eduKateSG · Why Science?
Connect every proteomics data file to the sample and design that produced it—without losing factors, controlled terms, validation or privacy
Choose the closest route, then return to the full index whenever you need the wider evidence chain.
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. SDRF-Proteomics is the HUPO-PSI Sample and Data Relationship Format for proteomics. The official page says it is tab-delimited, with each row representing a relationship between a sample and a data file; sample properties come first, followed by data-file properties. It was released as an official PSI specification on 24 May 2023. SDRF-Proteomics captures sample metadata and experimental design needed for reuse and reprocessing, but it does not aim to encode downstream comparisons, false-discovery thresholds or every statistical decision. 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: mzML and open mass-spectrometry data; mzQuantML and quantitative proteomics data; mzTab and tabular proteomics results; mzQC and mass-spectrometry quality metrics; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.
Primary and official sources checked on 12 October 2026: HUPO-PSI official SDRF-Proteomics page and 24 May 2023 specification release, checked 12 October 2026; Official SDRF-Proteomics specification repository and validation guidance, checked 12 October 2026; Official sdrf-pipelines validator repository, 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, repository services and recommendations can change, so preserve dated documentation, exact source data and software versions.
A deeper reading route
A raw proteomics file can contain millions of measurements and still be scientifically difficult to reuse if no one knows which organism, tissue, condition or preparation produced it. SDRF-Proteomics addresses that missing bridge. Its rows connect source material, samples, assays and data files, while columns express characteristics, factor values and technical details in a consistent table.
The official PSI page gives a clear row rule: each row corresponds to a relationship between a sample and a data file. Each column describes a property of the sample or file, and each cell carries the value for that row. This simple design is powerful because it supports both human inspection and automated workflows, provided that repeated samples, fractions and runs are represented consistently.
Column order follows experimental flow. Sample properties such as organism, disease or phenotype precede data-file properties such as label, fraction identifier and file name. The order is not merely visual; it helps readers reconstruct how biological material became measurements. A spreadsheet that rearranges or renames headers casually can break that shared meaning.
The standard deliberately has boundaries. The official page says it does not aim to capture downstream analysis such as which samples were compared, how samples form study variables, or thresholds for false discovery or p-values. Those details belong in analysis metadata, result formats and methods. Keeping the boundary prevents one convenient table from becoming an incomplete substitute for the whole workflow.
Controlled terms improve reuse. An organism, tissue, instrument or modification should use a declared ontology accession where the specification calls for one. A human-friendly label stays useful, but the accession reduces spelling and synonym problems. Ontology releases and local mappings still need provenance because metadata can drift even when file names do not.
Unknown and not-applicable states require care. Blank cells, the word unknown and a deliberately inapplicable field are not equivalent. The official requirements include support for unknown values and characteristics. A submission should follow the specification’s exact conventions and explain why critical metadata are missing rather than filling them with a convenient guess.
Fractions, labels and technical repeats complicate row counting. Several files may come from one sample, and one multiplexed file can represent labelled material from several samples. The table must represent the real relationships rather than imply one row equals one independent biological replicate. Statistical software needs the experimental unit, not a count of filenames.
Validation helps catch header syntax, term structure, required fields and cross-field rules. The project provides an official validator through sdrf-pipelines and shows a command-line validation route. Passing validation is valuable, but it does not prove that a tissue label is true, sample identities were not swapped or consent permits open release.
A practical audit chooses one raw file and traces backwards through every SDRF field to the source sample, characteristics, factors, preparation and instrument. Then it checks the repository landing page, mzML header and laboratory manifest. Disagreements should become visible corrections or documented limitations, not silent spreadsheet edits.
The cheerful promise is that good metadata lets old experiments ask new questions. A well-built SDRF can help automated reprocessing, multi-omics integration, teaching and quality review. It turns careful laboratory organisation into reusable scientific infrastructure while respecting that privacy, governance and interpretation still require human judgement.
Round-trip testing is essential when SDRF moves through spreadsheets. Import identifiers as text, protect tabs and leading zeroes, export to a new file, and compare the parsed table with the original. Visual similarity is not enough: header strings, column count, row order, accessions and filenames must retain exact meaning.
Design diagrams complement the table. A small graph of sources, samples, labelling channels, fractions and files can expose relationships that are difficult to spot across hundreds of rows. The diagram should be generated from the SDRF or checked against it, so it becomes a validation view rather than a separate manual story.
Privacy review should examine combinations, not only individual columns. A rare disease, age range, institution and sampling date can identify someone even when names are removed. Use the least identifying labels that still support the scientific purpose, document transformations and place sensitive fields under appropriate access controls.
