eduKateSG · Why Science?
Follow biodiversity evidence from encounter to sample to collection record—without losing place, process or uncertainty
Choose the route closest to your specimen, field observation or environmental sample, then keep every identity and collection step visible enough to audit.
Science learning becomes powerful when students can trace a biodiversity claim back to what was encountered, collected, prepared and identified. The Biological Collections Ontology, or BCO, is described by the OBO Foundry as an application ontology supporting interoperability of biodiversity data, including museum collections, environmental and metagenomic samples, and ecological surveys. Its goal is not merely to label a specimen: it helps connect organisms, material samples, collection or observation processes and the resulting evidence records. A shared class can link a field notebook, museum object and sequence-derived sample, but it cannot repair an unknown location, a broken chain of custody or a biased survey. This guide supports Primary Science, PSLE Science, Secondary Science, O-Level Science, O-Level Biology 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: Environment Ontology and comparable environmental context; NCBI Taxonomy and stable organism identifiers; Uberon and cross-species anatomy; Ontology for Biomedical Investigations; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.
Primary and official sources checked on 13 October 2026: OBO Foundry Biological Collections Ontology overview, checked 13 October 2026; Official Biological Collections Ontology source repository; Canonical Biological Collections Ontology OWL release; BCO introductory publication listed by OBO Foundry; 2026 Singapore–Cambridge O-Level Biology syllabus; MOE G2/G3 Lower Secondary Science teaching and learning syllabus; 2026 MOE G2 Computing syllabus. Ontologies, definitions and release products can change, so preserve dated documentation, exact source records and file versions.
How biodiversity evidence moves through collections
A preserved specimen is not the organism itself. It is material derived from an organism through a collection process, carrying evidence about identity, place and time while losing other aspects of the living encounter.
BCO exists because biodiversity records span museums, herbaria, field surveys, environmental samples and metagenomic studies. Shared modelling helps these communities connect without pretending their evidence types are identical.
The OBO Foundry description names museum collections, environmental and metagenomic samples, and ecological surveys as target domains. Each has a different chain from encounter to claim.
Collection events matter. Who collected what, where, when, by which method and under which permit can change whether a sample supports distribution, abundance, trait or genomic research.
Material samples can be whole organisms, parts, tissues, environmental material or subsamples. Parent–child links protect against treating several derived tubes as independent encounters.
Institutional accessions create persistent collection identities. Catalogue numbers should be paired with institution codes and source databases so a reader can resolve the intended object.
Taxonomic identification is an interpretation that may change after expert review or sequencing. Preserve the original identification, determiner, date, evidence and later revisions.
Environmental context belongs beside the collecting process. ENVO terms can describe biome, feature or material, while coordinates, depth, substrate and weather retain measured detail.
Absence data require explicit effort. Failing to collect a species does not establish absence unless method, area, timing and detection probability support that conclusion.
Metagenomic samples may contain DNA from many organisms and sources. Sequence assignment is not the same evidence as observing or collecting a whole organism.
Museums and field collections carry legal, ethical and cultural responsibilities. Permits, access and benefit-sharing, sensitive-locality rules and community governance are part of responsible reuse.
The optimistic promise is discoverability with custody: a researcher can find comparable records while every sample still points back to the event, collection and uncertainty that gave it meaning.
Start with the evidence type. Observation, photograph, whole specimen, tissue voucher, environmental sample and sequence record support overlapping but different conclusions.
Keep event and object identifiers separate. One collecting event can produce several specimens, and one specimen can yield many preparations or subsamples.
Coordinates need datum, precision and uncertainty. A rounded public locality can protect a sensitive species but should not masquerade as the exact collection point.
Time needs resolution and timezone. “Spring” or a month may be the best honest record for legacy material; software should not invent a precise day.
Sampling methods create filters. Nets, traps, quadrats, swabs and water filtration capture different organisms and sizes even at the same place.
Preservation changes evidence. Formalin, freezing, drying and storage duration can affect morphology, chemicals, DNA and microbial composition.
