eduKateSG · Why Science?
Development becomes easier to compare when the structure keeps its stage
Choose the closest question, or read the guide from the beginning. Each chapter keeps its explanation, example and next step visible.
Development is a moving story. A structure has a name, but that name becomes more useful when the observation also tells us which developmental stage was examined. The Zebrafish Anatomy Ontology, or ZFA, helps organise anatomical meaning, while stage information helps readers place evidence in that story. Why learn Science? Because we learn to compare the right things at the right point, rather than letting similar labels hide different questions.
The OBO Foundry describes ZFA as a vocabulary for zebrafish anatomy and development and identifies ZFIN's use in expression and phenotype annotation. Van Slyke and colleagues' 2014 primary publication describes ZFA used with the Zebrafish Stage Ontology, or ZFS, and relationships connecting anatomy to stages. The current ZFIN repository provides the maintained project route. Historical publication facts are not presented as current term counts or promises about a release schedule.
This guide owns stage-aware anatomical structure. The existing eduKate ZP article owns phenotype descriptions; Uberon owns cross-species anatomy; the spatial and term-lookup guides support separate review steps. All stage labels, diagrams, counts and numerical results in the exercises below are fictional teaching material, not real ZFA or ZFS terms or reported zebrafish findings. Learners can practise with paper timelines and published information without conducting animal experiments.
Follow the connected evidence: Zebrafish Phenotype Ontology; Uberon and Cross-species Anatomy; Biological Spatial Ontology; Ontology Lookup Service. For education and school-choice guidance, use the Education Hub; for possible work directions, continue to Careers by Subject Capability. These routes are separate from the specialist topic here.
| Layer | Question | Record | Boundary |
|---|---|---|---|
| Structure | What anatomical entity? | Reviewed ZFA term and source | Not a phenotype by itself |
| Stage | Which developmental context? | Staging vocabulary and evidence | Not elapsed time alone |
| Age | When was it observed? | Recorded duration and reference event | Not an inferred stage |
| Observation | What happened in the specimen? | Image, method, result and comparator | Not every instance of the structure |
Inside this guide
Chapters 1–8 · Meaning and evidence
- 1. Did You Know? Age and stage answer different questions
- 2. ZFA and ZFS contribute different information
- 3. Anatomy across development needs explicit relationships
- 4. The organism and specimen remain part of the record
- 5. A worked stage-aware observation record
- 6. Stage assignment deserves its own evidence note
- 7. Anatomical existence and measured absence are not interchangeable
- 8. Expression location is not a complete functional explanation
Chapters 9–16 · Comparison and learning
- 9. Match the developmental question before matching records
- 10. Work through a fictional detection proportion
- 11. Repeated observations do not create independent specimens
- 12. Distinguish anatomy, quality and experimental condition
- 13. A developmental query should expose its expansion rules
- 14. Use historical publications and current releases for different jobs
- 15. Build a paper developmental timeline with meaningful comparisons
- 16. A final stage-aware evidence check
CHAPTER 1 OF 16 · Connect anatomy and development
1. Did You Know? Age and stage answer different questions
Age describes elapsed time from a stated event. Stage describes a developmental category established by a staging system and its evidence. These fields may be related, but one should not be silently substituted for the other. A recommended record stores both when available and explains how the stage was assigned. This keeps a time measurement separate from a biological interpretation.
Imagine fictional observations taken twelve hours after a local reference event. One record has a reviewed stage label; another has only the time. Assigning the first record's stage to the second would require evidence that is not supplied by the matching timestamp. The example illustrates record design only; it does not claim a real zebrafish timing or stage correspondence.
Students can explore the distinction with two paper-building projects started at the same time. One reaches a defined construction milestone earlier. Their elapsed time matches while their milestone status differs. This analogy is not a biological model, but it makes the logical distinction accessible. Write two sentences: when was it observed, and what stage evidence was recorded? A clearer answer to each question improves the comparison without needing a more complicated calculation.
CHAPTER 2 OF 16 · Connect anatomy and development
2. ZFA and ZFS contribute different information
ZFA names anatomical structures; ZFS supplies developmental stage vocabulary. The 2014 primary publication describes their joint use for zebrafish expression and phenotype records and links that define stages during which anatomical entities exist. Source: Van Slyke and colleagues, 2014, as recorded by ZFIN. Anatomy and stage are therefore complementary, not competing labels for the same field.
For a fictional record, keep the anatomical structure in one column and the stage in another. The observation might concern a region at a particular milestone, with an image and method explaining the evidence. A structure identifier cannot replace the stage, and a stage identifier cannot reveal which structure was examined. Each field reduces a different ambiguity.
Use a simple reading test: can another learner state both the entity and developmental context without guessing? If not, identify which field needs clarification. The vocabulary pair helps organise records, but it does not generate missing observations. A tidy table remains incomplete if its stage was inferred from an unsupported assumption. The useful next move is to preserve the gap and locate the source evidence, making the record more faithful before asking it to support a broad conclusion.
