biological units under the design; technical replicates repeat processing or measurement of material. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A replicated injection is not an additional biological subject. Inflating sample size in translation changes how convincing the statistical analysis appears. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve sample hierarchy, n at each level, randomization and paired structure. Name replicates according to actual independence. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Recreate the experimental design from the translated text. If the count of independent units changed, block release. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

61. Sample metadata in SDRF-Proteomics

Definition and evidence level. Sample and Data Relationship Format tables connect sample attributes, experimental factors, protocols and assay files in a structured representation. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translated sample label may no longer match an assay file; a renamed factor value may split groups that were originally identical. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep factor names, material identifiers and file references stable as required by the exchange workflow. Store translated descriptions in controlled presentation layers. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Validate an SDRF table with an appropriate validator and compare each assay row against the original sample sheet. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

62. File-level data relationships

Definition and evidence level. A dataset may combine vendor raw, converted mzML, processing parameters, identification results and quantification summaries with explicit references among them. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Renaming files for aesthetics without updating links can sever provenance; silently discarding a processing file can prevent reproducibility. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Maintain stable file mappings and checksums when appropriate. Document a canonical manifest, derivative relationships and release version. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Follow a peptide result back to the correct raw run. If its file link points elsewhere, the translation bundle is not complete. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

63. ProteomeXchange and PRIDE accession meaning

Definition and evidence level. Repository project identifiers such as ProteomeXchange accessions support discovery and citation, but the identifiers alone do not establish the confidence of every result. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A dataset accession is not equivalent to a journal endorsement or confirmation of an individual protein. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep accessions exact and separate public availability, deposition status, release version and scientific interpretation. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Use the accession to locate the correct dataset and confirm that target-language prose describes the deposited version. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

64. Complete and partial submission terminology

Definition and evidence level. Repository compliance classifications may depend on the standards-based files supplied; at PRIDE, guidance distinguishes submissions including validated standard result formats from those using only native analysis outputs. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. The everyday adjective partial can misleadingly imply scientifically incomplete work or missing raw data in every case. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Translate the repository’s technical submission type and explain its administrative meaning. Do not conflate classification with research validity. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Compare the actual file inventory with the repository’s current rules before assigning any status in target language. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

65. mzIdentML result representation

Definition and evidence level. mzIdentML links peptide, spectrum, database, protein detection and analysis protocol entities. It carries scores and threshold semantics alongside relationships. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Copying a peptide table while losing spectrum references converts a traceable identification into an unsupported list. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Maintain identifiers and references as intact structured data. Translate interpretation in a separate layer that preserves relationships. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Validate structure and resolve at least one SpectrumIdentificationResult through its PeptideEvidence and SpectraData links. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

66. mzTab-P, mzTab-M and result summaries

Definition and evidence level. Tabular exchange formats can summarize downstream findings, but a readable summary is not the same as the raw spectra or complete identification evidence. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. Treating metabolomics mzTab-M and proteomics output as interchangeable, or promoting a table-only result to raw evidence, confuses standards and analysis stages. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Record the exact format, version and intended domain, then translate column explanations under its field definitions. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Check whether each column denotes a feature, molecule, peptide, protein, abundance or statistic. Labels alone are insufficient. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

67. mzQC quality metrics

Definition and evidence level. mzQC represents quality-related metrics for mass spectrometry workflows in structured form. QC evidence describes instrument or data fitness, not necessarily biological truth. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translated QC-pass statement may be mistaken for confirmation that every protein identification is correct. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Keep QC metric identifiers, calculation method, thresholds and instrument/run scope. Separate dataset-level quality from identification confidence. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Can the reviewer reproduce which QC test passed and identify what that test does not assess? The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

68. Cross-linking mass spectrometry

Definition and evidence level. Cross-linked peptide evidence involves linked residues, linker chemistry, possible alternative sites and specialized scores. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translated “contact confirmed” may turn a cross-link constraint into a direct structural proof or overstate site localization. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve linked sequences, residue candidates, linker identities, spectrum evidence and model limitations. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Trace each partner peptide and candidate link site. Require specialized review before making a mechanistic structural conclusion. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

69. Top-down, middle-down and bottom-up workflows

Definition and evidence level. Proteomics strategies differ in the molecular entities entering the mass spectrometer: intact proteoforms, larger peptide fragments or digested peptides. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A target-language sentence that describes intact-protein measurements as tryptic-peptide identification can completely misstate the method. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Name the strategy and sample-preparation level, then keep identification and modification claims at that molecular scale. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Ask what physically entered the source analyzer. If the translated description implies another substrate, correct it. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

70. Native files and evolving formats

Definition and evidence level. Instrument-native data formats and community exchange representations coexist. Some open formats, converters and future drafts continue to evolve. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A source may call a converter output raw in a casual sense while the exported file has undergone processing. A draft format must not be described as universally stable. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve format name, version, conversion tool and validation status. Distinguish stable releases from experimental format work. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Check which reader was used, whether conversion was lossless for required properties, and what source data remain archived. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

71. Release versions and errata

Definition and evidence level. Data repositories and software packages can update records, metadata and interpretation after initial release. A citation can identify a different state from the latest accessible files. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translated document may cite the same accession yet describe an obsolete manifest or earlier confidence threshold. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Include release date, version, relevant change log and the exact files consulted. Avoid silently blending old and new outputs. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Compare the translated numerical summary with the same version of the files, not a later export from the same project. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

72. Uncertainty in biological conclusion

Definition and evidence level. A statistically accepted peptide observation is one link in a larger biological argument. Cellular localization, function and causation require additional evidence. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A translator might strengthen “associated with,” “suggests” or “consistent with” to “proves” because the target sentence sounds more decisive. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Preserve hedging verbs, effect estimates, intervals, analysis scope and study limitations. Escalate any sentence that appears more certain than the source. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Translate the same conclusion back into a plain evidence statement and check whether it claims a new mechanism or causality. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

73. Privacy and sensitive sample annotations

Definition and evidence level. Human-derived datasets may include clinical, donor or other sensitive metadata. Translation does not remove governance obligations around data use. Before drafting the target sentence, identify the smallest source entity that supports it: spectrum, precursor feature, peptide candidate, group, sample, experiment or publication-level interpretation. The distinction controls both technical vocabulary and the verbs that accurately describe the finding.

A natural-sounding but wrong translation. A target-language note can accidentally reidentify a participant by combining rare descriptors or inventing unapproved demographics. This failure can survive ordinary proofreading because the target language remains fluent. It should be treated as a scientific meaning error, not as stylistic preference, whenever it changes the claimed object, statistical scope, provenance or experimental comparison.

Repair method. Follow data access and deidentification controls; translate only approved descriptors. Avoid embellishment, even when attempting to clarify. Lock any accession, numerical measure or machine key that participates in a join. Localize the explanation outside the data structure, document any proposed alternate term, and retain the source uncertainty. For recurrent reports, store this decision with an example and a counterexample in the translation termbase.

Diagnostic and reviewer question. Review the translated metadata as a whole for identifying combinations, not just the spelling of personal fields. The expert reviewer should be able to point to the original field, protocol or analysis report that settles the decision. When the evidence is missing, describe the uncertainty plainly instead of filling it with a likely scientific assumption.

The repeatable 30-gate translation workflow: from source file to defensible release

The workflow works as a sequence of evidence gates, not a ceremonial checklist. A team may combine adjacent gates for a small figure caption, while a large multilingual repository release may need an assigned owner and test artifact for each. What matters is that no one can satisfy a later gate by assuming an earlier question was answered. Every gate produces a decision, a record of supporting evidence and a defined reason to stop or consult an expert.

Gate 01 — Define the deliverable

Action. Separate translatable narrative, graph labels, glossary definitions, instrument method descriptions, curated sample metadata and machine-readable files. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A written brief listing the required languages, audiences, repository formats, supplied source versions and field classes. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not let a general request to translate data authorize changes to structured measurements or code. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 02 — Identify the primary scientific owner

Action. Establish who can approve wording about experimental design, uncertainty and statistical claims, separate from the linguistic reviewer. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. Names or roles of the sample owner, analyst, format specialist, language reviewer and final release authority. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Stop when no one can adjudicate an ambiguous protein inference or missing threshold. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 03 — Freeze the source snapshot

Action. Copy the exact version of manuscript, tables, schemas, spectra and analysis outputs selected for translation, preserving their release identifiers. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A manifest of filenames, dates, hashes where appropriate and links to source records. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not blend a new analysis export with an earlier methods paragraph without recording the change. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 04 — Classify every content element

Action. Mark tokens as ordinary prose, stable technical term, symbol, accession, file key, scientific value, formula or inferred statement. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A field map assigning translate, explain, lock, convert under authorization, or escalate status to each class. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. A field with machine-join semantics must not be translated as if it were ordinary text. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 05 — Disambiguate translation itself

Action. Decide whether any appearance of translation refers to language transfer or ribosomal protein synthesis. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A two-sense glossary note and source sentence examples. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Reject a headline that attracts readers expecting a gene-expression explanation while promising bilingual translation. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 06 — Establish the evidence hierarchy

Action. List instrument data, converted data, search results, summary tables, narratives and publication claims in order of provenance. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A diagram or ledger linking sample to run, spectrum, assignment, quantitative output and conclusion. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not let an editorial summary outrank the evidence it was meant to describe. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 07 — Select official terminology sources

Action. Check PSI-MS, mzML, mzIdentML, repository metadata standards and any instrument-specific controlled terms for the relevant release. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. Source URLs, vocabulary version references and approved bilingual renderings with accession identifiers. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Escalate a term that has multiple plausible target-language translations but different technical meanings. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 08 — Audit unit-bearing fields

Action. Record m/z, Da, ppm, minutes, seconds, intensity scale, concentration, charge and other relevant unit conventions. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. An inventory of numeric columns, axis titles, unit fields and stated conversion rules. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not silently substitute a unit, alter numerical precision or localize an encoded binary array. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 09 — Rebuild the experiment hierarchy