A voucher anchors future re-identification. When destructive analysis consumes material, record what remains and which derivatives carry the evidence forward.
Duplicate or exchanged specimens need explicit links so institutions do not count them as separate encounters without understanding their shared origin.
For classroom work, a fictional schoolyard survey can demonstrate event, observation, sample and catalogue identity without disturbing wildlife.
Biodiversity interoperability succeeds when differences stay visible: harmonised identifiers should connect records, not erase how each was produced.
Legacy labels are evidence artifacts. Transcribe the original text, preserve images and add interpreted fields separately so later readers can distinguish source wording from modern georeferencing or taxonomy.
Georeferencing is a scientific interpretation. Record the gazetteer, map, method, uncertainty radius and reviewer rather than replacing a historical locality with one unexplained coordinate.
Sensitive data may need tiered access. A public record can use a generalised locality while authorised researchers retain a governed route to precise information.
Collection bias shapes biodiversity maps. Roads, cities, charismatic species, seasonal fieldwork and accessible habitats can appear as biological patterns unless sampling effort is modelled.
A specimen can support many studies, but every new preparation or measurement should cite the source accession and retain its own method, date and quality metadata.
Environmental DNA provides powerful detection evidence while introducing transport, persistence, contamination and reference-database questions. The organism may not have been present exactly where or when the molecule was collected.
Taxonomy releases and name changes require stable concept or taxon identifiers. Updating the displayed name should not sever earlier determinations and publications.
Before pooling collections, build a provenance matrix listing event, object, derivative, institution, taxon, locality, method, preservation, permit and uncertainty. It exposes hidden pseudoreplication and incompatible evidence.
A responsible map legend states whether points represent observations, specimens, samples, sequence detections or inferred ranges. Similar symbols can otherwise imply equal evidence.
BCO encourages a generous scientific habit: value the people, processes and collections that make a biodiversity claim checkable decades after the original fieldwork.
Field notes and collection databases answer complementary questions. Notes can preserve weather, behaviour, substrate and decision context that a catalogue schema omits, while structured records make identifiers and coordinates searchable. Digitisation should link both rather than treating transcription as a complete replacement.
Specimen condition is part of evidence quality. Damage, mould, fading, contamination, missing parts and previous destructive sampling can change which taxonomic or chemical conclusions remain defensible. Record condition assessments through time instead of presenting the accession as permanently unchanged.
Collections also document non-detections and failed attempts when surveys preserve them. Empty traps, contaminated controls and aborted preparations may feel unproductive, yet they define effort and reveal why surviving specimens are not a random sample of nature.
A barcode sequence and a museum voucher strengthen each other only when the linkage is secure. The sequence needs extraction and library identifiers; the voucher needs accession and tissue-sampling records; both need checks against swaps and contamination.
When records from many institutions are aggregated, catalogue numbers must remain qualified by collection identity. A bare number such as 12345 can legitimately exist in several museums, so interoperability requires globally resolvable or compound identifiers.
A biodiversity dashboard should expose the path behind its totals. Readers need to know whether a count represents records, collection events, individual organisms, specimens, derivatives or accepted taxa before comparing regions or years.