CHAPTER 3 OF 16 · Connect anatomy and development
3. Anatomy across development needs explicit relationships
The 2014 ZFA paper describes subclass, part and developmental organisation. Source: ZFIN's primary-publication record. These relationships answer different questions: what kind of structure, what it belongs to and how developmental descriptions connect. Preserve the relation rather than interpreting every nearby term as a synonym or a stage-independent equivalent.
Consider an invented paper timeline in which a temporary component is replaced by a later component. A “develops into” connection in the exercise is different from “is a kind of” and “is part of”. If all arrows are unlabeled, a reader could mistake an earlier component for a subtype of its later form. Explicit local rules keep the interpretation clear.
For actual term review, inspect the selected release and its definitions and relationships. Do not manufacture a developmental link from similar words. A recommended annotation note states which relationship mattered and what evidence supported the choice. Learners can practise by turning each arrow into a complete sentence and checking whether the sentence follows from the diagram. This builds a habit that transfers into biological data: the link's meaning is essential to the story the record tells.
CHAPTER 4 OF 16 · Connect anatomy and development
4. The organism and specimen remain part of the record
ZFA is zebrafish-specific. Keep the organism identity beside the annotation and retain the actual specimen or group identifier. A term for an anatomical class does not identify the individual fish, material or observation. Combining these layers into one label can make a record difficult to trace, especially when several specimens share the same stage and location.
Our fictional dataset contains Specimens A, B and C, all assigned to a local stage category. Two have images of a selected region; the third has only a broad source note. The shared category makes them discoverable together, but it does not make their evidence equal. A review should preserve the distinct specimen records and the information actually available for each.
Recommended fields include organism, specimen identifier, relevant source metadata, anatomy, stage and observation reference. Use non-sensitive invented records for classroom work. If a professional record uses pooled material, document the pooling unit rather than treating it as a single specimen. The important learning question is “whose observation is this?” Asking it early keeps stage and anatomy attached to a legitimate subject instead of floating as unexplained descriptors in a spreadsheet.
CHAPTER 5 OF 16 · Read a staged observation
5. A worked stage-aware observation record
Build a fictional record for Specimen A with a local structure label, Stage Two, an elapsed duration from a defined starting event and an image identifier. The source note says a marker was detected in the inspected structure. Add the method and the rule used to decide detection. None of these invented fields is a reported ZFIN result or a real ontology identifier.
Read the record in order: organism, specimen, anatomical meaning, stage evidence, age reference, observation and method. Identify what the source actually supports. A marker detected in one inspected region at Stage Two does not establish its presence at every stage or in every specimen. The limited statement can still be scientifically useful because its scope is clear.
Now create a second record with the same structure label but no stage evidence. Keep its annotation review separate. If the comparison requires matched stages, the second record is a candidate needing clarification, not an automatic member of the matched set. This case gives students a practical way to see why metadata matter: the missing field changes which question the observation can answer, even when the anatomical name looks identical.
CHAPTER 6 OF 16 · Read a staged observation
6. Stage assignment deserves its own evidence note
A recommended record explains how a stage label was obtained. Was it supplied by the source, assigned through a documented review or inferred by a local rule? These routes should not be indistinguishable in the final table. Preserve the supporting description and any uncertainty, and avoid pretending that every stage field was directly observed in the same way.
For a fictional classroom example, one image meets two stated milestone features, while another shows only one because part of the drawing is obscured. The first supports the defined local stage. The second may remain uncertain. A reviewer should not assign the complete stage merely because it is the most likely option or because the timestamp resembles other records.
Keep candidate and accepted stages separate if review is incomplete. A short rationale such as “source explicitly states the stage; image unavailable for independent review” communicates a useful limitation. This is an editorial recommendation, not an official ZFIN submission rule. It helps learners distinguish the evidence they possess from evidence they assume exists elsewhere. A stage-aware dataset becomes more reusable when another reader can see both the assigned category and the route that justified it.
CHAPTER 7 OF 16 · Read a staged observation
7. Anatomical existence and measured absence are not interchangeable
A developmental model may describe when a structure is expected to exist. An experiment or image record describes what was assessed in a particular specimen. Keep those claims separate. A term's stage relationship should not be used to generate a negative observation automatically, and an unobserved structure should not be called absent merely because it was outside the evidence's coverage.
In our fictional timeline, Structure X is defined for Stage Three. A Stage Two image has no assessment of X. That record cannot be converted into a measured absence of X. The exercise's model provides one kind of information; the image provides another. If an actual observation appears to conflict with a model, review source stage, anatomical identification and method before altering either.
