Why Science? | Zebrafish Phenotype Ontology and Comparable Developmental Phenotypes

Zebrafish make development visible: transparent embryos, rapid early growth and well-curated genetics let researchers connect genes with observable changes. The OBO Foundry says the Zebrafish Phenotype Ontology, or ZP, formally defines zebrafish phenotypes and records Monarch as a user of ZP-based phenotype querying. Its official dependency list includes anatomy, spatial, process, quality and relation ontologies. ZFIN explains that its phenotype annotations draw on the Phenotype and Trait Ontology, Zebrafish Anatomy Ontology and Gene Ontology, with evidence curated from figures and direct submissions. These views are complementary: formal ZP classes can organise compositional phenotype meaning, while source records still need anatomy, quality, stage, genotype, image and evidence. This guide supports Primary Science, PSLE Science, Secondary Science, O-Level Science, O-Level Biology, data literacy and STEM education. It is not medical advice, a diagnosis, an animal-use protocol, an admission promise or a career guarantee.

Related eduKate reading: Phenotype and Trait Ontology and computable qualities; Uberon and cross-species anatomy; Mammalian Phenotype Ontology and comparable phenotypes; Evidence and Conclusion Ontology and computable evidence; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.

Primary and official sources checked on 13 October 2026: OBO Foundry Zebrafish Phenotype Ontology overview, checked 13 October 2026; Official Zebrafish Phenotype Ontology repository; ZFIN phenotype annotation explanation; Canonical Zebrafish Phenotype Ontology OWL release; 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 a zebrafish observation becomes comparable developmental evidence

A phenotype is an observed characteristic or state, not simply a gene name. In zebrafish work, the same gene can be associated with different observations across alleles, stages, tissues and experimental conditions.

The OBO Foundry describes ZP as formally defining zebrafish phenotypes. A formal class allows software to recognise that different labels can represent related anatomy–quality combinations.

ZP depends on several other ontologies, including PATO, Gene Ontology, Zebrafish Anatomy, Uberon, BSPO and the Relation Ontology. This modular design keeps anatomy, quality, process and relationship meaning connected rather than copied inconsistently.

ZFIN says it annotates phenotype data using PATO, the Zebrafish Anatomical Ontology and Gene Ontology. Source curation can therefore remain compositional even when downstream systems also use pre-composed ZP classes.

Anatomy and quality do different jobs. The anatomical entity states what was observed; the quality states how it differed, such as size, shape, position, amount or presence.

Developmental stage is essential because an absent structure may be expected before it forms, transient during morphogenesis or abnormal after a defined milestone.

Associated figures are valuable evidence when specimen identity, orientation, scale, stage, imaging method and annotation target remain attached. A cropped image alone can lose these boundaries.

Genotype context includes allele, zygosity, genetic background and intervention. Mutant, knockdown and edited models can have different specificity, timing and off-target risks.

Phenotype similarity can connect zebrafish observations with other species through shared qualities and anatomy mappings. Similarity ranks useful candidates; it does not prove disease identity or causal equivalence.

Hierarchy and logical definitions support broader searches, but results inherit every annotation and mapping decision beneath them. The source figure and curator note remain part of review.

Ontology releases change. Revised dependencies or definitions can alter inferred parents and similarity results even when the original fish records do not change.

The optimistic promise is a developmental evidence map: students and scientists can connect visible changes to anatomy, timing, genotype and comparative questions without erasing the experiment that made the observation.

Begin with a specimen-level record. Keep fish or clutch identity, genotype call, treatment, stage, image and observation linked before producing any summary phenotype.

Use stage names with actual timing and rearing conditions where relevant. Temperature can change developmental rate, so hours post-fertilisation are not always a complete stage description.

Record orientation and laterality. Dorsal, ventral, anterior, posterior, left and right can change what an image or position phenotype means.

A missing structure must be distinguished from a structure outside the field of view, obscured by preparation or not yet expected at the sampled stage.

Quantitative phenotypes should retain values and units. Converting continuous measurements into normal or abnormal categories can hide effect size and overlap.

Report penetrance as numerator and denominator, not only as present. A phenotype observed in part of a clutch may reflect biology, genotype uncertainty or variable conditions.

Use appropriate controls for injected, edited or treated fish. Handling, reagent exposure and background mutations can create observations unrelated to the intended perturbation.

