Why Science? | mzTab-M and Tabular Metabolomics Results

Science learning becomes powerful when students can tell a measured pattern from a model, a statistical decision from a mechanism, and a useful lead from a finished conclusion. mzTab-M is the HUPO Proteomics Standards Initiative lightweight, tab-delimited format for reporting mass-spectrometry-based small-molecule metabolomics and lipidomics results. The official project lists ratified version 2.0.0 from March 2019 and describes it as a non-backwards-compatible specification distinct from proteomics-focused mzTab 1.0. The current official repository identifies 2.1.0 as in development in 2026, so production reports must state whether they use stable 2.0.0 or a development snapshot. mzTab-M makes results accessible to people and software, but a valid table does not prove molecular identity, concentration, pathway mechanism or clinical meaning. This guide supports Primary Science, PSLE Science, Secondary Science, O-Level Science and STEM education as a longform evidence-reading route; it is not medical advice, a laboratory protocol, an admission promise or a career guarantee.

Related eduKate reading: mzTab and tabular proteomics results; mzML and open mass-spectrometry data; mass spectrometry, ionisation and mass-to-charge evidence; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.

Primary and official sources checked on 12 October 2026: HUPO-PSI mzTab project page and ratified mzTab-M 2.0.0 specification, checked 12 October 2026; HUPO-PSI mzTab-M official development repository, including 2.0 resources and 2.1.0 development status, checked 12 October 2026; HUPO-PSI jmzTab-m reference implementation and validator, checked 12 October 2026; 2026 Singapore–Cambridge O-Level Biology syllabus; MOE G2/G3 Lower Secondary Science teaching and learning syllabus; 2026 MOE G2 Computing syllabus. Specifications, vocabularies, tools and recommendations can change, so preserve dated documentation, exact source data and software versions.

A deeper reading route

Metabolomics workflows can produce feature tables, tentative annotations, confirmed compounds and quantitative values across many samples. mzTab-M places metadata and typed sections in one tab-delimited file so collaborators can inspect the broad result without proprietary software. Its value is a shared map of the result, not a replacement for source spectra, chromatography, standards or laboratory notes.

The small-molecule section can describe reported chemical entities, while feature and evidence layers keep links to measured signals. One chemical entity may correspond to several features because of adducts, isotopes, fragments or chromatographic peaks. Conversely, one unresolved feature may have several candidate identities. The data model should preserve those relationships rather than force a one-row-one-molecule story.

Metadata gives context to every later value: instruments, software, databases, samples, assays, study variables, spectra references, score definitions and units. Numbered keys and controlled terms make this context machine-readable. A reviewer should resolve every referenced object before comparing columns across files.

Mass-to-charge ratio and retention time help locate a feature, but they do not uniquely identify a compound. Isomers can share formula and mass; adducts shift observed mass; chromatography and ionisation depend on method. Identification confidence needs its evidence category, database, spectral match, standard or orthogonal confirmation.

Spectra references connect summary results to observations. The reference must identify the exact source file and scan or spectrum under a declared convention. If public, a durable identifier can improve retrieval. A missing or broken reference turns a plausible table cell into an uninspectable claim.

Adduct and charge annotations change the relationship between neutral mass and measured mass-to-charge value. Parsing them as decoration can assign the wrong formula or merge distinct ions. Keep the notation, charge, mass calculation and software rule together.

Quantitative columns need a clear level, unit and normalisation method. Peak area, peak height, relative abundance, concentration and fold change answer different questions. Study-variable summaries are derived values, not raw observations. Reports should keep assay-level evidence and calculation scripts alongside the table.

Missing values can reflect non-detection, extraction failure, filtering, unmatched features or unavailable reporting. Those states are scientifically different. Imputation is a model and must be named, parameterised and separated from measured values.

Lipidomics adds structural ambiguity. A lipid class, sum composition, fatty-acyl composition and fully resolved molecular species represent different confidence levels. mzTab-M can report structured names and identifiers, but the author must not promote a partial annotation into a precise structure without supporting fragmentation or standards.

Controlled vocabularies reduce ambiguity for instruments, software, assays, databases, units and scores. Syntax validation checks rows and columns; semantic validation checks allowed terms, references and value rules. A file opening in a spreadsheet proves neither type of validity.

