eduKateSG · Why Science?
Cross-species comparison works when the gene, organism, relationship and evidence type stay visible
Use the Alliance as a comparative map: begin with the exact gene record, inspect orthology separately from phenotype or disease evidence, and follow links back to organism-specific knowledgebases.
The Alliance of Genome Resources brings comparative information from major model-organism knowledgebases and the Gene Ontology resource into a shared portal. Its value is not that every species becomes interchangeable. Its value is that a reader can move across gene identity, orthology, function, expression, phenotype and disease information while keeping the organism and evidence route visible.
This guide is for students and families curious about how experiments in yeast, worms, flies, fish, frogs, mice and rats can inform biological questions. It uses the Alliance as a comparison map, not as a source of medical advice. Fictional genes and classroom tables are labelled as such. Live resource claims are tied to the official portal, Alliance papers and open-data documentation accessed on 24 October 2026.
Chapter index
Chapters 1–9: from gene identity to comparative views
- 1. Begin with the exact gene and organism
- 2. Read the gene page as an integration hub
- 3. Understand orthology as an evolutionary relationship
- 4. Treat stringency as an analysis choice
- 5. Compare function annotations with their evidence
- 6. Keep expression tied to anatomy and stage
- 7. Separate alleles and variants from genes
- 8. Read phenotype terms as descriptions, not diagnoses
- 9. Distinguish disease associations by route
Chapters 10–17: from model-organism evidence to scoped conclusions
- 10. Follow the source knowledgebase link
- 11. Keep human relevance separate from human equivalence
- 12. Compare one evidence layer at a time
- 13. Work a fictional cross-species evidence chain
- 14. Track missing data without turning it into absence
- 15. Cite a comparative claim so another reader can rebuild it
- 16. Use the Alliance as a science-learning bridge
- 17. Connect comparative genomics to future pathways
CHAPTER 1 OF 17 · Find the gene and organism
1. Begin with the exact gene and organism
A gene symbol is meaningful only with its organism and identifier. The same letters can name different genes, and related genes may have different symbols across species. Start on a gene page and record the species, primary symbol, full name and source identifier. Do this before opening disease, expression or orthology panels.
Try this
Imagine a fictional mouse gene `Luma1` and a zebrafish gene `luma1a`. Similar names suggest a possible relationship but do not establish it. Your first notes should say “mouse Luma1—fictional” and “zebrafish luma1a—fictional”, each with a separate identifier. Only then should you inspect an orthology view.
Create a gene identity card with organism, symbol, identifier, name and source knowledgebase. If any field is unclear, pause. A cross-species comparison built on the wrong gene cannot be rescued by a sophisticated chart.
Evidence checkpoint
Write organism and primary identifier beside every symbol. Cross-species reasoning fails quickly when a familiar symbol is attached to the wrong record.
CHAPTER 2 OF 17 · Find the gene and organism
2. Read the gene page as an integration hub
An Alliance gene page brings several data types into one view. It may connect function, orthology, expression, alleles, phenotype, disease and interactions. These panels answer different questions and can draw on different evidence. The page is a hub, not a single experiment. Keep the panel name beside every copied statement.
Try this
For fictional Luma1, suppose the function panel has Gene Ontology annotations, the expression panel shows embryonic tissue records and the disease panel lists an orthology-based association. A sentence that combines all three as “Luma1 causes disease in the embryo” would overreach. Each panel needs separate reading.
Make a page map with one box per panel. Under each box, write “object”, “evidence type” and “source”. This turns a long page into manageable scientific questions.
Evidence checkpoint
Label each panel by data type and source. A gene page integrates evidence; it is not a single experiment with one universal method.
CHAPTER 3 OF 17 · Find the gene and organism
3. Understand orthology as an evolutionary relationship
Orthologs are genes related through speciation, but orthology is inferred rather than directly watched through evolutionary time. Different algorithms, evidence combinations and stringency settings can produce different relationship sets. An ortholog is a strong comparison lead; it does not guarantee identical function, expression or phenotype.
Try this
In the fictional example, mouse Luma1 may have two zebrafish co-ortholog candidates after a duplication. Choosing the one with the most familiar name would be unsafe. Inspect the orthology method or stringency and keep the one-to-many relationship visible.
Write a sentence using “is listed as an ortholog under…” and include the selected setting. Then add: “This relationship motivates comparison; it does not by itself establish conserved function.” This is precise comparative science.
Evidence checkpoint
Report orthology as an inferred evolutionary relationship under a named method or setting. Do not translate it into guaranteed functional equivalence.
CHAPTER 4 OF 17 · Find the gene and organism
4. Treat stringency as an analysis choice
Some Alliance views allow ortholog comparisons at different stringency levels. A stringent set may prioritise relationships supported more consistently, while a broader set can recover additional candidates. Neither setting should be hidden. The setting changes the question from “which relationships are most strongly agreed?” to “which relationships might be worth exploring?”