Action. Distinguish donor, biospecimen, aliquot, fraction, analytical injection, run and processed file. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A design table mapping each biological unit to its technical acquisitions. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Block any translated claim in which technical injections become independent biological subjects. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 10 — Map raw to converted data

Action. Record when vendor output became mzML or another exchange representation and what processing occurred. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. Converter name, version, settings and source-to-derivative file mapping. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not call a centroided export an untouched detector signal. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 11 — Resolve spectrum identity

Action. Test whether each spectrum ID or scan reference used by derived reports points to the intended file and acquisition event. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A reproducible mapping of at least a representative PSM to a native spectrum identifier. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not rename or localize an identifier embedded in a machine link. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 12 — Verify acquisition semantics

Action. Document polarity, MS level, isolation, fragmentation method, ion mobility and related measurement settings. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A methods extraction sheet anchored in instrument or file metadata. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Reject target text that describes the wrong ion stage or attributes every fragment to a unique precursor. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 13 — Fix terminology for charge and ions

Action. Choose controlled target-language explanations for m/z, charge state, precursor, product ion, adduct and isotope pattern. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A small bilingual termbase with examples and explicit non-equivalences. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. A translation that merges neutral mass with observed m/z must be revised. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 14 — Resolve chromatographic context

Action. Identify separation technique, retention-time units, extraction windows, feature integration and run alignment where used. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A figure-data crosswalk for time axes, peak descriptions and source methods. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not describe a chromatographic peak as a fragment-ion peak or an uncalibrated intensity as concentration. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 15 — Identify search scope

Action. Preserve sequence database, version, enzyme rules, missed-cleavage settings, modifications and search tolerances. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. An analysis-protocol record and its corresponding translated methods sentences. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not fill an absent database release or search parameter from memory. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 16 — Separate identifications by level

Action. Identify PSMs, distinct peptides, modified peptidoforms, protein groups, isoforms and biological conclusions. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A result-level classification for each identification statement. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not translate a representative group accession into a unique protein claim. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 17 — Audit statistical language

Action. Record score definitions, score direction, PEP, q-value, FDR, threshold flags and levels at which they apply. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A table connecting each statistical phrase to a named analysis field. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not turn a set-level FDR threshold into a probability of truth for each identification. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 18 — Preserve site localization uncertainty

Action. Check whether modification identities and residue positions are supported uniquely or remain ambiguous. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A list of modified peptide examples with site probabilities or alternative sites where given. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. If the source does not support a unique residue, the target must not create one. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 19 — Review quantitative scale

Action. Identify whether values are raw signal, normalized precursor intensities, peptide summaries, protein rollups or calibrated amounts. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A units-and-scale map for all exported numerical figures. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Reject arbitrary units presented as absolute concentrations. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 20 — Protect comparison structure

Action. Validate labels for control groups, batches, technical replicates, randomization and time points. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A target-language experimental design table consistent with the source run map. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not swap treatment and baseline while translating concise axis labels. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 21 — Handle missingness explicitly

Action. Differentiate zero, non-detect, below LOQ, missing, excluded and imputed quantities. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A missing-data legend and documented imputation or filtering decision. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. A missing value is not automatically an observed absence. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 22 — Check metadata and data relations

Action. Verify sample attributes, SDRF columns, references and file names remain resolvable. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A passing relevant metadata validation where available and a manually checked sample mapping. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not permit translated factor values to create new categories accidentally. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 23 — Translate the human-facing text

Action. Draft natural sentences that preserve technical qualifiers, evidence level, causality limits and the source’s logical connections. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. Paired source-target segments with a reason recorded for important nonliteral choices. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Naturalness alone is not acceptance if the scientific meaning moves. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 24 — Run bilingual factual comparison

Action. Review each number, unit, sign, accession, condition, omission, addition and statement of uncertainty against source evidence. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A discrepancy log classifying each change by severity and resolution owner. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. A small stylistic improvement cannot justify an extra biological claim. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 25 — Run target-only comprehension

Action. Give an independent qualified target-language reader the translated report without the original and ask them what they believe happened. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A structured comprehension record about samples, acquisition, identifications and conclusions. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. A reader who infers a different experiment exposes a translation problem even when individual terms look correct. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 26 — Validate file structures separately

Action. If data files are part of the release, run appropriate schema, CV and metadata validators without treating them as language validators. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. Versioned validator output and a record of any waived warnings. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. A passing XML parser does not mean that a caption preserved the statistical meaning. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 27 — Review accessibility and presentation

Action. Check unit symbols, superscripts, tables, plots, abbreviations, reading order and alt text in the target publication. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A target-language visual review using actual rendered figures and accessible text. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not let formatting turn m/z or charged ion notation into an ambiguous expression. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 28 — Confirm repository submission language

Action. Read current PRIDE or other repository rules and distinguish file inventory, public access and the technical meaning of submission classifications. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A submission manifest and link to the current instructions used. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Never equate a repository’s administrative label with scientific confirmation. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 29 — Sign off with a scientific reviewer

Action. Resolve high-stakes discrepancies about identification, quantification, sample identity and statistical interpretation before release. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. An approval ledger recording accepted uncertainty and unresolved limitations. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Where expertise is missing, mark the content as not ready rather than creating false confidence. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

Gate 30 — Archive and monitor revisions

Action. Retain the released source and target, terminology decisions, validator outputs, DOI or accessions, and a maintenance owner. The point is to make the choice explicit before translating a sentence that depends on it. Assign the step to someone who can distinguish an editorial issue from an instrument, data or inference issue.

Required evidence. A release package that can be compared with later revised datasets or errata. Record the decision in a form another reviewer can inspect, rather than relying on a memory that the issue was checked. If the source uses a controlled vocabulary or repository schema, cite the relevant version or field definition.

Stop condition. Do not update target-only claims after a source revision without testing equivalence again. In a small project, this may mean one question to the source author. In a large archive, it may mean quarantining a subset of fields while safe paragraphs proceed. The rule prevents one doubtful term from becoming the template for hundreds of translated records.

The failure-mode atlas: 40 ways accurate-looking translations become wrong

A failure mode is not merely a mistranslated word. It is a predictable path by which a user, editor, parser or downstream analyst receives a different scientific assertion from the original one. Each example below describes a symptom, how the problem emerges, the specific diagnostic and the repair standard. Treat dangerous errors as release blockers even if the rest of the article is fluent.

Failure 01 — m/z becomes molecular weight

Symptom. An axis titled m/z is relabeled molecular weight, even though detected ions may carry multiple charges. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Check charge-state and isotopic-spacing evidence beside the axis. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore m/z and explain the quantity, without asserting neutral mass. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 02 — Ion charge becomes oxidation state

Symptom. The source’s z=3 is rendered with terminology meant for formal oxidation number. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Examine whether the source describes a gas-phase ion or a redox convention. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Use charge-state language consistent with the instrument measurement. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 03 — Retention time loses its unit

Symptom. An export in seconds appears in target prose as the same number of minutes. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Pick a peak and compare its chromatographic position in the original plot. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore units or explicitly document a correct conversion. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 04 — Intensity turns into concentration

Symptom. A relative peak height becomes a stated mass-per-volume concentration. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Look for calibration curves, internal standards and measurement units. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Use intensity or relative abundance unless concentration is supported. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 05 — Data processing disappears

Symptom. A centroided file is called untouched raw instrument data. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Read converter and processing metadata, including peak picking operations. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Distinguish raw acquisition from derived spectra and identify processing. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 06 — Polarity is omitted

Symptom. Positive and negative mode features are discussed as though identical ions were measured. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Check acquisition mode and proposed adduct formula for each row. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Retain polarity and ion-species context in captions. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 07 — Precursor and product ions swap

Symptom. A translated method reverses the ions selected and generated during MS/MS. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Follow the source acquisition event through precursor and fragment fields. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Correct both labels and their causal order. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 08 — Isolation becomes purification

Symptom. The target sentence says a unique peptide was purified when the instrument isolated an m/z window. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Inspect window width and the potential for coisolation. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Describe isolation accurately; do not imply chemical purity. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 09 — DIA is described as DDA

Symptom. A broad-window acquisition is translated as selected top-N precursor fragmentation. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Read the isolation schedule and reported DIA method. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore the correct acquisition strategy and its evidence limits. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 10 — The wrong scan is joined

Symptom. A spreadsheet editor tidies native identifiers and breaks the spectrum link. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Resolve the PSM’s scan reference against original data. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore the exact identifier and test the join. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 11 — A controlled accession is translated

Symptom. The display label is localized by changing a PSI-MS accession string. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Query the accession in the controlled vocabulary and inspect the source field. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Leave the accession fixed; place translated prose beside it. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 12 — Search tolerance loses ppm

Symptom. A 10 ppm precursor tolerance is presented as 10 Da. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Compare the method’s units and one mass-window example. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore the correct tolerance and explain the scale. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 13 — Observed and theoretical mass exchange places

Symptom. A report labels the theoretical sequence value as an experimentally measured value. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Trace the theoretical calculation and instrument-observed peak. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Use separate observed and calculated labels. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 14 — A PSM becomes a protein

Symptom. One matched peptide spectrum is translated as a definitive unique protein discovery. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Inspect whether peptides are shared across candidate proteins. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Keep PSM, peptide and protein inference as separate levels. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 15 — A shared peptide becomes unique

Symptom. A shared sequence’s supporting role is erased in a localized results table. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Map the sequence against all matched database proteins. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore shared-peptide language and group ambiguity. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 16 — A 1% FDR becomes 99% certainty

Symptom. The statistical group threshold is rendered as a claim about each individual hit. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Read the analysis-level error-control procedure. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. State FDR and the level to which it applies, with no invented individual probability. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 17 — A q-value becomes a p-value

Symptom. The column heading is translated into a familiar but different statistical label. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Verify the algorithm’s output field and documentation. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Keep the original statistic and its interpretation. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 18 — Score direction reverses