Reuse readiness improves when the SDRF is bundled with a repository manifest, acquisition files, open spectra, results and analysis workflow. The table then acts as a map rather than a lonely attachment. File checksums and stable PXD references help future analysts test that every row still points to the same evidence.
A final scientific review should ask whether factor values match the stated hypothesis, whether biological replication exists at the correct level and whether batch variables are represented. These questions go beyond syntax. They determine whether an automated reanalysis can make a fair comparison instead of amplifying an undocumented design confounder.
Repository preparation benefits from a preflight inventory. Count expected sources, samples, labels, fractions and files, then compare those counts with distinct SDRF values and the upload manifest. Differences may be legitimate, but each should have an explanation. This simple reconciliation catches omissions and duplicate relationships before a public accession makes the table part of a long-lived scientific record.
Metadata revisions should be versioned rather than silently overwritten. Preserve the submitted SDRF, corrected SDRF, validator reports and a change table describing affected rows and reasons. If a sample label or factor value changes, downstream analysis may need to be rerun; the correction is scientific provenance, not merely editorial housekeeping. Keep every revision clearly dated and attributable.
| Stage | Question | Evidence | Responsible output |
|---|---|---|---|
| Describe | What material was studied? | Source names, sample characteristics and controlled terms | An interpretable biological context |
| Connect | Which preparation produced which file? | Rows, labels, fractions, assays and filenames | A sample-to-data relationship |
| Compare | Which intended factors vary? | Factor values and experimental-design context | A design that can be reconstructed |
| Govern | Can the metadata be shared and trusted? | Validation, provenance, privacy review and hashes | A reusable metadata package |
Inside this guide
1–12 · Foundations and design
- 1. Start with the design question
- 2. Use the official specification
- 3. Keep the file tab-delimited
- 4. Treat one row as a relationship
- 5. Treat each column as a property
- 6. Follow experimental flow
- 7. Name sources and samples consistently
- 8. Describe organism and material
- 9. Describe factors separately
- 10. Describe technical preparation
- 11. Describe instruments
- 12. Link data filenames exactly
13–24 · Models and estimation
- 13. Represent fractions carefully
- 14. Represent labels carefully
- 15. Represent replicates honestly
- 16. Use controlled terms
- 17. Handle unknown values explicitly
- 18. Respect scope boundaries
- 19. Validate headers
- 20. Validate required information
- 21. Validate controlled values
- 22. Validate row consistency
- 23. Use the official validator
- 24. Compare with source records
25–36 · Diagnostics and interpretation
- 25. Protect sensitive metadata
- 26. Separate metadata from measurement
- 27. Separate design from conclusion
- 28. Connect to Primary Science
- 29. Connect to Secondary Science
- 30. Connect to mathematics
- 31. Connect to computing
- 32. Connect to pathways
- 33. Did you know?
- 34. Write a reader-ready methods note
- 35. Build a validation ladder
- 36. Finish with a reproducible bundle
Section 1 of 36
1. Start with the design question
Name the biological units, conditions, treatments, technical steps and expected data files before filling the table.
Working checkpoint — SDRF-Proteomics: Draw the sample-to-file graph first. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Section 2 of 36
2. Use the official specification
The PSI page records SDRF-Proteomics as an official specification released on 24 May 2023.
Working checkpoint — sample metadata: Pin the specification and validator versions. Read every row beside the experimental unit, label, fraction and controlled terms.
For future study, connect biology to data stewardship while checking official pathways separately.
Section 3 of 36
3. Keep the file tab-delimited
Tabs define columns. Visual alignment with spaces or spreadsheet formatting does not preserve the machine-readable structure.
Working checkpoint — experimental design: Validate raw text after every export. Preserve missing and conflicting metadata; a tidy table is not the original laboratory record.
For scientific writing, cite specification, samples, design, files, validator and limitations.
Section 4 of 36
4. Treat one row as a relationship
Each row connects a sample with a data file under the stated preparation and design attributes.
Working checkpoint — data file: Check that repeated rows represent real relationships. Separate format validity from sample identity, design quality and biological conclusion.
For students, ask which sample produced each fictional file.
Section 5 of 36
5. Treat each column as a property
Headers describe a sample or data-file attribute; cells give the value for the row.
Working checkpoint — factor value: Use specification-compliant header names. Treat identifiers, tabs and filenames as scientific inputs: one mismatch can detach evidence.
For parents, ask whether reusable data keep privacy and design visible.
Section 6 of 36
6. Follow experimental flow
Sample properties should precede properties of generated data files, reflecting the natural workflow.
Working checkpoint — tab-delimited table: Review the column order as a process map. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Section 7 of 36
7. Name sources and samples consistently
Source and sample identifiers support joins across laboratory manifests and repositories.