| Stage | Question | Evidence | Responsible output |
|---|---|---|---|
| Encounter | What organism or environmental material was encountered? | Field observation, location, time and method | A bounded occurrence claim |
| Collect | What material was removed or preserved? | Collection process, collector, permit and sample identity | A traceable material sample |
| Accession | How did it enter a collection or dataset? | Institution, catalogue, preparation and custody history | A persistent collection record |
| Interpret | What conclusion does the evidence support? | Identification, sequencing, context, uncertainty and review | A revisable biodiversity statement |
Inside this guide
1–12 · Encounters, samples and identity
- 1. Begin with the biodiversity question
- 2. Use the BCO identifier
- 3. Name the evidence type
- 4. Identify the encounter
- 5. Identify the collecting process
- 6. Identify the material sample
- 7. Link parent and child samples
- 8. Create collection identity
- 9. Preserve original labels
- 10. Record collector and team
- 11. Record time honestly
- 12. Record locality honestly
13–24 · Collections, mappings and validation
- 13. Connect environmental context
- 14. Connect taxonomic identity
- 15. Handle multiple determinations
- 16. Distinguish organism and specimen
- 17. Distinguish specimen and derivative
- 18. Record preparation history
- 19. Record chain of custody
- 20. Model duplicates and exchanges
- 21. Describe survey effort
- 22. Challenge absence claims
- 23. Challenge collection bias
- 24. Challenge environmental DNA
25–36 · Learning, stewardship and pathways
- 25. Protect sensitive localities
- 26. Respect permits and communities
- 27. Review ontology and taxon migrations
- 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 collection note
- 35. Build a validation ladder
- 36. Finish with a reproducible collections bundle
Section 1 of 36
1. Begin with the biodiversity question
Distribution, taxonomy, traits, ecology and genomics require different evidence and collection detail.
Working checkpoint — BCO identifier: Write the target claim before selecting records. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Section 2 of 36
2. Use the BCO identifier
A stable ontology ID distinguishes the intended class from local museum or survey language.
Working checkpoint — collecting process: Store identifier, label, definition and release. Record location precision, method, effort, preservation and permits before comparing records.
For computing, validate identifiers, parent–child sample links and coordinate uncertainty.
Section 3 of 36
3. Name the evidence type
Observation, image, specimen, tissue and environmental sample do not support identical claims.
Working checkpoint — material sample: Classify the evidence before analysis. Separate organism observation from sample-derived or sequence-derived detection.
For writing, say whether a record is an observation, specimen, sample or inferred occurrence.
Section 4 of 36
4. Identify the encounter
The starting event records an organism or material encountered at a place and time.
Working checkpoint — organismal entity: Keep observer, date and locality source. Preserve original labels and determinations beside every modern interpretation.
For students, ask what physical evidence exists and how it reached the record.
Section 5 of 36
5. Identify the collecting process
Collection is an action using a method that selects and transforms what becomes available for study.
Working checkpoint — occurrence evidence: Record protocol, equipment and effort. Pin BCO, ENVO and taxonomy releases with validation and migration notes.
For parents, the useful checkpoint is whether a map point has a traceable source and method.
Section 6 of 36
6. Identify the material sample
A sample is material obtained through a process and may be whole, partial or mixed.
Working checkpoint — provenance chain: Describe physical composition and condition. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Section 7 of 36
7. Link parent and child samples
Subsamples, extracts and preparations inherit origin but remain distinct material objects.
Working checkpoint — BCO identifier: Preserve the complete derivation chain. Record location precision, method, effort, preservation and permits before comparing records.
For computing, validate identifiers, parent–child sample links and coordinate uncertainty.
Section 8 of 36
8. Create collection identity
An accession or catalogue record makes an object findable within an institution.
Working checkpoint — collecting process: Store institution, collection and catalogue identifiers. Separate organism observation from sample-derived or sequence-derived detection.
For writing, say whether a record is an observation, specimen, sample or inferred occurrence.
Section 9 of 36
9. Preserve original labels
Historical wording can contain locality, collector and taxonomic clues not captured by modern fields.
Working checkpoint — material sample: Transcribe and image the source label. Preserve original labels and determinations beside every modern interpretation.
For students, ask what physical evidence exists and how it reached the record.
Section 10 of 36
10. Record collector and team
People and organisations provide provenance, expertise and responsibility for field records.
Working checkpoint — organismal entity: Use stable agent identifiers where appropriate. Pin BCO, ENVO and taxonomy releases with validation and migration notes.
For parents, the useful checkpoint is whether a map point has a traceable source and method.
Section 11 of 36
11. Record time honestly
Legacy records may provide only a year, season or range, which should not be converted to false precision.