Symptom. A lower-is-better error score is translated as if higher were stronger evidence. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Compare accepted and rejected examples by native score. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Explain the correct direction and rerun result-order comprehension. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 19 — Threshold pass becomes experimental proof

Symptom. A software boolean is described as confirmation by an independent laboratory method. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Check whether there is any orthogonal evidence outside the algorithm. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Keep the threshold-pass decision at its correct analytical level. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 20 — Modification identity becomes site certainty

Symptom. An ambiguous phosphopeptide is assigned a single precise residue in the target caption. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Inspect localization probabilities and alternative sites. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore site ambiguity and conditional wording. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 21 — Variable modification becomes measured chemistry

Symptom. A search option is reported as a proven chemical state of every peptide. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Compare search parameters against supported matched spectra. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Distinguish hypothetical search space from observed support. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 22 — Protein group representative becomes the only member

Symptom. An accession chosen for display is translated as the sole identified protein. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Read the protein group membership and uniqueness evidence. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Name the ambiguity group or qualify the representative. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 23 — Isoform inference is overstated

Symptom. A gene-level or group-level finding is reported as a particular isoform. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Look for isoform-discriminating peptides in the dataset. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Use the narrowest supported protein or gene description. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 24 — Reporter channels are swapped

Symptom. Target-language figure labels accidentally reverse treatment and baseline tags. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Trace reagent channel to sample through the design table. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Correct mapping and recheck every ratio sign. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 25 — Relative quantity becomes absolute amount

Symptom. A normalized LFQ-like value gains ng/mL units during internationalization. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Inspect the calibration and scale definitions. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Keep relative, arbitrary or normalized units without invention. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 26 — Missing signal becomes zero

Symptom. An empty cell is rewritten as a measured absence. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Check missingness codes, imputation and detection limits. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Retain unknown, non-detect or below-limit status as applicable. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 27 — Technical repeats inflate sample count

Symptom. Three injections from one sample are described as three independent biological replicates. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Rebuild sample-to-run hierarchy. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Correct n, replicate descriptions and inference limits. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 28 — Control type is mistranslated

Symptom. A solvent blank becomes an untreated biological control in a chart. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Consult the acquisition sequence and sample sheet. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore blank, reference standard or biological-control meaning. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 29 — The wrong database release is supplied

Symptom. A translator fills in a plausible protein database version not present in the source. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Compare actual search files, fasta metadata and manuscript. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Remove invented version and seek source-owner confirmation. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 30 — Peak area becomes peak height

Symptom. An integrated chromatographic area is translated as the maximum point intensity. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Check feature integration software and axes. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Preserve the quantitative operation that generated the value. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 31 — DIA library provenance disappears

Symptom. A predicted library entry is rendered as a measured reference standard. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Identify library construction and predictive components. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Specify predicted or empirical references and their limits. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 32 — A dataset accession becomes validation

Symptom. Repository availability is summarized as official confirmation of all identifications. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Read repository guidance and status separately from scientific evidence. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Describe deposition as deposition, not certification. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 33 — Submission classification is moralized

Symptom. A repository technical label such as partial is translated as incomplete or poor science. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Check current repository submission-format policy. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. State the administrative meaning and avoid judging research quality. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 34 — Sample attribute changes category

Symptom. Inconsistent translations of a factor value create a new group in metadata. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Compare unique values and assay-row counts before and after. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Use approved controlled terms and stable file joins. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 35 — An amino-acid sequence is modified typographically

Symptom. Editorial hyphens or substitutions enter a sequence string or ProForma annotation. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Attempt parsing and compare the exact source bytes. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore the unchanged scientific notation and move explanation outside it. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 36 — Fragment labels are prettified

Symptom. An ion-series b/y index, charge or neutral-loss annotation is reformatted as prose. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Recalculate the assignment or inspect the software annotation. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Protect ion notation and translate only the explanatory caption. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 37 — Two data versions are combined

Symptom. A translated figure uses latest intensities with an older stated FDR procedure. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Compare file timestamps, versions and processing manifests. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Synchronize all claims to one approved release state. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 38 — QC pass becomes identification certainty

Symptom. The experiment met run-quality metrics and is described as proof each biological claim is correct. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Inspect the QC metric definition and identification validation stage. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Distinguish instrument or run fitness from inference confidence. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 39 — Uncertainty hedges disappear

Symptom. Words like suggests, consistent with or candidate are removed for a more confident sentence. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Compare causal and statistical commitments across source and target. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Restore source-supported uncertainty without unnecessary vagueness. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Failure 40 — Privacy grows through added detail

Symptom. A localized narrative adds personal or rare descriptors that were absent from approved metadata. This can happen during machine translation, table relabeling, manual paraphrase, automated export or the final editorial pass. The error may leave numbers and surrounding grammar apparently clean while changing what they describe.

Diagnose. Review combined identifying information at dataset level. The relevant comparison must follow the claim back to its source entity. A reviewer should record the precise field, axis, reference or decision whose meaning moved, instead of marking the sentence as merely awkward.

Repair and acceptance. Remove unapproved enrichment and follow access rules. Recheck all related figure labels, glossary entries, captions and summary statements for the same error. A corrected sentence is only fully repaired when the downstream reader will make the same scientific interpretation that the source supports.

Worked translation diagnostics: 60 fictional records that expose different failure modes

Every case in this section is an invented, simplified teaching record. The numbers, labels and sample identities are not a real experiment, instrument export, clinical finding or repository submission. Where chemistry or statistics requires calibration or a particular method, the case describes that requirement rather than pretending a single number proves a result. Each diagnostic includes a source fragment, a tempting but unsafe interpretation, a faithful target-language objective and the evidence that would settle the choice. Treat these as exercises in preserving meaning; they are not directions for overriding scientific software.

Diagnostic 01 — A two-charge peptide axis

Hypothetical source record. A hypothetical MS1 peak has m/z 601.31 and assigned charge z=2, with an accompanying spectrum. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The draft caption says the peptide has molecular mass 601.31 Da because it confuses the horizontal coordinate with neutral mass. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep “m/z 601.31, charge state 2+.” If neutral mass is required, calculate using the stated adduct convention and mass precision, not a dictionary substitution. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check the charge assignment and ion formula; the corrected target caption makes no unlicensed neutral-mass claim. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 02 — The isotope-spacing clue

Hypothetical source record. An illustrative isotope envelope contains adjacent isotopic peaks about 0.50 Th apart under an assigned charge state. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated paragraph calls the spacing a chromatographic retention-time interval or assumes z=1. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Explain that isotope spacing on the m/z axis can inform charge-state assignment; do not use spacing alone as molecular identity. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. The MS spectrum, not the chromatogram, must contain the spacing and the charge assignment used in the explanation. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 03 — A 6 ppm error

Hypothetical source record. A result reports measured m/z, theoretical m/z and a mass error of +6 ppm. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. An editor changes +6 ppm to 6 Da or deletes the plus sign as irrelevant punctuation. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Retain ppm and the signed comparison; if the prose does not specify the sign convention, refer back to the software rather than defining it by guess. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A table note must let the reader distinguish the size and direction of the deviation. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 04 — A chromatographic unit conversion

Hypothetical source record. The scan-start time is stored as 180 seconds; the human chromatogram axis is labeled in minutes. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated figure reports 180 minutes without converting the magnitude. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Write 180 seconds or 3 minutes, making the presentation conversion explicit and retaining the underlying metadata. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A reader should locate the same point on the original and target chromatogram after conversion. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 05 — Two ion adducts

Hypothetical source record. The same illustrative compound appears as a protonated ion in one analysis and a sodium adduct in another. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation assigns both observations the same bare molecular-mass label and erases adduct chemistry. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep the appropriate ion formulas and distinguish molecular mass from measured ion m/z. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check whether each proposed ion composition explains the coordinate, not merely whether the target label sounds familiar. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 06 — Negative ion mode

Hypothetical source record. One reference figure shows [M−H]− under negative mode, with a separate positive-mode result for another run. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The caption says only “protonated molecule” for both peaks. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Preserve deprotonation and polarity language as described by the source, and keep the two acquisition conditions distinct. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. The proposed label must agree with the ion formula and instrument polarity field. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 07 — An MS2 product spectrum

Hypothetical source record. A peptide has one precursor scan reference and a later fragment spectrum. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The editor calls the fragment spectrum the original molecular-weight measurement and ignores precursor selection. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. State that a selected precursor generated an MS/MS product-ion spectrum, with the source’s fragmentation method. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. The translated report must allow a reviewer to connect the precursor reference to the correct MS2 spectrum. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 08 — A broad isolation window

Hypothetical source record. An illustrative DIA window covers m/z 500–525 and contains signals from several possible precursors. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target text says a single purified peptide was fragmented within that range. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Explain the isolation interval and the possibility of cofragmented species; use the specific analysis result only as a candidate identification. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A method reviewer should reconstruct the same acquisition window, not a fictional single-molecule purification. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 09 — Normalized collision energy

Hypothetical source record. A methods table says normalized collision energy 28 under a specific instrument method. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target sentence adds “28 electron volts” without evidence that the numerical scale is eV. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Retain normalized collision-energy wording and instrument context; do not invent a physical unit for a vendor setting. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check the source method and software documentation before converting any energy representation. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 10 — Centroided versus profile data

Hypothetical source record. A converter exported centroided peaks from a vendor acquisition file, and the original profiles remain separately archived. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation describes the exported centroided array as the unprocessed raw signal. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Name the derived centroid representation and keep source file, converter version and processing steps traceable. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Compare the data-processing list with the translated word “raw”; remove it where unjustified. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 11 — A TIC and a BPC

Hypothetical source record. A report plots total ion current above a base-peak chromatogram for the same run. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translator gives both plots the identical target title “total signal” and implies equal quantities. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Define how each trace was computed and retain TIC/BPC distinctions even when their shapes look similar. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A reviewer should identify which plot sums contributions and which follows a maximum signal. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 12 — Chromatographic peak area