Working checkpoint — SDRF-Proteomics: Keep governed identifiers stable through de-identification. Read every row beside the experimental unit, label, fraction and controlled terms.
For future study, connect biology to data stewardship while checking official pathways separately.
Section 8 of 36
8. Describe organism and material
Organism, tissue, cell type and disease characteristics give biological meaning to files.
Working checkpoint — sample metadata: Use appropriate controlled terms and accessions. Preserve missing and conflicting metadata; a tidy table is not the original laboratory record.
For scientific writing, cite specification, samples, design, files, validator and limitations.
Section 9 of 36
9. Describe factors separately
Factor values identify intended experimental variables, while characteristics describe the material.
Working checkpoint — experimental design: Do not turn every descriptive field into a comparison factor. Separate format validity from sample identity, design quality and biological conclusion.
For students, ask which sample produced each fictional file.
Section 10 of 36
10. Describe technical preparation
Cleavage, labelling, fractionation and enrichment information shape what spectra can represent.
Working checkpoint — data file: Match metadata against the laboratory protocol. Treat identifiers, tabs and filenames as scientific inputs: one mismatch can detach evidence.
For parents, ask whether reusable data keep privacy and design visible.
Section 11 of 36
11. Describe instruments
Instrument-model metadata aids reprocessing and quality comparisons but should agree with source-file headers.
Working checkpoint — factor value: Cross-check against mzML or vendor records. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Section 12 of 36
12. Link data filenames exactly
A one-character filename mismatch can detach metadata from measurements.
Working checkpoint — tab-delimited table: Resolve every file and preserve hashes. Read every row beside the experimental unit, label, fraction and controlled terms.
For future study, connect biology to data stewardship while checking official pathways separately.
Section 13 of 36
13. Represent fractions carefully
Multiple fractions from one preparation are technical partitions, not independent biological samples.
Working checkpoint — SDRF-Proteomics: Keep fraction identifiers and parent sample visible. Preserve missing and conflicting metadata; a tidy table is not the original laboratory record.
For scientific writing, cite specification, samples, design, files, validator and limitations.
Section 14 of 36
14. Represent labels carefully
Multiplex labels can connect several samples to signals in one acquired file.
Working checkpoint — sample metadata: Validate label uniqueness and channel mapping. Separate format validity from sample identity, design quality and biological conclusion.
For students, ask which sample produced each fictional file.
Section 15 of 36
15. Represent replicates honestly
Biological and technical replicates answer different uncertainty questions.
Working checkpoint — experimental design: Record the experimental unit outside filename patterns. Treat identifiers, tabs and filenames as scientific inputs: one mismatch can detach evidence.
For parents, ask whether reusable data keep privacy and design visible.
Section 16 of 36
16. Use controlled terms
Ontology accessions reduce ambiguity for organisms, tissues, instruments and methods.
Working checkpoint — data file: Store accession, label and ontology release. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Section 17 of 36
17. Handle unknown values explicitly
Unknown, unavailable and not applicable represent different evidence states.
Working checkpoint — factor value: Follow specification conventions and explain important gaps. Read every row beside the experimental unit, label, fraction and controlled terms.
For future study, connect biology to data stewardship while checking official pathways separately.
Section 18 of 36
18. Respect scope boundaries
SDRF-Proteomics does not encode every downstream contrast, study variable or statistical threshold.
Working checkpoint — tab-delimited table: Link separate analysis metadata instead of overloading cells. Preserve missing and conflicting metadata; a tidy table is not the original laboratory record.
For scientific writing, cite specification, samples, design, files, validator and limitations.
Section 19 of 36
19. Validate headers
A validator checks whether columns follow recognised patterns and ordering rules.
Working checkpoint — SDRF-Proteomics: Reject improvised headers before repository submission. Separate format validity from sample identity, design quality and biological conclusion.
For students, ask which sample produced each fictional file.
Section 20 of 36
20. Validate required information
Different experiment types can require different metadata to support reuse.
Working checkpoint — sample metadata: Choose the relevant template and document exceptions. Treat identifiers, tabs and filenames as scientific inputs: one mismatch can detach evidence.
For parents, ask whether reusable data keep privacy and design visible.
Section 21 of 36
21. Validate controlled values
An accession should exist in the pinned ontology and describe the accompanying label.
Working checkpoint — experimental design: Resolve representative values independently. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Section 22 of 36
22. Validate row consistency
Repeated identifiers, label conflicts and impossible file relationships can reveal design errors.
Working checkpoint — data file: Run uniqueness and cross-field checks. Read every row beside the experimental unit, label, fraction and controlled terms.
For future study, connect biology to data stewardship while checking official pathways separately.