Try this
Suppose the fictional strict view shows one fish candidate and the broader view shows three. Report both counts and the setting used. Do not say that two genes appeared biologically only after changing a menu; the interface exposed a different inference set.
Create a two-row comparison with stringent and broader settings. Add candidate count, shared candidates and interpretation. This teaches how parameter choices shape a result without implying that one menu option is universally correct.
Evidence checkpoint
Record the stringency setting and candidate count. A broader menu choice exposes a different inference set; it does not create new biology.
CHAPTER 5 OF 17 · Compare orthology and function
5. Compare function annotations with their evidence
Gene Ontology annotations describe molecular function, biological process and cellular component. They can be supported by experiments, computational inference or curator statements. Similar ontology terms across orthologs make comparison easier, but the evidence codes and source species still matter. Shared annotation is not automatically shared direct experiment.
Try this
For fictional Luma1, imagine the mouse record has an experimentally supported function term while the fish record receives a related annotation by sequence similarity. A safe sentence distinguishes them: “The records share related functional annotation, with direct experimental support shown for mouse and inferred support shown for fish.”
Build a function table with gene, term, aspect, evidence and source. If the term is broad, do not rewrite it as a detailed mechanism. Controlled vocabulary gives consistency, not unlimited specificity.
Evidence checkpoint
Compare function terms with their aspect, evidence and source species. Related ontology labels can hide different evidence routes.
CHAPTER 6 OF 17 · Compare orthology and function
6. Keep expression tied to anatomy and stage
Expression records depend on tissue, cell type, developmental stage, assay and result. Comparing only a coloured summary can erase those conditions. Open the underlying annotations where available and keep anatomy and stage beside the gene. “Expressed in brain” can hide very different developmental windows and anatomical granularity.
Try this
Imagine mouse Luma1 is reported in adult retina while zebrafish luma1a is reported in an embryonic eye field. Both involve visual anatomy, but the stages and structures differ. The records support a comparison question, not the conclusion “same expression pattern”.
Create an expression card with species, gene, anatomy, stage, assay and result. Link to GXD and stage-specific mouse expression evidence when the mouse developmental context needs deeper inspection.
Evidence checkpoint
Keep anatomy, developmental stage, assay and result together. Similar organ names at different stages do not establish an identical expression pattern.
CHAPTER 7 OF 17 · Compare orthology and function
7. Separate alleles and variants from genes
A gene is a genomic entity; an allele or variant is a particular change or state associated with it. Phenotypes often belong to a genotype involving a specific allele, zygosity or combination, not to the abstract gene in every context. Keep the allele identifier and genotype visible when using phenotype evidence.
Try this
In a fictional fly study, allele `luma1-x1` produces a wing phenotype only when homozygous. Writing “the Luma1 gene causes abnormal wings” discards allele and genotype. A stronger note says that the observed phenotype is associated with the specified genotype in the cited experiment.
Draw a chain: gene → allele → genotype → observed phenotype. Place the evidence source under the final link. This keeps a gene-level summary from swallowing the experimental design.
Evidence checkpoint
Trace gene, allele, genotype and phenotype as separate objects. An observed genotype effect does not describe every state of the gene.
CHAPTER 8 OF 17 · Compare orthology and function
8. Read phenotype terms as descriptions, not diagnoses
Phenotype ontologies standardise observed features so records can be compared. A term describes an observation in an organism or model. It does not automatically diagnose a human condition or establish causality. Read the term, organism, genotype, assay and evidence together.
Try this
Suppose a fictional mouse model has “reduced activity” and a fish model has “abnormal movement”. The terms are related in everyday language but may come from different ontologies and assays. Do not merge them without checking mappings and definitions.
Use a phenotype comparison table with exact term, ontology, experiment and your broader category. Keep your category in a separate column so readers can distinguish database language from your interpretation.
Evidence checkpoint
Use exact phenotype terms and keep your broader grouping in a separate column. Database vocabulary and student interpretation should remain distinguishable.
CHAPTER 9 OF 17 · Inspect phenotype and disease
9. Distinguish disease associations by route
Disease associations can be supported directly, through experimental models, by orthology or by curated literature. Interfaces may show qualifiers that explain the route. That qualifier is essential. “Associated with disease” is too broad when one record represents a human gene finding and another represents a model-organism phenotype connected through orthology.
Try this
For fictional Luma1, imagine the human gene has a curated literature association while the mouse gene appears through orthology. A correct report names the human association and says the mouse connection is inferred via orthology. It does not label the mouse as a patient model without experimental evidence.
Write the association as a sentence with subject, disease term, relation or qualifier, evidence and source. If one element is missing, say what the page displays and avoid supplying the missing interpretation.