Hypothetical source record. The quantification output integrates signal over an elution peak and reports area in instrument response units. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target caption says “peak height, ng/mL,” introducing both a different operation and concentration units. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep integrated peak area and its scale; do not transform the output into absolute concentration without calibration. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. The reader should recover the integration method and units from the localized table. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 13 — Three traces in an XIC

Hypothetical source record. An extracted-ion chromatogram is shown for three precursors within specified extraction windows. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target figure calls the traces “three mass spectra” and reads the x-axis as m/z. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Use extracted-ion chromatogram language, explain the ion extraction windows and preserve retention time as the horizontal axis. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check that each translated line corresponds to the intended precursor and chromatographic interval. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 14 — A DDA top-N method

Hypothetical source record. Acquisition software selects certain precursor candidates according to intensity and exclusion rules. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target explanation says the instrument measured every peptide’s fragment spectrum in the sample. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Describe the data-dependent selection strategy and its sampling limits; do not infer complete coverage. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A method expert should be able to identify which precursor opportunities might have been missed. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 15 — DIA pseudo-spectra

Hypothetical source record. A DIA analysis generates a software-assembled fragment-ion representation from multiplexed acquisition data. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translator describes it as a raw MS2 spectrum from one physically isolated unique peptide. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep the distinction between acquired multiplexed scans and algorithmically extracted or assembled evidence. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Trace the candidate and its score through the DIA processing record, not only the displayed figure. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 16 — Transition monitoring

Hypothetical source record. A targeted method records precursor/product ion transitions for a known compound with designated channels. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target text uses “complete unknown-molecule scanning,” obscuring targeting. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Name the predefined transitions, identifiers and quantitative or qualifier purpose if documented. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A second analyst must be able to locate the exact transitions in the method file. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 17 — An ion mobility dimension

Hypothetical source record. A feature carries a retention time, m/z and a mobility-related coordinate. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target table translates mobility as chromatographic time and merges the two axes. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Maintain separate fields, units and physical processes; label a computed CCS only if the analysis actually produced one. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. The same feature should occupy the same three-dimensional coordinates after language conversion. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 18 — A predicted spectral library

Hypothetical source record. A library search used a set of predicted fragment intensities rather than an empirical physical standard for every candidate. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated conclusion calls each hit “confirmed by a measured reference spectrum.” A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Report predicted-library evidence and the subsequent scoring and threshold procedures. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Ask which reference entries were directly measured; uncertainty must survive when they were not. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 19 — A spectrum with a missing native ID

Hypothetical source record. An exported search result references scan=4217 in an original run, while a human figure uses a different sequential image number. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. An editor replaces scan=4217 with the figure number, breaking the evidence link. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Preserve the native ID exactly and keep the picture identifier as a separate editorial reference. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Resolve one PSM back to the original run in the source and translated manifests. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 20 — A controlled vocabulary name

Hypothetical source record. The source metadata uses a PSI-MS accession to declare spectrum type, plus a display name in English. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target translation overwrites the accession with a localized word and breaks machine interpretation. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Leave the accession untouched; supply the approved target-language explanation outside the encoded field. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Run a format-aware validator and a separate bilingual semantic review. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 21 — A peptide assignment, not a protein discovery

Hypothetical source record. An MS2 spectrum has an accepted peptide-spectrum match to peptide A; the peptide occurs in several proteins. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated abstract announces identification of a unique protein from that scan. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. State the supported peptide match and qualify the protein group where necessary. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check the protein inference output, not merely the presence of an accepted PSM. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 22 — Two indistinguishable isoforms

Hypothetical source record. The observed peptides are common to isoforms X-1 and X-2 and no isoform-specific sequence was detected. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The editor renders the top-listed isoform as the one experimentally proven. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Translate the ambiguity group and call the accession representative when that is its role. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Review presence or absence of sequence-discriminating evidence before asserting isoform specificity. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 23 — A unique-peptide boundary

Hypothetical source record. A peptide is unique in the searched reference set, but the paper does not claim uniqueness across every possible biological sequence. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target sentence says the peptide exists in no other protein anywhere. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. State uniqueness relative to the sequence database and relevant grouping convention. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check the exact scope and database release behind the unique label. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 24 — A razor peptide allocation

Hypothetical source record. A protein summarization tool allocates a shared peptide to one protein group by a documented heuristic. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target text states that the peptide came only from the selected protein. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Describe computational allocation rather than exclusive biological origin. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Inspect the quantification algorithm and other compatible protein groups. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 25 — The score-direction trap

Hypothetical source record. One illustrative search score ranks 0.002 as stronger evidence than 0.2 because it is an error-oriented score. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. A translator calls 0.2 “higher-confidence” based only on numerical size. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep the score’s name and its documented direction; avoid inventing percentage interpretations. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check examples of accepted and rejected PSMs under the actual software version. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 26 — FDR at one percent

Hypothetical source record. A dataset accepts PSMs at an estimated 1% FDR, using a declared target-decoy workflow. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target language claims each PSM has 99% personal certainty, which does not follow from the set-level procedure. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. State “PSMs were filtered at an estimated 1% FDR” and keep the analysis level explicit. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. The target must not use universal individual-probability language unsupported by the source. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 27 — A protein-level threshold

Hypothetical source record. The analysis has separate PSM-level and protein-level error-control procedures with distinct thresholds. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation mentions only a single “global 1% accuracy” condition. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Retain both levels and their thresholds, including any protein grouping rules. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A reviewer should identify which table is filtered under which threshold. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 28 — Posterior error probability

Hypothetical source record. The source reports an illustrative PEP value for one match and a separate q-value for ranked results. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. A target heading merges both into one “confidence percentage” column. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep PEP and q-value distinct, with the source method’s definition and score direction. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Use the processing report to confirm neither number has been mistaken for a simple probability of biological truth. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 29 — A threshold pass flag

Hypothetical source record. An mzIdentML item has passThreshold=true and points to a protocol threshold statement. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated row reads “experimentally verified protein” solely because the Boolean is true. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Describe it as passing the configured identification threshold, not as orthogonal laboratory confirmation. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Inspect the threshold element and whether an independent verification exists. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 30 — A decoy sequence hit

Hypothetical source record. A search uses synthetic or reversed decoy sequences to estimate false matches. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target-language method says researchers injected a decoy substance as a negative control. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Explain computational decoys in the search database; keep physical blank and biological control language separate. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. The source data inventory should show whether decoys are database entries rather than experimental injections. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 31 — A database contaminant

Hypothetical source record. The search includes a common contaminant list and some peptides are classified accordingly. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation reports every identified peptide as originating from the intended biological sample. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Retain contamination flags and their analytical role; avoid implying every matched protein is endogenous. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check sample preparation, contaminant database entries and reporting filters before interpreting findings. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 32 — Trypsin missed cleavages

Hypothetical source record. An enzymatic digestion method used trypsin but the database search allowed two missed cleavage sites. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated sentence says the sample was fully cleaved with zero missed cleavages. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Distinguish experimental digestion procedure from search allowance for incompletely cleaved peptides. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Compare source methods with the search protocol and avoid inventing digestion efficiency. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 33 — Fixed carbamidomethylation

Hypothetical source record. A search model applies carbamidomethylation as a fixed modification for eligible cysteine residues. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target claims every molecule’s modification was separately verified by direct site evidence. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Call it a fixed search modification and distinguish the alkylation step from individual observation. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check eligibility, preparation chemistry and search settings. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 34 — Oxidation as a variable candidate

Hypothetical source record. Methionine oxidation was allowed as a variable modification in the search space. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated results claim that oxidation was detected in every peptide examined. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. State that it was considered in the search, and only call it assigned where identification evidence supports it. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Inspect which accepted candidates actually carry the annotation. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 35 — An ambiguous phosphate location

Hypothetical source record. A peptide contains two plausible phosphorylatable residues and a source score does not uniquely assign the site. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target-language figure places the phosphate exclusively at one residue and removes “potentially.” A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep candidate sites and the localization limit explicit; use an uncertainty-aware caption. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check site-determining fragment evidence and the reported localization metric. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 36 — An unassigned mass shift

Hypothetical source record. An open search reports an unexplained mass difference of approximately +80 Da on a peptide. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target manuscript identifies a specific phosphate site without any supporting site or modification assignment. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Report an observed mass shift and the hypothesis status, not a confirmed chemical identity. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Seek supporting evidence before converting a nominal mass difference into a site-specific modification. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 37 — A ProForma syntax marker

Hypothetical source record. A modified sequence uses structured modification brackets and accession tags. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The editorial process replaces ASCII brackets with typographic punctuation inside the sequence. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Protect the notation as immutable code and translate the explanation beside it. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Run a parser or exact-string comparison to detect subtle markup changes. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 38 — Reporter-channel mapping

Hypothetical source record. A multiplexed experiment assigns channel A to treatment and B to control in its design. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The figure is localized with A marked control and B marked treatment, reversing the ratio story. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Restore channel-to-condition mapping and the order of the reported fold change. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Compare the manifest to all target-language legends, labels and conclusions. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 39 — Ratio compression caveat

Hypothetical source record. An isobaric-label analysis includes an interference-correction step and notes possible ratio compression. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation presents all ratios as exact intrinsic biological differences without method limitations. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Preserve the stated measurement caveat, corrections and comparison basis. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Review the source uncertainty note alongside reported ratios before changing evaluative language. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 40 — Label-free output units

Hypothetical source record. An LFQ table reports normalized arbitrary intensity values, not calibrated concentration. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated caption gives ng/mL because the target journal favors physical concentration units. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep the quantitative scale and its method name, even if the unit seems less intuitive. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Require calibration evidence before using physical amount or concentration language. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 41 — A protein rollup

Hypothetical source record. A result is a protein-level summary of several peptide or precursor features after normalization. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target says the number is an directly measured instrument signal from the intact protein. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Explain the aggregation level and underlying measurement chain. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Inspect which features entered the rollup and whether any shared peptides affect inference. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 42 — Below the quantification limit