Working checkpoint — occurrence evidence: Store value and temporal uncertainty. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Section 12 of 36
12. Record locality honestly
Place names, coordinates and habitat descriptions have different precision and can change over time.
Working checkpoint — provenance chain: Keep source locality, datum and uncertainty. Record location precision, method, effort, preservation and permits before comparing records.
For computing, validate identifiers, parent–child sample links and coordinate uncertainty.
Section 13 of 36
13. Connect environmental context
Biome, feature, material, depth and substrate help explain what was sampled.
Working checkpoint — BCO identifier: Pair ENVO terms with measured context. Separate organism observation from sample-derived or sequence-derived detection.
For writing, say whether a record is an observation, specimen, sample or inferred occurrence.
Section 14 of 36
14. Connect taxonomic identity
A taxon determination links evidence to a biological concept but can be revised.
Working checkpoint — collecting process: Record identifier, determiner, date and confidence. Preserve original labels and determinations beside every modern interpretation.
For students, ask what physical evidence exists and how it reached the record.
Section 15 of 36
15. Handle multiple determinations
Experts may disagree or update an identification after new characters or sequences appear.
Working checkpoint — material sample: Keep every determination and its status. Pin BCO, ENVO and taxonomy releases with validation and migration notes.
For parents, the useful checkpoint is whether a map point has a traceable source and method.
Section 16 of 36
16. Distinguish organism and specimen
The living organism and preserved material derived from it are not the same entity.
Working checkpoint — organismal entity: Avoid transferring every organism property to the specimen. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Section 17 of 36
17. Distinguish specimen and derivative
A tissue slide, DNA extract or image may derive from one accession without being interchangeable with it.
Working checkpoint — occurrence evidence: Give every derivative its own identifier. Record location precision, method, effort, preservation and permits before comparing records.
For computing, validate identifiers, parent–child sample links and coordinate uncertainty.
Section 18 of 36
18. Record preparation history
Fixation, drying, freezing and mounting affect which analyses remain reliable.
Working checkpoint — provenance chain: Link method, date and responsible laboratory. Separate organism observation from sample-derived or sequence-derived detection.
For writing, say whether a record is an observation, specimen, sample or inferred occurrence.
Section 19 of 36
19. Record chain of custody
Transfers, loans and destructive sampling can change location, condition and available material.
Working checkpoint — BCO identifier: Maintain event-based custody records. Preserve original labels and determinations beside every modern interpretation.
For students, ask what physical evidence exists and how it reached the record.
Section 20 of 36
20. Model duplicates and exchanges
Several institutions may hold duplicates from one event or organism.
Working checkpoint — collecting process: Link shared origin to prevent accidental independence. Pin BCO, ENVO and taxonomy releases with validation and migration notes.
For parents, the useful checkpoint is whether a map point has a traceable source and method.
Section 21 of 36
21. Describe survey effort
Area, duration, observers, traps and visits determine the chance of finding organisms.
Working checkpoint — material sample: Store denominators needed for abundance or absence claims. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Section 22 of 36
22. Challenge absence claims
No record can mean true absence, failed detection, unsuitable method or missing digitisation.
Working checkpoint — organismal entity: Require explicit effort and detection reasoning. Record location precision, method, effort, preservation and permits before comparing records.
For computing, validate identifiers, parent–child sample links and coordinate uncertainty.
Section 23 of 36
23. Challenge collection bias
Accessible places, seasons and conspicuous taxa are often overrepresented in collections.
Working checkpoint — occurrence evidence: Map effort and metadata completeness beside occurrences. Separate organism observation from sample-derived or sequence-derived detection.
For writing, say whether a record is an observation, specimen, sample or inferred occurrence.
Section 24 of 36
24. Challenge environmental DNA
DNA detection can reflect transport, persistence, contamination or incomplete reference libraries.
Working checkpoint — provenance chain: Report controls and assignment confidence. Preserve original labels and determinations beside every modern interpretation.
For students, ask what physical evidence exists and how it reached the record.