Keep direct observation separate from curator inference and computational mapping. Each can be useful, but their provenance and confidence should remain distinguishable.

For classroom learning, students can annotate a sequence of embryo drawings and discover why anatomy, stage and viewpoint are required before comparing shape.

A reader-ready phenotype summary pairs formal terms with plain language, representative evidence, counts, stage, genotype, comparator and limitations.

Pre-composed ZP classes can make recurring phenotype meanings convenient to query, while post-composed anatomy–quality expressions may preserve source detail. Record which representation and release a dataset uses.

Anatomy mappings across species are hypotheses with scope. A shared parent or homologous structure can support comparison without making two developmental systems identical.

Gene Ontology terms can describe affected processes or functions, but a process annotation and an observed morphology are separate statements with separate evidence.

Phenotype profiles are rarely all-or-nothing. One model can match some features, miss others and show additional effects. Represent the full profile rather than declaring simple validity.

Morpholino, CRISPR and stable mutant studies can differ in timing, specificity, genetic compensation and experimental design. Do not merge them under one gene label without reviewing method.

Imaging pipelines introduce segmentation thresholds, focus, illumination and registration error. Store the raw image and analysis settings before converting a shape into a categorical annotation.

Clutch and tank effects matter. Siblings share parental and environmental context, so statistical models should not treat every larva as an unrelated independent replicate.

Blinded scoring and predefined criteria reduce interpretation drift when researchers classify subtle morphology or behaviour.

Negative phenotype evidence is credible only when the feature was assessed at the appropriate stage with a method sensitive enough to detect the expected change.

Release migration should rerun logical classification, ancestor queries and phenotype-similarity fixtures. Report whether changed results came from dependency updates or new annotations.

Ethics and welfare remain outside the ontology. Structured data never substitutes for appropriate review, humane endpoints or responsible animal care.

The strongest ZP workflow is reversible: every comparison can be traced from a similarity result to the formal class, anatomy, quality, fish, stage, image and curator decision.

A robust zebrafish phenotype matrix should keep the formal class beside the compositional evidence that supports it. Useful columns include fish or clutch, genotype, allele and zygosity, background, intervention, rearing condition, developmental stage, anatomy identifier, quality identifier, ZP identifier where used, image or measurement reference, comparator, penetrance, severity rule, curator, source figure, annotation date and ontology releases. Keeping these fields separate lets a project distinguish a new observation from a new interpretation of an old figure. It also prevents a convenient phenotype label from concealing missing stage or specimen information. When a formal class changes, the source composition and image still provide a route for review and migration.

A good classroom or curation audit begins with one figure and asks students to create two records: a literal observation record and an interpretation record. The literal record describes anatomy, position, stage, measurement and visible quality. The interpretation record proposes genotype relevance, pathway or model-organism significance and labels its evidence. Compare annotations across readers, resolve differences with definitions, and test whether a broader query retrieves the record. Finally, remove the image or stage field and notice how confidence changes. This exercise shows why ZP is not a shortcut around observation. Its formal structure is most valuable when it carries the observation forward, exposes disagreements and allows new comparative questions without rewriting uncertain evidence as fact.

Before publication, an annotation team can run a compact audit across ten records: confirm fish identity, stage, anatomy, quality, genotype, comparator, source figure, curator, ZP representation and dependency releases. Count how many records lack one of those fields and explain how the omission changes possible queries. Then compare a literal source annotation with the broader class returned through reasoning. If the broader result is useful, keep it as a derived view rather than overwriting the source. This preserves the ability to update mappings, challenge an inference and explain the record to a student. Formal phenotype infrastructure earns trust when it makes every transformation visible and reversible.

A four-stage evidence ladder for zebrafish phenotype records

Stage	Question	Evidence	Responsible output
Observe	What changed in which fish and at what stage?	Image, measurement, anatomy, quality, time and comparator	A bounded phenotype observation
Compose	Which formal phenotype meaning represents it?	ZP identifier or anatomy–quality composition and release	A machine-readable phenotype class
Contextualise	Which genotype and experiment produced it?	Allele, zygosity, background, treatment, assay and provenance	Traceable zebrafish evidence
Compare	Which mutant, pathway or cross-species claim is justified?	Statistics, mappings, replication and domain review	A calibrated model-organism conclusion