The 2.0.0 format is ratified and supported by a reference implementation. The repository says 2.1.0 is in development in 2026, with goals including modularisation, FAIR use cases and intermediate outputs. That status matters: a development snapshot may change, so it should never be labelled as the stable standard.

mzTab-M is not backwards-compatible with mzTab 1.0. The shared family name can mislead users into applying proteomics row rules to metabolomics files. Always state the domain and version in the filename, methods and reader guidance.

For students, mzTab-M resembles a museum catalogue for chemical signals. The catalogue makes many finds browsable, but each label should point back to the object, the measurement and the curator’s confidence. A readable name is still an evidence claim.

A classroom model can use fictional features with mass, retention time and candidate compounds. Students add an adduct, score and confidence level, then see why equal mass does not prove equal identity. The exercise joins chemistry, ratios, tables and uncertainty.

A reproducible bundle includes mzTab-M, source mzML or repository records, extraction and alignment parameters, libraries and database releases, standards, controlled vocabularies, validator output, scripts and hashes. These receipts let another team reconstruct both the table and its limits.

The cheerful practical benefit is shared visibility. A good mzTab-M report invites chemists, biologists and data scientists into the same result while keeping source links and uncertainty close. Accessibility makes better collaboration possible when it is paired with disciplined evidence.

A cautious reviewer begins with one small-molecule row and follows every linked feature and spectrum. The row should reveal how names, identifiers, adducts, charge, formula, m/z, retention and reliability relate. A broken link or unexplained score should pause interpretation rather than being filled by a familiar database guess.

Spreadsheet round trips are particularly risky for chemical identifiers and scientific notation. Import columns as declared text or numbers, save to a new file and compare with a standards-aware parser. The human-readable design invites inspection, but original bytes and typed validation protect the scientific record.

Between-study merging needs a common model for feature detection, alignment, adduct grouping, identification confidence, normalisation and missingness. Two rows with the same compound name may represent different isomers or confidence levels. Prefer accessions, spectra and explicit mappings to name-only joins.

A compact publication note can state stable mzTab-M version, source data, extraction software, databases and libraries, confidence policy, quantitative units and validator. The reader then gets both a welcoming overview and a precise route into deeper evidence.

Feature grouping should remain auditable. Isotopes, adducts and in-source fragments may be combined into one putative chemical entity, but that grouping depends on tolerances, correlation and method assumptions. Preserve the original features and the grouping rule so another analyst can test whether the same evidence supports the same entity.

Blank controls and pooled quality-control samples help reveal contaminants, drift and batch effects. Their roles belong in metadata and the methods narrative, while filtering decisions should remain reproducible. A valid result row should not hide that its signal was also strong in a procedural blank.

Identification levels should be reported without marketing shortcuts. Exact standard confirmation, library match, formula-level annotation and unknown feature are useful but different outcomes. Map the project’s terminology to a documented confidence framework and retain the underlying score and evidence rather than publishing labels alone.

Quantitative comparison also needs transparent denominators. Normalising to total signal, internal standards, sample mass or another reference produces different questions. State the denominator in words and structured metadata, then keep pre-normalisation values so future readers can evaluate the choice.

Long-term reuse requires a release-aware reader. The tab-delimited surface may still look familiar years later, while controlled terms, validators and software expectations have moved. Archive the stable 2.0.0 specification, validation receipts and exact dependency versions with the file. If a future team migrates the result toward 2.1, keep the original and publish a field-by-field transformation log. That approach preserves the historic scientific claim while allowing newer tools and FAIR workflows to benefit from carefully reviewed improvements.

A final spot check should open the file in both a plain-text editor and an independent parser. People can catch surprising names or scales, while software catches shifted tabs and broken references. Agreement between those views gives the readable format its full value.

A four-step claim ladder for this method

Stage	Question	Evidence	Responsible output
Describe	Which study and method produced the result?	Metadata, assays, study variables, software and CV terms	An interpretable context
Observe	Which signal was measured?	Features, m/z, retention time and spectra references	A traceable observation
Annotate	Which chemical entity may explain it?	Identifiers, adducts, scores and confidence	A bounded identification
Quantify	How was amount compared?	Units, values, normalisation and missingness	A reproducible result