Hypothetical source record. A feature is detected but is below the declared LOQ; it is not reliably quantified. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated table says “absent” or inserts an exact zero concentration. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep detected-below-LOQ distinct from non-detection and missingness. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check the laboratory threshold definition and ensure the target statement does not erase low-level evidence. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 43 — Imputed sample value

Hypothetical source record. An analysis pipeline imputes missing intensities before statistical testing. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation describes every displayed number as directly measured. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Identify imputed fields and their stated method, while preserving which rows are observations. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Compare the raw and processed matrix plus imputation log. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 44 — Two injections from one specimen

Hypothetical source record. One biological specimen is measured twice as technical injections, while two other independent specimens are also in the study. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated methods call four independent biological replicates. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. State the actual biological n and which measurements are technical repeats. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Reconstruct the experimental design table and examine independence of observations. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 45 — A solvent blank

Hypothetical source record. A blank injection was used to assess carryover after a highly concentrated sample. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The caption translates it as a negative biological control with no treatment. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Name the blank’s purpose and location in injection order, not an imagined biological condition. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check whether any unexpected peak indicates residual carryover rather than sample biology. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 46 — A conversion manifest

Hypothetical source record. RawA.raw and RawA.mzML are versions of the same run, with a documented conversion step. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation counts them as two independent experiments and cites separate sample numbers. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Retain parent–derivative relationships and name the converter and settings. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A manifest must link both files to one run and sample. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 47 — An SDRF factor value

Hypothetical source record. An SDRF file uses a stable experimental factor value across three assays. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. A translator uses two synonymous target phrases, creating separate categories in the localized table. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep the machine-facing factor value consistent, using a separate display translation where appropriate. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check unique factor categories and required column values with validator-supported procedures. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 48 — A repository complete label

Hypothetical source record. A project includes appropriate standard result files and receives a repository-specific compliance classification. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation tells readers the science was proven complete and true by repository approval. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Explain the technical meaning of the submission classification and list the file types it covers. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Read the current PRIDE guidance, not a generic dictionary meaning of “complete.” A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 49 — A PXD identifier

Hypothetical source record. A paper refers to a ProteomeXchange project accession and links it to a data release. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target text translates the identifier letters or formats it as a local publication number. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Preserve exact accession and link; ensure it points to the intended dataset. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Resolve the accession and compare version and file inventory against the paper. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 50 — A later reanalysis

Hypothetical source record. An updated search pipeline yields a revised protein group count for the same archived raw runs. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translated abstract retains the earlier count while the latest table is inserted. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Choose one declared analysis version or explicitly compare editions, with dated processing details. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check that tables, narratives and methods refer to the same computation. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 51 — An instrument QC pass

Hypothetical source record. The run satisfies retention-time stability and other quality metrics under the method’s criteria. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target states every peptide identification is correct because the instrument passed QC. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Distinguish run fitness from identification FDR and biological confirmation. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Compare what the actual QC metrics assess with the claim. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 52 — A high-confidence claim in the abstract

Hypothetical source record. The source abstract says data are consistent with increased pathway activity, subject to further verification. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translation says the experiment proves activation of the pathway and its cause. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Retain associated-with or consistent-with uncertainty and the study’s evidentiary limits. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Read the abstract alongside the statistical analysis and validation experiments. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 53 — A protected donor record

Hypothetical source record. An international study shares a limited sample metadata description with no detailed donor identifiers. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target-language article adds rare patient traits to make the case seem vivid. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Translate only approved metadata; do not enrich with identifiers or infer demographic attributes. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Review the full localized dataset for combinations that might increase reidentification risk. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 54 — A target-language m/z glossary

Hypothetical source record. A multilingual guide provides a local expression for mass-to-charge ratio and retains the symbol m/z. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The editor treats the translated expansion as permission to relabel every numeric m/z field. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Place explanations in tooltips or glossary prose while preserving schema names and numeric arrays. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Check machine-readability and comprehension with separate tests. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 55 — A methods note about fragmentation

Hypothetical source record. The source distinguishes HCD from ETD in describing different fragmentation runs. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target uses a single generic collision-fragmentation phrase for both methods. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Preserve the distinct method names and, when needed, explain their role rather than their full underlying physics. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Inspect acquisition metadata to ensure each run receives its method. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 56 — A sample with isotope label swap

Hypothetical source record. Two replicate sets use reversed light/heavy labels to counter labeling confounds. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target summary assumes light always means control and heavy always means treatment. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Track label-to-condition mapping per run, including swaps. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Verify every summarized ratio uses the intended direction after matching the design. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 57 — A cross-linked peptide

Hypothetical source record. One spectrum supports a candidate pair of cross-linked peptide chains with ambiguous site alternatives. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target declares an exact three-dimensional protein contact as experimentally proven. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Report the cross-link candidate and residue constraints, not a unique structure or mechanism unless additional evidence exists. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Review specialized scoring, alternative assignments and independent structural validation. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 58 — An intact proteoform study

Hypothetical source record. The instrument measures intact protein ions in a top-down workflow and reports candidate proteoforms. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translator rewrites the methods as trypsinized bottom-up peptides because that language is more familiar. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Preserve molecular input scale, fragmentation method and proteoform-level interpretation. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Confirm the specimen entered the instrument as an intact protein species rather than a peptide digest. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 59 — A metabolomics table boundary

Hypothetical source record. A tabular result uses mzTab-M for small-molecule feature and identification reporting. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. A translated methods page calls it peptide mzIdentML and explains every row as protein evidence. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Name mzTab-M and its intended molecular domain, preserving separate feature and identification concepts. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Cross-check the file specification and sample analytical method. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 60 — A mobile-phase blank and solvent label

Hypothetical source record. A chromatography method reports solvent A and solvent B composition, and uses a blank run as a control. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The translator mistakes “B” for a biological group and confuses blank with sample B. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Keep solvents, blank injection and biological group identifiers separate. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. Recreate the gradient and run table from the translated methods; they should agree with the source. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Diagnostic 61 — A native-score software update

Hypothetical source record. The same search engine reports a similarly named score under different versions with changed processing details. The case deliberately provides less information than a full laboratory dataset: missing fields must not be supplied by guesswork.

Tempting mistranslation. The target handbook copies the older score definition into a report built with the newer release. A bilingual dictionary could produce all the individual words in this rendering, yet the scientific message would move. The reviewer should state which claim, measurement or evidence level was changed.

Faithful target-language strategy. Check software version and current field documentation, preserving the exact reported name. This wording goal can be implemented in the target language’s natural grammar without modifying scientific identifiers, numbers or inferred relationships.

Independent diagnostic. A reviewer should be able to interpret score direction and scope using the proper release. A passing case leaves the target reader able to reconstruct the supported scientific event and its limitations. If the original evidence is unavailable, mark the statement unverified instead of certifying an invented interpretation.

Transfer workshops: use the method beyond a single spectrum

The point of a technical translation system is transfer. A reader who knows one mzML field but cannot apply the distinction to a manuscript, a table, a repository manifest or a cross-language meeting has not yet learned a dependable method. Each workshop changes the document setting while preserving the same discipline: source entity, units, provenance, uncertainty, audience and a test for whether the target language would support the original inference.

Transfer workshop 01 — A multilingual journal methods section

The assignment. Take three source paragraphs that describe sample preparation, instrument settings and database searching. Mark all experimental actions, software products, fixed parameters and conditional verbs. Translate the prose for clarity while leaving the sample-to-data chain intact. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Ask another reader to reconstruct the procedural order from target text alone, then compare with the actual protocol. In particular, inspect whether digestion and database search conditions were confused, whether the instrument model survived, and whether any source omission was silently repaired. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. A publishable result has an accountable list of unresolved method details. Linguistic smoothing is welcome only when it preserves what the authors actually did. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 02 — A spectrum figure and its legend

The assignment. Select a hypothetical m/z–intensity plot with a labeled precursor, fragments and an isotope pattern. Prepare target-language axis descriptions, a legend and alt text. Keep m/z, ion series and numerical values intact. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Give a reviewer the localized caption but not the source. Ask which ions were observed, how many charges were assigned and what claim the annotations support. Compare answers with the source spectrum metadata. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. If the legend claims a unique peptide identity from peaks alone, downgrade the language to the evidence actually provided. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 03 — A chromatography-panel localization

The assignment. Use an XIC plot containing several traces, time axis in minutes, an integration window and a peak-area table. Separate visual labels from machine files and calculation results. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Compare each trace’s extraction definition and the reported integrated area with the localized caption. Verify that retention time is not called m/z and that signal is not represented as a concentration without calibration. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. A clean visual translation enables readers to associate the same peak with the same sample and experimental operation. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 04 — A cross-language laboratory handoff

The assignment. Imagine two teams exchange original raw files, converted mzML files and search results. Produce a handoff note that defines exactly which files are source records, derivatives and final summaries. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Ask the receiving team to locate a candidate PSM in the correct original run using only the documentation. Any ambiguous filename, translated accession or missing converter version is a handoff defect. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The handoff is successful when the target-language reader can retrace rather than merely admire the experiment. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 05 — A controlled-vocabulary glossary

The assignment. Choose a set of PSI-MS entries for MS level, polarity, instrument model, spectral representation and unit-bearing parameters. Add target-language explanations in a parallel table. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Check every accession against the official term, its value type and unit where applicable. Put translation options in a human-readable layer; do not rewrite machine keys or imply a synonym is the same accession. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. This builds a glossary that supports both people and software, rather than a bilingual list that silently changes ontology identity. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 06 — A repository project summary

The assignment. Write a concise multilingual project description that links raw data, analysis outputs, publication and any standardized results files. Keep the accession and file manifest aligned with the repository. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Compare public availability statements with current submission guidance. Ensure the summary does not infer that a technical repository classification certifies the biological conclusions. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. Readers should know what can be accessed, what was deposited and what is still uncertain without needing the source-language abstract. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 07 — A database-search reproducibility note