Section 25 of 36
25. Protect sensitive localities
Exact sites for threatened or culturally sensitive material can create harm if disclosed openly.
Working checkpoint — BCO identifier: Apply governed generalisation and access rules. Pin BCO, ENVO and taxonomy releases with validation and migration notes.
For parents, the useful checkpoint is whether a map point has a traceable source and method.
Section 26 of 36
26. Respect permits and communities
Collection and reuse may be constrained by permits, laws, agreements and community expectations.
Working checkpoint — collecting process: Keep governance documentation with the sample. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Section 27 of 36
27. Review ontology and taxon migrations
Updated classes and taxonomy can change labels and grouping without changing the physical specimen.
Working checkpoint — material sample: Maintain a reproducible migration table. Record location precision, method, effort, preservation and permits before comparing records.
For computing, validate identifiers, parent–child sample links and coordinate uncertainty.
Section 28 of 36
28. Connect to Primary Science
A careful schoolyard observation log teaches habitat, classification and respect for living things.
Working checkpoint — organismal entity: Use non-destructive fictional or photographic examples. Separate organism observation from sample-derived or sequence-derived detection.
For writing, say whether a record is an observation, specimen, sample or inferred occurrence.
Section 29 of 36
29. Connect to Secondary Science
Sampling, variation, ecosystems and classification show why collection method shapes evidence.
Working checkpoint — occurrence evidence: Compare two survey methods at the same site. Preserve original labels and determinations beside every modern interpretation.
For students, ask what physical evidence exists and how it reached the record.
Section 30 of 36
30. Connect to mathematics
Sampling units, effort, detection probability and pseudoreplication explain why counts need design.
Working checkpoint — provenance chain: Draw the event-to-sample hierarchy before calculating. Pin BCO, ENVO and taxonomy releases with validation and migration notes.
For parents, the useful checkpoint is whether a map point has a traceable source and method.
Section 31 of 36
31. Connect to computing
Persistent identifiers, graphs, geospatial fields and provenance rules make collections searchable.
Working checkpoint — BCO identifier: Test parent links, coordinates and duplicate origin. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Section 32 of 36
32. Connect to pathways
Collections connect ecology, taxonomy, museums, genomics, conservation and data stewardship.
Working checkpoint — collecting process: Check current official programme details separately. Record location precision, method, effort, preservation and permits before comparing records.
For computing, validate identifiers, parent–child sample links and coordinate uncertainty.
Section 33 of 36
33. Did you know?
The OBO Foundry explicitly places museum collections, environmental and metagenomic samples, and ecological surveys inside BCO’s interoperability goal.
Working checkpoint — material sample: Compare the evidence chain for one example from each domain. Separate organism observation from sample-derived or sequence-derived detection.
For writing, say whether a record is an observation, specimen, sample or inferred occurrence.
Section 34 of 36
34. Write a reader-ready collection note
Name the event, object, derivative, taxon, locality, method, preservation, institution, identifiers, uncertainty and governance.
Working checkpoint — organismal entity: Make the physical evidence route visible. Preserve original labels and determinations beside every modern interpretation.
For students, ask what physical evidence exists and how it reached the record.
Section 35 of 36
35. Build a validation ladder
Check identity, event links, labels, locality, method, taxon, preparation, custody and independent review in sequence.
Working checkpoint — occurrence evidence: Stop the conclusion at the highest verified rung. Pin BCO, ENVO and taxonomy releases with validation and migration notes.
For parents, the useful checkpoint is whether a map point has a traceable source and method.
Section 36 of 36
36. Finish with a reproducible collections bundle
Deliver event records, specimen and derivative IDs, images, mappings, releases, permits, validation, code and limitations.
Working checkpoint — provenance chain: Make every biodiversity point traceable to evidence. Keep encounter, collection event, object, derivative, institution and taxonomic determination linked.
For mathematics, distinguish repeated measurements from genuinely independent collection events.
Contents · Previous section · Continue to the Science Learning Hub