Section 23 of 36
23. Use the official validator
The project documents sdrf-pipelines for checking SDRF files against rules.
Working checkpoint — factor value: Keep command, version and full output. Preserve missing and conflicting metadata; a tidy table is not the original laboratory record.
For scientific writing, cite specification, samples, design, files, validator and limitations.
Section 24 of 36
24. Compare with source records
Validation cannot detect a confidently typed but wrong tissue or swapped sample.
Working checkpoint — tab-delimited table: Reconcile against laboratory and instrument manifests. Separate format validity from sample identity, design quality and biological conclusion.
For students, ask which sample produced each fictional file.
Section 25 of 36
25. Protect sensitive metadata
Disease, phenotype and sample labels can expose personal or governed information even without names.
Working checkpoint — SDRF-Proteomics: Apply consent, de-identification and access review. Treat identifiers, tabs and filenames as scientific inputs: one mismatch can detach evidence.
For parents, ask whether reusable data keep privacy and design visible.
Section 26 of 36
26. Separate metadata from measurement
SDRF explains relationships and design; mzML and result formats carry measurements and analyses.
Working checkpoint — sample metadata: Keep links strong while preserving format roles. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Section 27 of 36
27. Separate design from conclusion
A clearly annotated factor does not prove that it caused an observed molecular change.
Working checkpoint — experimental design: Use controls, statistics and biological evidence separately. Read every row beside the experimental unit, label, fraction and controlled terms.
For future study, connect biology to data stewardship while checking official pathways separately.
Section 28 of 36
28. Connect to Primary Science
A fictional sample-to-file table lets younger learners separate an observation from the label attached to it. The goal is careful comparison, not memorising specialist jargon.
Working checkpoint — data file: Use invented records, define every column and ask what remains unknown. Preserve missing and conflicting metadata; a tidy table is not the original laboratory record.
For scientific writing, cite specification, samples, design, files, validator and limitations.
Section 29 of 36
29. Connect to Secondary Science
SDRF-Proteomics table supports lessons about molecules, measurement, controls and the difference between evidence and explanation.
Working checkpoint — factor value: Trace one claim to its supporting record and state one limitation. Separate format validity from sample identity, design quality and biological conclusion.
For students, ask which sample produced each fictional file.
Section 30 of 36
30. Connect to mathematics
Counts, coordinates, scores, ratios and uncertainty show why scientific data need units, definitions and compatible denominators.
Working checkpoint — tab-delimited table: Recalculate one simple example and explain every symbol. Treat identifiers, tabs and filenames as scientific inputs: one mismatch can detach evidence.
For parents, ask whether reusable data keep privacy and design visible.
Section 31 of 36
31. Connect to computing
Typed columns, ontology accessions and validators show how tables can drive reproducible workflows.
Working checkpoint — SDRF-Proteomics: Break one reference or field, run a check and repair it transparently. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Section 32 of 36
32. Connect to pathways
Proteomics metadata connects biology, laboratory science, data stewardship and software and can illuminate STEM education and career pathways.
Working checkpoint — sample metadata: Check current official education and qualification routes separately before making decisions. Read every row beside the experimental unit, label, fraction and controlled terms.
For future study, connect biology to data stewardship while checking official pathways separately.
Section 33 of 36
33. Did you know?
A short, readable SDRF-Proteomics table record can depend on a much longer chain of source data, identifiers, versions and validation receipts.
Working checkpoint — experimental design: Follow one record backwards until every dependency is named. Preserve missing and conflicting metadata; a tidy table is not the original laboratory record.
For scientific writing, cite specification, samples, design, files, validator and limitations.
Section 34 of 36
34. Write a reader-ready methods note
State the specification or service, release, inputs, software, validation, uncertainty and the narrow claim the evidence supports.
Working checkpoint — data file: Place the note beside the machine-readable material. Separate format validity from sample identity, design quality and biological conclusion.
For students, ask which sample produced each fictional file.
Section 35 of 36
35. Build a validation ladder
Check bytes, structure, identifiers, controlled meanings, cross-references, scientific plausibility and a small known fixture in that order.
Working checkpoint — factor value: Stop at the first unresolved layer instead of polishing a later result. Treat identifiers, tabs and filenames as scientific inputs: one mismatch can detach evidence.
For parents, ask whether reusable data keep privacy and design visible.
Section 36 of 36
36. Finish with a reproducible bundle
Deliver source data or stable identifiers, metadata, software versions, configuration, validator output, checksums and a plain-language limitations note.
Working checkpoint — tab-delimited table: Make the published conclusion traceable and reviewable. Keep source, sample, characteristics, factors, technical steps and file links together.
For a classroom model, map invented samples through preparation to filenames.
Contents · Previous section · Continue to the Science Learning Hub