The assignment. Localize a paragraph naming a protein sequence database, enzyme, missed cleavages, modifications, search tolerances and analysis software. Keep all numbers and settings within the original protocol. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Try to set up an equivalent hypothetical search using the localized note. If multiple interpretations of enzyme specificity or modification type remain, note the uncertainty rather than selecting defaults. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. Reproducibility is a comprehension test: the target language must recover the same computational search assumptions. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 08 — An FDR and q-value training card

The assignment. Create a target-language explainer with separate definitions for set-level FDR, per-item PEP where reported, a q-value and a threshold decision. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Test a few hypothetical accepted and rejected PSMs. Ask learners which quantity is a probability of an individual error under a model and which governs a set or ranked results. Do not provide numerical certainty not supported by the analysis. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The training succeeds when a reader stops converting “1% FDR” automatically into “99% certain for this protein.” Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 09 — A protein-group results table

The assignment. Present a fictional peptide list shared by two isoforms, with one group representative. Localize column headings and a results summary. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Ask the target-language reader whether the experiment uniquely identified either isoform. If the answer is yes but the source has no discriminating peptides, the translation overstates evidence. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The finished table shows the group structure and inference boundary explicitly without replacing it with a convenient gene name. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 10 — An ambiguous phosphorylation site

The assignment. Use a hypothetical phosphopeptide with two possible positions and a localization score that does not resolve them. Translate the caption without obscuring the uncertain residue. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Have a qualified reader annotate the positions that remain plausible. The correct answer is not necessarily the first residue mentioned by the original author. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. A faithful translation can be clear and uncertain simultaneously. Uncertainty is part of meaning, not a failure of writing. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 11 — A quantitative fold-change statement

The assignment. Imagine a report compares normalized protein-level values between two groups, with confidence intervals and filtering criteria. Localize the results paragraph. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Check the numerator/denominator, transformed scale, direction of change and significance language. Ensure the ratio is not inverted when the target language changes sentence order. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The independent target reader should recover the same comparison and the same statistical limits. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 12 — A TMT-style channel legend

The assignment. Translate a multi-channel experimental legend with sample assignments, normalizations and an interference note. Keep channel names consistent across every caption and table. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Verify one treatment-to-control ratio against the sample manifest and correct any swapped labels. A visual or prose translation that reverses experimental groups is a scientific error. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. A working bilingual legend should let readers locate each labeled sample without consulting the original author’s private notes. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 13 — A label-free quantification report

The assignment. Take a matrix of normalized intensities and an explanation of how missing features were handled. Translate the report for a non-specialist research collaborator. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Ask whether the target-language reader believes the results are absolute concentrations, raw detector readings or processed relative quantities. Only the original quantity level is acceptable. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. Explain normalization and missingness in accessible prose without pretending the numerical values are intrinsically comparable under every condition. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 14 — A data-independent acquisition comparison

The assignment. Prepare an explanation comparing DDA and DIA with common terms such as precursor, isolation, fragment evidence, search and library. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Check whether the target reader thinks every DIA fragment spectrum came from a single isolated molecular species. Correct that implication using method-specific explanation. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The aim is an accurate conceptual contrast, not a generic definition of “advanced mass spectrometry” that erases analytical architecture. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 15 — A clinical research methods excerpt

The assignment. Use a deidentified proteomics study excerpt with sample categories, instrument methods and a qualified protein-expression finding. Keep the clinical inference intentionally narrow. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Review whether sample descriptors are approved, biological replication is correct, and association has been promoted into clinical diagnosis or causation. Scientific translation is not permission to give individual medical advice. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The translated article must remain a research report, not a patient-level diagnostic promise. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 16 — An environmental metaproteomics abstract

The assignment. Imagine a complex community sample where multiple proteins share peptides across taxa and the database is incomplete. Draft a multilingual summary of observed peptides and tentative assignments. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Test whether the translation invents species-specific source attribution from shared sequences or describes functional annotation as directly observed biological activity. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. Keep taxonomic, functional and protein inference uncertainties distinct; this is especially important when results are used in cross-disciplinary discussions. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 17 — A targeted small-molecule assay

The assignment. Localize a report containing multiple reaction monitoring transitions, calibration, limits of quantification and an internal standard. This is not automatically bottom-up peptide identification. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Check whether each translated unit, transition, calibration result and qualifier remains linked to the intended analyte. Missing a qualifier transition changes what the method claims. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The exercise demonstrates how a mass spectrometry translation method transfers beyond protein identification without importing inappropriate proteomics assumptions. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 18 — A metabolomics results file

The assignment. Read an mzTab-M result summary in a small-molecule context and identify feature, identification and abundance fields. Translate descriptions without changing the tabular schema. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Ask whether the target makes a feature-level signal sound like a confirmed named compound, or misrepresents a lipid as a peptide. Track adduct and identification confidence language. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The result is an interoperable explanation that respects the format’s own molecular domain. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 19 — An ion mobility dataset

The assignment. Translate a figure caption that includes drift time or mobility, m/z, chromatographic retention and optionally collision cross section. Do not flatten three axes into one. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Have a specialist verify the source units and any calibration behind reported CCS. If the target treats mobility as chromatography, the axes and conclusions will be wrong. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The translation should convey why each dimension separates ions and which quantities were measured or inferred. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 20 — A high-resolution isotope profile

The assignment. Use a hypothetical isotope envelope with an assigned charge and a calculated theoretical pattern. Draft a caption distinguishing observation from model fit. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Test whether the translated description calls the predicted theoretical envelope directly observed, or mislabels isotope spacing as mass error. Compare the source and predicted traces. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. A scientifically safe caption can describe agreement without implying chemical identity is unique in the absence of supporting evidence. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 21 — A cross-linking proteomics report

The assignment. Translate a methods-and-results pair about linked peptides, candidate site pairs and spatial constraints. Keep terms such as cross-link, site, score and restraint distinct. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Ask whether the target implies an exact protein structure was demonstrated when the source supported only a proximity-related constraint under assumptions. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The transfer skill is preserving the strength of structural inference and keeping the candidate alternatives visible. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 22 — An intact-protein analysis

The assignment. Localize a top-down proteomics figure with intact charge envelopes, deconvolution and a proposed modified proteoform. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Check that the target language does not introduce tryptic digestion, peptide-level PSM assumptions or a unique site claim absent from the source. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. Readers should learn which molecular entity was directly measured and which proteoform properties were inferred by processing. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 23 — A FAIR data-management package

The assignment. Build a translation-ready folder with a data manifest, machine-immutable files, multilingual README, glossary, source citations and versioned QC reports. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Give the package to another reviewer and ask them to find one sample, raw file, derivative spectrum, identification and corresponding published statement. Record failed links and missing provenance. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The goal is findable, accessible under stated terms, interoperable and reusable research evidence; translated documentation should improve those properties without altering the data. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

Transfer workshop 24 — A machine-translation quality audit

The assignment. Feed a harmless invented methods excerpt to an automated translator and compare the output with a human review of ions, units, exclusions, confidence and experimental roles. Before drafting, separate what the source explicitly states from what the reviewer would merely expect to be true of a typical experiment.

The falsification test. Use an error classification with separate penalties for style and scientific harm. A polished phrase with wrong sample identity deserves much more attention than a minor awkward collocation. A useful test tries to expose an incorrect inference rather than merely checking that all familiar nouns appear somewhere in the page.

What transfers. The improvement loop stores specific bilingual counterexamples in a termbase, then retests on unseen material rather than congratulating itself for fluency. Apply the same discipline to a new document and record any term that required an expert decision. The goal is future competence with less guesswork, not a single attractive translation.

A working terminology glossary for international readers

The entries below are concise distinctions for language transfer, not substitutes for full schema definitions or instrument documentation. When an identifier has an official definition, that definition controls the encoded field. The wording here helps a human reader keep nearby concepts apart.

01. m/z. Mass-to-charge coordinate for detected ions; do not equate directly with neutral molecular mass. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

02. Thomson (Th). A conventional unit used to express mass-to-charge ratios; preserve the scientific context. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

03. Da. Dalton, a unit of mass; not interchangeable with ppm or a retention-time unit. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

04. ppm. Parts per million; in mass error contexts, a relative deviation with a specified reference. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

05. Precursor ion. Ion selected or considered for subsequent fragmentation or analysis. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

06. Product ion. Ion generated from precursor fragmentation, supporting structural interpretation. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

07. Charge state. Assigned net ion charge, commonly indicated by z or a superscript sign. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

08. Adduct. Ion form that includes an added or associated ion or group relative to a neutral species. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

09. Monoisotopic mass. Mass calculated from a specified isotope composition. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

10. Isotope envelope. Distribution of isotope-related peaks, with spacing influenced by charge. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

11. Centroid spectrum. Processed spectrum represented by discrete peak positions and intensities. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

12. Profile spectrum. Spectrum retaining sampled peak shapes rather than only centroided peaks. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

13. MS1. A mass spectrometry stage commonly used to observe precursors or survey ions. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

14. MS2. A tandem mass spectrometry stage commonly containing product-ion information. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

15. MSn. Mass spectrometry experiment with multiple stages of ion selection and fragmentation. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

16. Polarity. Positive or negative ion acquisition mode; not a quality label. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

17. Ion source. Instrument component or method responsible for generating ions. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

18. ESI. Electrospray ionization, an ion-generation technique often used with LC-MS. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

19. MALDI. Matrix-assisted laser desorption/ionization, a distinct ionization approach. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

20. Collision energy. Setting or physical quantity related to activation or fragmentation under a method. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

21. HCD. Higher-energy collisional dissociation, a named fragmentation method. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

22. ETD. Electron-transfer dissociation, a distinct fragmentation method. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

23. Isolation window. m/z interval selected for an acquisition event; it may include multiple species. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

24. DDA. Data-dependent acquisition, where selection responds to observed signal or rules. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

25. DIA. Data-independent acquisition, typically covering predefined ranges for fragmentation evidence. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

26. SRM. Selected reaction monitoring of planned precursor/product transitions. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

27. MRM. Multiple reaction monitoring, often used for targeted quantitative analysis. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

28. TIC. Total ion current, a scan-related aggregate signal represented across time. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