Evidence checkpoint
Preserve qualifiers such as direct evidence, experimental model or orthology-based association. They explain how a disease link was constructed.
CHAPTER 10 OF 17 · Inspect phenotype and disease
10. Follow the source knowledgebase link
The Alliance harmonises views, while organism-specific knowledgebases retain deep community curation and context. Follow the source link when a claim needs assay details, literature notes, allele nomenclature or species-specialist interpretation. The two resources work together: the Alliance supports comparison; the source knowledgebase supports depth.
Try this
In the fictional workflow, the Alliance points you from mouse Luma1 to an MGI record. The source page clarifies the allele and experiment. Cite the Alliance for the comparative route and the source record for the detailed evidence. Do not describe information as Alliance-native if you obtained it only after following the source.
Create a two-source note with “comparison view” and “organism detail”. This is a simple provenance pattern for integrated databases.
Evidence checkpoint
Follow the organism-specific knowledgebase when the claim needs allele, assay or literature detail. Cite the comparative route and the detailed source separately.
CHAPTER 11 OF 17 · Inspect phenotype and disease
11. Keep human relevance separate from human equivalence
Model organisms are valuable because shared biology can make experiments informative. They are not small humans. Differences in gene duplication, anatomy, development, physiology and environment can limit transfer. Cross-species evidence supports hypotheses about human biology; it does not remove the need for human evidence.
Try this
Suppose a fictional worm gene affects cell division and has a human ortholog. This supports investigating the conserved process. It does not show that a worm phenotype predicts a patient outcome. The level of conservation and the measurement need to be specified.
Write two columns: “conserved question” and “species-specific caution”. Fill both before drafting a human-relevance sentence. This makes optimism and restraint work together.
Evidence checkpoint
State the conserved question and a species-specific caution together. Model-organism relevance is valuable without implying human equivalence.
CHAPTER 12 OF 17 · Inspect phenotype and disease
12. Compare one evidence layer at a time
A cross-species table becomes confusing when one row is expression, another is phenotype and a third is a pathway prediction. Compare like with like first. Align orthology relationships, then function terms, then expression contexts, then phenotype or disease evidence. Integrate the layers only after each one is understood.
Try this
For fictional mouse, fish and fly genes, start with an identity and orthology table. Build separate sheets for function and phenotype. If the fly record lacks a phenotype annotation, write “not found in the inspected view” rather than “no phenotype”.
Use consistent columns and include a source link for every row. A neat table is only as reliable as its provenance.
Evidence checkpoint
Build separate identity, orthology, function, expression and phenotype tables before integrating them. This reveals which evidence layer drives the conclusion.
CHAPTER 13 OF 17 · Build a careful cross-species claim
13. Work a fictional cross-species evidence chain
The classroom question is: “Could Luma1 participate in visual-system development across vertebrates?” The fictional Alliance view lists mouse Luma1 and zebrafish luma1a as ortholog candidates. Mouse expression is adult retina; fish expression is an embryonic eye field. Function annotations overlap at a broad process term but have different evidence routes.
Try this
This supports a comparative hypothesis: related genes may participate in visual-system biology. It does not establish identical developmental roles. The difference between adult retina and embryonic eye field is especially important. A next step might inspect stage-matched data or primary experiments.
Write the conclusion in four clauses: identity, orthology setting, comparable evidence, limitation. This structure is clear enough for a school report and disciplined enough to scale into research.
Evidence checkpoint
In the fictional chain, treat related expression and function as a hypothesis lead. Stage mismatch remains a material limit until checked with comparable evidence.
CHAPTER 14 OF 17 · Build a careful cross-species claim
14. Track missing data without turning it into absence
An empty panel can reflect coverage, curation status, filtering or true lack of known records. Report the interface state you observed and the date. Search the source knowledgebase and relevant literature before making a biological absence claim. “No record found” and “feature absent” are different statements.
Try this
Imagine the fish phenotype panel is empty. You can say that no phenotype association was found in the inspected Alliance view on 24 October 2026. You cannot conclude that altering the gene produces no phenotype in fish.
Add a status code to your table: present, not found, not checked or not applicable. This makes uncertainty sortable and prevents blank cells from acquiring accidental meaning.
Evidence checkpoint
Code blank cells as present, not found, not checked or not applicable. Empty space should never silently become biological absence.
CHAPTER 15 OF 17 · Build a careful cross-species claim
15. Cite a comparative claim so another reader can rebuild it
A reproducible comparative claim includes each gene identifier, species, relationship setting, evidence panels and access date. Add source knowledgebase records or papers for the substantive biological statements. Screenshots can illustrate the route, but identifiers and URLs should carry the citation.
Try this
For a three-species comparison, give each species its own source line. Then cite the Alliance orthology view used to connect them. If the interface changes, the identifiers still let another reader rebuild the comparison.