Use separate status fields for assessed and detected, assessed and not detected, unassessed and unresolved. Define them locally for the learning exercise. This gives an apparent disagreement a route to investigation instead of forcing it into a binary answer. Scientific reasoning becomes productive when it can identify which layer needs checking: stage assignment, structure meaning, image coverage, detection rule or the model's documented boundary.
CHAPTER 8 OF 16 · Read a staged observation
8. Expression location is not a complete functional explanation
The OBO Foundry identifies ZFIN's use of ZFA for expression and phenotype annotation. Source: OBO Foundry ZFA overview. An expression-location record helps describe where evidence was reported. It does not alone establish what a gene does, why a developmental change occurred or whether the same conclusion applies across all contexts. Keep location evidence distinct from a functional interpretation.
For a fictional example, a marker is detected in a region at Stage Two and not detected under the stated rule at Stage Three. The pattern can motivate a question about timing and location. It cannot by itself prove a mechanism or explain the difference. The report still needs specimen context, method comparability and the evidence supporting any proposed causal claim.
Ask learners to separate observation from interpretation in their writing. “Detected in this region under this method” is an observation statement; “causes this developmental outcome” is a stronger claim requiring other evidence. A clear distinction does not make the observation less valuable. It makes the next scientific question visible and prevents a useful anatomy annotation from carrying a conclusion the source did not demonstrate.
CHAPTER 9 OF 16 · Compare developmental evidence
9. Match the developmental question before matching records
Two records can share anatomy while serving different developmental questions. A study may compare specimens at the same stage, follow observations over elapsed time or ask when a structure becomes detectable. Define the intended comparison first. The correct grouping depends on the question, and a convenient stage label should not decide the analysis without explanation.
Suppose a fictional question asks whether a marker is detected in a defined structure at Stage Two. Include records with supported Stage Two assignments and suitable observation evidence. A record at Stage Three may be useful for another question, but it does not belong in the same stage-matched count simply because the structure name matches. An uncertain stage should remain a separate review category.
Write inclusion criteria before summarising the dataset. Keep rejected or unresolved candidates visible with reasons. This makes the comparison reproducible and reduces the temptation to select records after seeing a desired result. Students can practise by sorting the same fictional cards for two different questions, then explaining why the sets differ. The exercise shows that a good dataset is organised around a scientific question, not only around a common keyword.
CHAPTER 10 OF 16 · Compare developmental evidence
10. Work through a fictional detection proportion
Imagine ten fictional specimens with reviewed Stage Two assignments and adequate images of Structure X. A defined marker is detected in six. The observed detection proportion is six divided by ten, or 60%, for this teaching set and its stated method. Two additional records with uncertain stage are listed separately. They are not silently added to the denominator or counted as negative.
If the two uncertain records are both assigned to Stage Two without evidence, the calculation changes. That change reflects an annotation decision, not new observation. A percentage can look precise while relying on unsupported membership in the group. Keep the counts, eligibility criteria and unresolved records visible so another reader can inspect the arithmetic and the meaning of its denominator.
This is a fictional arithmetic exercise, not a biological estimate or ZFIN performance claim. It illustrates why stage-aware description belongs before calculation. Ask students which new information would justify including the unresolved records and whether the detection methods are comparable. The mathematical answer is easy once the counted set is defined; the scientific work is to define that set honestly and explain which records can support the question.
CHAPTER 11 OF 16 · Compare developmental evidence
11. Repeated observations do not create independent specimens
A time series may contain several records for the same specimen. The number of rows is therefore not automatically the number of independent specimens. Preserve specimen identifiers and the relationship between observations. A recommended analysis states whether it counts specimens, images, time points or anatomical assessments, with a rule for repeated observations.
Our fictional Specimen A has four images at four local stages. That is one specimen represented by four observations, not four different specimens. If a report counts detections per image, say so. If it counts whether a specimen ever had a detected marker, define that aggregation. The same rows can support different summaries, but the summaries answer different questions.
Students can demonstrate this using one paper model photographed repeatedly during assembly. The model identity stays constant as its observations change. Keep the image identifiers separate from the model identifier and add the stage record for each photograph. This safe activity connects developmental thinking with data organisation. It shows why clear anatomy and stage annotations need an observation structure that preserves dependence, rather than letting spreadsheet row count stand in for scientific replication.
CHAPTER 12 OF 16 · Compare developmental evidence
12. Distinguish anatomy, quality and experimental condition
A structure is the entity described. A quality records a characteristic of that entity. An experimental condition describes the circumstances under which evidence was obtained. These layers can be connected without being merged into one term. The existing eduKate ZP guide examines phenotype meaning; this ZFA guide focuses on the anatomy and stage that keep such meaning grounded.
For a fictional note, “Structure X is shorter at Stage Two under Condition C” contains several claims. The structure label does not encode the length measurement or comparator. The stage does not encode the condition. The condition does not prove that it caused the difference. Put the layers into explicit fields and preserve the method and reference used for the comparison.