29. BPC. Base peak chromatogram, following the strongest peak under a stated definition. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

30. XIC. Extracted ion chromatogram, tracing selected ion signal across chromatographic time. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

31. Retention time. Chromatographic elapsed time; distinguish from biological time points. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

32. Peak area. Integrated signal under a defined chromatographic peak or feature. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

33. Peak height. A maximum or selected signal amplitude, not the same as integrated area. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

34. Signal intensity. Detector or processed response; not automatically concentration. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

35. Resolving power. Instrument or method property describing separation of close mass signals. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

36. Mass accuracy. Agreement of measured and reference or theoretical mass under a defined comparison. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

37. Ion mobility. A separation dimension related to ion motion under electric fields and gas conditions. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

38. CCS. Collisional cross section, usually a derived physical descriptor with method dependence. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

39. mzML. Community XML representation for mass spectrometric spectra and metadata. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

40. mzIdentML. Community format for peptide/protein identification evidence and relationships. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

41. PSI-MS. Controlled vocabulary supporting standardized mass spectrometry terminology. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

42. CV accession. Stable controlled-vocabulary identifier; machine key should not be translated. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

43. PSM. Peptide-spectrum match: an analytical association between a spectrum and peptide candidate. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

44. Peptide. Amino-acid sequence entity, not automatically a uniquely inferred protein. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

45. Protein inference. Reasoning from peptide evidence to protein candidates or groups. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

46. Protein group. Set of compatible protein assignments under an analysis procedure. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

47. Representative accession. Chosen display member of a group, not automatically a uniquely proven member. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

48. Unique peptide. Sequence considered unique within the stated database or inference context. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

49. Shared peptide. Peptide sequence compatible with more than one protein candidate. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

50. Razor peptide. Shared peptide assigned under a defined protein summarization rule. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

51. Isoform. Sequence variant of a protein product; often difficult to distinguish by shared peptides. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

52. Proteoform. Specific molecular form of a protein, including sequence and modifications. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

53. ProForma. Structured notation for modified peptide and protein sequence representation. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

54. PTM. Post-translational modification; site identity and localization can remain uncertain. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

55. Fixed modification. Search setting applying a modification to eligible sites under the computational model. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

56. Variable modification. Modification considered among search candidates, not automatically experimentally confirmed. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

57. Site localization. Inference about the residue position of a modification. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

58. Neutral loss. Loss of an uncharged species during a fragmentation interpretation. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

59. Fragment-ion series. Structured annotations such as b and y ions used to interpret spectra. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

60. Search database. Sequence or spectral reference collection used for candidate matching. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

61. Enzyme specificity. Cleavage constraints in the sample preparation or search analysis. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

62. Missed cleavage. Peptide containing an uncleaved expected protease site under a model. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

63. Mass tolerance. Allowed difference between observed and candidate masses, with stated units. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

64. Target-decoy method. Procedure using decoy matches to estimate false discovery rates under assumptions. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

65. Decoy. Search/reference construct for error estimation, not necessarily a physical negative control. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

66. FDR. Expected false-discovery proportion across a selected set under an estimation procedure. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

67. q-value. Multiple-testing metric linked to an FDR threshold under a stated analysis method. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

68. PEP. Posterior error probability for an individual candidate under a calibrated model. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

69. passThreshold. Reported Boolean decision under an identification protocol, not proof of absolute truth. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

70. LFQ. Label-free quantification; the associated output may be normalized relative intensity. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

71. Reporter ion. Fragment ion often used to distinguish labeled samples in multiplexed workflows. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

72. TMT. Tandem mass tag approach for multiplexed relative quantification. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

73. Internal standard. Reference material introduced or used for calibration or quality assessment. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

74. LOD. Limit of detection under a stated analytical definition. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

75. LOQ. Limit of quantification under a stated analytical definition. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

76. Missing value. Unreported data entry; its reason may be unknown, filtered, censored or unavailable. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

77. Imputation. Replacement of a missing entry by a model-based estimate under assumptions. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

78. Biological replicate. Independent biological unit under the experimental design. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

79. Technical replicate. Repeated processing or measurement rather than a new biological unit. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

80. SDRF-Proteomics. Structured sample and data relationship metadata representation for proteomics. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

81. PRIDE. EMBL-EBI proteomics archive providing dataset records and associated files. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

82. ProteomeXchange. Consortium framework supporting dissemination of proteomics datasets. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

83. PXD. Project accession prefix in ProteomeXchange; not a synonym for validation. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

84. mzTab-M. Tabular reporting format for mass spectrometry-based metabolomics results. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

85. mzQC. Structured quality-control metrics format in the proteomics standards ecosystem. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

86. mzSpecLib. Developing standardized representation for spectral library information; verify its maturity. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

87. Checksum. Digest used to detect byte-level changes in files when calculated and compared properly. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

88. Provenance. Documented origin and processing history of a result or data artifact. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

89. Data lineage. Traceable parent–derivative relationships among samples, acquisitions, files and outputs. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

90. Semantic fidelity. Preservation of the meaning and evidentiary scope of source statements across languages. A translator should also check its actual schema field, measurement context or protocol usage before committing it to a bilingual termbase.

Frequently asked questions about translating mass spectrometry and proteomics evidence

1. Is mass spectrometry data translation the same as translating a chemistry textbook?

No. A textbook usually explains established concepts in prose, while a mass spectrometry release also includes instrument output, encoded metadata, identifications, statistical decisions and relationships between files. A fluent translation can still be dangerous if it changes those relationships. Start by deciding which elements are language content and which are scientific evidence.

2. Should I translate mzML files themselves?

Do not replace machine-readable keys, accession identifiers, binary arrays or numbers with localized text. Human-facing descriptions and documentation can be translated in a separate layer or by an explicitly supported localization feature. If a format conversion is required for a technical reason, use suitable scientific software and preserve a traceable source copy.

3. What is the simplest explanation of m/z?

It is an ion’s mass-to-charge coordinate. The same neutral compound can produce more than one ion form and therefore more than one measured coordinate. Keeping m/z as a symbol and explaining it in the target language is normally safer than substituting an everyday phrase such as molecular weight.

4. Why does charge state matter in a translation?

Because charge changes the relationship between the observed coordinate and the molecule. A multiply charged peptide is not interpreted the same way as a singly charged ion. Even a visually small omission of z or a sign can change a mass assignment. Ask whether the source supports the stated charge before rewriting the sentence.

5. Can I translate instrument abbreviations such as MS2 or HCD?

You can explain them in the target language, but keep standardized abbreviations or exact method labels where they are identifiers in reports or metadata. Expansion should not erase differences between acquisition stages or fragmentation methods. If instrument-specific terminology differs, the source documentation should settle the choice.

6. What is the difference between a mass spectrum and a chromatogram?

A mass spectrum represents ion information as a function of mass-to-charge, while a chromatogram represents a separation-related signal across time or another separation axis. Extracted-ion chromatograms track selected ion signals through the chromatographic dimension. The plots answer different questions, so translating both as the same graph is not adequate.

7. Does a match prove that the protein exists?

Not by itself. A PSM links a spectrum and candidate peptide under an analysis; peptides can be shared by proteins, and protein inference may retain ambiguity. Report the actual level of identification and the filtering criteria. Strong biological conclusions may require additional evidence.

8. What does one percent FDR really mean?

Under the stated estimation procedure, it controls or estimates the false-discovery proportion among a reported set. It should not automatically be described as 99% certainty for each individual peptide or protein. The analysis level—PSM, peptide, protein or site—must appear in a careful translation.

9. Why do q-values and PEP values not mean the same thing?

They are produced for different statistical interpretations. A PEP, when calibrated, concerns a candidate’s modeled chance of being wrong, while a q-value is tied to multiple testing or error-rate thresholds under an analysis procedure. Translating both as generic confidence removes critical information.

10. What if a source manuscript is scientifically inconsistent?

Do not silently invent a correction. Record the conflicting statements, identify the authoritative data or method field, and refer the issue to the scientific owner. The translation should carry only confirmed corrections and may need a documented note explaining unresolved ambiguity.

11. Can I translate amino-acid sequence annotations?

Translate descriptions and headings around them, not the machine-readable sequence or structured modification notation. Even typography changes can alter a parseable sequence representation. Verify exact characters in accession references, ProForma annotations and other structured terms.

12. How should an uncertain phosphorylation site be described?

Use wording that allows the same alternative residues or localization uncertainty as the source. The target language can make an explanation easier to read without choosing a residue the evidence does not distinguish. A site-localization score is not a license for unwarranted certainty.

13. What should I do with proprietary vendor raw files?

Preserve the originals and record their relation to converted representations. Translate methods and metadata descriptions rather than modifying proprietary binary content. If conversion to mzML or another format is performed, retain the tool, version, settings, file mapping and validation evidence.

14. Is mzIdentML equivalent to mzML?

No. mzML primarily represents mass spectrometric acquisition data and associated metadata, while mzIdentML represents downstream peptide and protein identification evidence with references to spectra and analysis protocols. A translation should preserve that source-to-result distinction.

15. What is the purpose of PSI-MS?

The controlled vocabulary provides standardized concept identifiers and definitions for mass spectrometry and associated informatics. It improves interoperability across software and records. Translate explanatory labels where useful, but preserve accession strings and check the original term’s scope, units and version.

16. What happens when a peptide belongs to multiple proteins?

The inference may identify a protein group or report representative accessions rather than a single proven member. Do not pick the best-known gene name for the target language simply because it is easier to recognize. A meaningful translation retains the group or explains its ambiguity.

17. Can I translate a relative intensity into a concentration?

Not without an appropriate calibrated method and an authorized, documented transformation. Relative intensity is a signal-related quantity; a concentration has physical units and may depend on standards, recovery and matrix effects. Changing the label without the evidence changes the scientific claim.

18. What is the safest way to handle missing values?

Find out whether an entry means missing, filtered, below detection, below quantification, zero or imputed. Those are not interchangeable. Preserve the source code and document the interpretation. If the reason is unknown, say that rather than replacing it with an observed absence.