Ask a partner to reproduce one row without your help. If they select a different gene or evidence type, revise the record identity or panel description.
Evidence checkpoint
Include gene identifiers, species, orthology setting, panel names, source links and access date so another reader can rebuild the comparison.
CHAPTER 16 OF 17 · Build a careful cross-species claim
16. Use the Alliance as a science-learning bridge
The Alliance makes curriculum ideas—inheritance, variation, evolution, gene function and experimental models—visible in a real information system. A tutor can ask a student to move from one organism to an ortholog, compare one ontology term and explain one limit. This is more valuable than collecting many database facts.
Try this
Choose a gene connected to a familiar topic. Build a one-page comparison between two species using identity, orthology, one function annotation and one evidence difference. Keep disease interpretation out unless the lesson explicitly covers the evidence routes.
The activity suits secondary science and science enrichment because it rewards vocabulary, careful reading and argument structure without requiring advanced laboratory access.
Evidence checkpoint
Use one gene pair to connect inheritance, evolution and evidence. A precise two-species explanation teaches more than a long untraceable list.
CHAPTER 17 OF 17 · Build a careful cross-species claim
17. Connect comparative genomics to future pathways
Alliance-style work combines evolutionary biology, genetics, ontologies, data engineering, curation and scientific communication. These capabilities appear in genomics, model-organism research, bioinformatics, biotechnology and research software. Students can notice whether they enjoy organism detail, comparison logic, coding or explanation.
Try this
Your next step is deliberately modest: compare two ortholog candidates, inspect one function annotation and follow one source knowledgebase link. End with a sentence about what the comparison suggests and a sentence about what it does not establish.
Continue through the Science Learning Hub, use the Education Hub for wider school choices, and explore careers by subject capability when you want to connect comparative reasoning to future options.
Evidence checkpoint
Connect the task you enjoyed—organism detail, comparative logic, data curation or writing—to a next subject choice, not a guaranteed occupation.
Contents · Previous chapter · Continue at the Science Learning Hub
A final two-species comparison routine
A reliable Alliance comparison can begin with only two species. Record the primary gene identifier and source knowledgebase for each. Inspect the orthology relationship and its stringency or method context. Then choose one shared evidence layer—function, expression or phenotype—and compare like with like. A small, balanced table is easier to audit than a wide table mixing every panel on the page.
Next, identify the route of each association. A human disease link, an experimental model annotation and an orthology-based connection are not interchangeable. Preserve qualifiers and follow the organism-specific knowledgebase when the claim needs allele, genotype, assay or literature detail. The Alliance supplies the comparative bridge; specialist records provide much of the depth.
Finally, write one conservation sentence and one caution sentence. The first names the relationship or comparable evidence that motivates the hypothesis. The second names the species, stage, anatomy, genotype or evidence difference that limits transfer. Holding those sentences together is the heart of responsible model-organism reasoning.
For a school project, stop there and invite a reproducibility check: can a classmate use your identifiers, settings and links to reach the same two records? If yes, you have built a real evidence route. If not, improve the route before adding another species.
If you later add a third organism, do not copy the first interpretation across it. Repeat the identity, orthology and evidence checks from the beginning. Gene duplication, developmental timing and data coverage can change the comparison. This deliberate reset keeps comparative genomics curious and fair: every species contributes evidence in its own context before joining the wider pattern. Record the new source knowledgebase and date, compare only the same evidence layer, and revise the caution sentence before expanding the conclusion.
Keep a final audit row for unresolved differences. One organism may have rich phenotype curation while another has stronger expression coverage; that imbalance reflects the inspected resources as well as biology. Reporting the difference openly helps a reader choose the next source or experiment instead of mistaking uneven data coverage for a species-level conclusion.
| Reading object | What it can support | Question to keep visible |
|---|---|---|
| Gene record | Species-specific identity, symbol and integrated annotations | Have I named the organism and record identifier? |
| Orthology view | A proposed evolutionary relationship between genes | Which method or stringency produced this relationship? |
| Phenotype or disease association | Curated links between genes, alleles, models and observed terms | Is the association direct, inferred through orthology or supported by a model? |
| Source knowledgebase link | A route to organism-specialist curation | Does the source page add experimental or literature context needed for my claim? |
Sources and next reading
Substantive resource claims in this article were checked against the primary pages below. A database record is a structured route into evidence; it does not by itself prove a new biological conclusion. Recheck the live record, evidence labels and release context before using it in assessed work or research.
Primary sources
- Alliance of Genome Resources portal — accessed 24 October 2026.
- Alliance of Genome Resources Consortium 2023 overview — accessed 24 October 2026.
- Alliance portal paper: unified model-organism research platform — accessed 24 October 2026.
- Alliance open-data documentation — accessed 24 October 2026.