Ask students to colour-code the entity, observation, stage and condition in a sentence. Then remove one part and discuss what becomes unclear. This is a reading exercise as much as a scientific one. It helps learners build explanations whose nouns, quantities and qualifiers remain connected. A well-organised record gives later readers a route from a phenotype statement back to the actual source evidence instead of leaving only a compressed label.
CHAPTER 13 OF 16 · Practise with a timeline
13. A developmental query should expose its expansion rules
A query may retrieve specific anatomical terms beneath a broader category or observations across several stages. These expansions are useful when their rules are visible. Preserve the original annotation on each returned record and explain which relationships the query followed. A broad result set is a set of candidates, not automatic proof that every row describes an interchangeable observation.
In a fictional catalogue, a broad-region search returns two component observations and one whole-region observation at different stages. A stage filter then selects only supported Stage Two records. Show both steps in the query log. If an unresolved stage disappears from the final set, explain why rather than letting its exclusion look like missing source data.
Validate the query against a small expected set that includes a duplicate image, a broader annotation, a later-stage record and an unresolved stage. Check both membership and counting. Computing learners can implement the filter; Science learners can review the inclusion logic. The useful product is an explainable query whose results answer the defined developmental question, with important differences preserved for the next stage of review.
CHAPTER 14 OF 16 · Practise with a timeline
14. Use historical publications and current releases for different jobs
The 2014 paper explains the design and scientific purpose of ZFA and ZFS. The current ZFIN repository provides the project route for inspecting ontology files and reporting concerns. Sources: ZFIN publication record and official repository. A historical term count or example is not a substitute for checking a current definition in the release used by an analysis.
Recommended annotation records retain the chosen release, term identifier, source wording and stage evidence. If a label or relationship changes, review whether the earlier decision still fits. Keep the original analysis history rather than silently replacing it with today's wording. The scientific meaning should remain reconstructable even when the vocabulary becomes more detailed or a source record is clarified.
For a student project, a dated vocabulary snapshot and a short change note may be sufficient. State which change caused a revised result: new stage evidence, a corrected anatomy assignment, a different inclusion rule or updated vocabulary. This makes improvement visible and accountable. It also prevents a common confusion in learning: a different result can come from better information rather than a calculation error or a newly discovered biological effect.
CHAPTER 15 OF 16 · Practise with a timeline
15. Build a paper developmental timeline with meaningful comparisons
Create a fictional model assembled through four locally defined milestones. Photograph or draw the model at each milestone and record elapsed time separately. Label a component that appears only at a later milestone. Students can then practise anatomy-like structure descriptions, stage assignment and observation status without pretending the model is a zebrafish or using living animals.
For Primary Science and PSLE Science learning, ask what was observed and what the picture cannot show. For Secondary Science, add two models built at different rates and compare them by both time and milestone. Write the question before choosing the comparison. A matched timestamp and a matched stage are different designs, and the learner should explain which design answers their particular question.
Use the Science Learning Hub for wider practice and the Education Hub for study planning and school-choice exploration. The career capability page offers a way to reflect on skills such as clear observation, numerical reasoning and data organisation. These links support options without promising a qualification or career outcome. The immediate achievement is a timeline whose evidence and assumptions can be understood by another learner.
CHAPTER 16 OF 16 · Practise with a timeline
16. A final stage-aware evidence check
Prepare one fictional record with structure, specimen, stage, age reference, observation, method and source evidence. Then prepare a second record that shares the anatomy label but lacks stage information. Ask a partner to decide whether both can answer a specified stage-matched question. They should explain the difference rather than merely accept or reject a row.
Finish with an unresolved-case note stating exactly what additional evidence would help. A useful next step might be a source-stage statement, a clearer image or a corrected anatomical definition. Do not fill the gap with a likely answer. The record is complete as a review product when it shows what is known, what is uncertain and how the decision could be improved.
Why Science? Because development becomes more understandable when we keep its moving parts in view. ZFA contributes anatomical meaning; ZFS contributes stage vocabulary; the source supplies the actual observation. A learner who can distinguish these layers can ask better questions, calculate more meaningful proportions and explain why a comparison fits. Begin with one structure and one supported stage, then let the evidence guide the next step.
Back to contents · Previous chapter · Continue to the Science Learning Hub
Sources and scope
Official and primary sources checked on 15 October 2026: OBO Foundry: Zebrafish anatomy and development ontology; ZFIN: official ontology repository; ZFIN: Van Slyke and colleagues, 2014 primary publication; OBO Foundry: Zebrafish Stage Ontology. Project descriptions above are attributed to their named sources. Worked datasets are fictional; record designs, review checklists and teaching activities are editorial recommendations, not official ontology requirements. No search-volume, clinical, admission or career-outcome claim is made.