19. Can a machine translator be used for a proteomics manuscript?

It can assist with drafting human-facing language, but it cannot be trusted to preserve every unit, protein-group limitation, score definition or controlled code without verification. Protect machine fields, maintain a domain-specific termbase, and require bilingual and scientific review commensurate with the stakes.

20. Should translation quality be measured only by grammar?

No. Language fluency, fidelity of scientific meaning, machine-data integrity and reproducibility are separate concerns. ISO 5060:2024 is useful for human translation-output evaluation, but it is not a replacement for schema, scientific and provenance checks.

21. Does a repository accession prove that a protein identification is correct?

No. The accession identifies a project or dataset within a repository ecosystem. It helps discovery and traceability; it does not automatically certify the correctness of each reported result. Verify what files were deposited and which scientific interpretation is supported.

22. Why is a reproducible sample manifest important?

Because sample labels, fractions, injections, conversion outputs and search results can look similar while representing different entities. A manifest allows a reviewer to follow a stated observation back to the right original material. Translating labels without preserving those joins can silently change the experiment.

23. What if a translated title must work for SEO?

Research how target-language audiences actually ask the relevant question, then write a truthful title that preserves the science. Do not insert “molecular weight” just because it is a common query if the article is about m/z. Search clarity and scientific precision should reinforce one another.

24. How can I decide whether to expand an abbreviation?

Expand on first meaningful use when this serves the target reader, preserve standardized acronyms and exact instrument or format names, and avoid assigning a broader meaning than the source. Abbreviations in structured data should remain unchanged even if a translated glossary gives a natural expansion.

25. What does a scientific release reviewer need?

The exact source version, a target-language draft, a bilingual terminology ledger, a list of locked fields, unit and identifier checks, format-validation results where applicable, evidence links and records of unresolved issues. The purpose is to make it possible for another qualified person to reconstruct the original claims.

26. How do I teach this to someone without proteomics training?

Begin with three distinctions: the machine observed signals, software assigned interpretations, and authors wrote conclusions. Then use an example where a precursor, peptide candidate and protein group are clearly different objects. Only introduce formal file formats and FDR after the learner can explain those boundaries.

27. Why are specialized scientific reviewers still necessary?

A bilingual expert may know the language but not the interpretation of a specific score, calibration procedure or protein-inference model. A scientist may know the instrument but overlook target-language ambiguity. High-quality release requires complementary expertise and a process for resolving disagreements from evidence.

28. Is this guide useful for other analytical sciences?

Yes, as a method for preserving physical quantities, processing history, confidence and record identity. But metabolomics, environmental chemistry, imaging mass spectrometry and targeted clinical assays have their own standards and validation expectations. Transfer the questions, not a set of proteomics-specific answers.

29. What does a fully defensible translation look like?

It lets a qualified target-language reader reconstruct the same experiment, the same computed analysis, the same limitations and the same scope of findings without changing source identifiers or data. Every controversial term has a reasoned choice and an accountable reviewer. Fluency is part of the finish, not the entire definition of correctness.

Four-week implementation plan: acquire the distinctions, practise the gates, test transfer

Week 1 — Map the record and establish the terminology boundary

Day 1: read the first three paragraphs and select three queries your target readers actually use. Day 2: distinguish m/z, ion mass, charge and isotope patterns. Day 3: compare a spectrum, chromatogram and an extracted-ion trace. Day 4: label MS1, MS2, precursor, product ion and acquisition modes. Day 5: inspect an mzML schema example without modifying machine fields. Day 6: create ten bilingual glossary entries with accession references. Day 7: test yourself by explaining, without notes, why accurate terminology requires both physical context and data structure. The deliverable is a one-page source-to-target field-class map, not a translated full data file.

Week 2 — Learn the measurement-to-inference chain

Day 8: map one hypothetical sample through preparation, injection and run IDs. Day 9: distinguish raw from centroided or otherwise processed spectra. Day 10: describe search databases, enzymes and mass tolerances. Day 11: inspect peptide-spectrum matches and protein ambiguity groups. Day 12: compare FDR, PEP and q-value claims. Day 13: practise translation of modification sites with unresolved localization. Day 14: run a blind target-only comprehension test. The deliverable is a short sample report in which every candidate identification has its source evidence and the right level of certainty.

Week 3 — Build repeatable QA and diagnostic literacy

Day 15: build a locklist for IDs, sequences, units and schema keys. Day 16: reconstruct a chromatogram’s time scale and concentration versus intensity distinction. Day 17: test reporter channels and biological replicate counts. Day 18: handle non-detects, missing values and imputation. Day 19: annotate a PRIDE-like file manifest and SDRF-like sample table. Day 20: perform a separate language review, data-format review and scientist review. Day 21: record every error by type, severity and repaired downstream occurrences. The deliverable is a release checklist that catches both fluent scientific errors and malformed structured content.

Week 4 — Demonstrate transfer and controlled publication

Day 22: localize a fictional LC-MS/MS methods paragraph. Day 23: draft a bilingual figure legend while preserving peak labels and numerical meaning. Day 24: write a repository description that distinguishes data availability from validation. Day 25: adapt the method to a small-molecule or ion mobility record while noting where proteomics assumptions stop applying. Day 26: ask a second reader to reconstruct the experiment from the translated documentation. Day 27: reconcile any mismatches and repeat the test. Day 28: archive the source snapshot, reviewed translation, glossary, version references and sign-off. The final standard is independent recoverability of meaning, not merely completion of a word quota.

Release scorecard: distinguish language, science and data-integrity errors

• Critical — block release: corrupted machine IDs, binary data or sequence notation; swapped sample groups; changed units with incorrect magnitude; new biological claims; FDR misrepresented as individual certainty; privacy violations.
• Major — repair and recheck: missing acquisition mode, ambiguous protein-group attribution, omitted method revision, wrong score direction, unstated quantification scale, overwritten sample or repository metadata.
• Moderate — revise for usability: inconsistent expanded acronyms, unexplained established symbols, target-language sentence structure that makes participant relationships unclear, incomplete but not misleading glossary explanations.
• Minor — polish: stylistic awkwardness or formatting differences that do not change scientific meaning, identifiers, source scope or reader action.
• Independent test: read source and target side by side; read target without source; validate data files with compatible scientific tools; trace sampled conclusions back through the manifest.
• Release authority: linguistic editor approves clarity and fidelity; subject specialist approves scientific meaning; data specialist approves intact links and format validation; project owner accepts residual limitations.

A numerical scorecard can make reviews consistent, but no average score should override a critical failure. A beautiful translation with one swapped treatment/control assignment is unsafe. Conversely, small improvements in style can be made after the scientific and structural invariants are secure. Where ISO 5060:2024 is used as translation-evaluation guidance, the specific scoring policy still belongs to the project and should not be misrepresented as an automatic scientific certificate.

Sources, version notes and further study

This article was researched for publication on 31 October 2026. Reference pages can change. The HUPO-PSI mzML repository provides the community format context and its historical 1.1.0 specification material. The mzIdentML specification identifies version 1.3.0 as the June 2024 final specification and provides detailed evidence and threshold semantics. The EMBL-EBI PRIDE guide describes the kinds of experimental and analytical files a proteomics dataset may include. The SDRF validator addresses structured sample and data relationships. The PRIDE submission guidance explains current distinctions among raw, native analysis and standardized result files.

For related result and QC formats, consult HUPO-PSI mzTab-M format documentation, mzQC examples and the mzSpecLib project documentation, noting whether a resource is a released standard, specification or work in progress. For a separate evaluation of human-facing translation quality, see ISO 5060:2024. These links are reference sources, not endorsements of invented workshop records. Use official release notes and validators for any real deployment.

Next routes: return to the translation root without replacing specialist owners

Continue to Master Art of Translation | The Complete System for Moving Meaning Between Languages when the question broadens from scientific evidence to the general principles of context, equivalence, audience, terminology, translation memory, revision and ethical limits. That page remains the canonical architectural root; this child handles the narrower problem of preserving mass spectrometry and proteomics evidence across languages.

For mzML fields, spectra, chromatograms, controlled vocabulary and binary-array preservation in a more focused workflow, use the already published Translate Like a Pro | Localize mzML Mass Spectrometry Spectra, Chromatograms, Controlled Vocabulary and Binary Arrays Without Changing Peaks. For the wider language of methods, peer review, citations, statistical evidence and scholarly claims, continue to Master Art of Translation | The Scientific Research & Scholarly Publishing Translation System. These are complementary specialist routes rather than pages to rewrite or absorb.

When the obstacle is learning and retrieving difficult scientific words, enter the protected Vocabulary Learning Hub; it owns vocabulary acquisition, memory, usage and transfer. When the obstacle is English grammar, reference, sentence structure or register within English, consult How English Works V1.1. Return here when the scientific meaning must cross between languages and a translator needs to protect units, identifiers, uncertainty and reproducibility.

Final principle: a translated result must still be the same evidence

In mass spectrometry, the evidence does not arrive as one finished sentence. It passes through a sample, instrument signals, processing transformations, candidate assignments, statistical controls, biological inference, written interpretation and sometimes a public repository. At every layer, a familiar word can conceal a technical distinction. The translator’s task is not to make the method sound more certain, simpler or more impressive than it was. It is to let another reader receive the same justified claim using the linguistic resources available to them.

Protect the data’s structure. Define the quantity. Preserve the original evidence level. Translate controlled terms with identifiers in view. Keep statistical limitations in the sentence. Reconstruct sample relationships. Test both the machine-readable record and the human-facing meaning. Record decisions so that a later editor can distinguish a translation correction from a new scientific analysis. These steps make a long guide useful in practice: they turn language work into a traceable part of the research evidence chain.

The decisive question is therefore not, “Does the target text sound like a scientific paper?” It is, “Could a qualified reader using this translation locate the same source data, understand the same operations and reach the same bounded interpretation without being misled?” If the answer is yes under an explicit review standard, the translation has done its job. If not, more elegant words are not the solution; the evidence link that broke must be restored.