eduKateSG · Why Science?
A pathway picture becomes scientific evidence when every event can be opened, located and traced
Follow the hierarchy from pathway to reaction, then check participants, cellular compartments, species status and references before interpreting an analysis result.
Reactome is a free, open-source, curated and peer-reviewed pathway database. Its diagrams are attractive, but their scientific value comes from the objects behind the picture: pathways arranged in a hierarchy, reactions with participants, cellular compartments, species context and literature references. When you open those layers, a pathway stops being a colourful poster and becomes a route through evidence.
This guide shows students, parents and educators how to read Reactome without turning a pathway label into an oversized claim. It supports science learning, secondary science enrichment and early bioinformatics practice. The miniature datasets are fictional teaching examples. They illustrate reasoning and arithmetic; they do not report a Reactome analysis or a biological discovery.
Chapter index
Chapters 1–9: from pathway maps to event details
- 1. Start with the biological question, then choose the pathway level
- 2. Read the hierarchy as containment, not a time line
- 3. Open the event instead of interpreting the icon
- 4. Keep physical entities and their forms distinct
- 5. Treat cellular compartment as part of the event
- 6. Trace catalysts and regulators without merging their roles
- 7. Use stable identifiers to keep the route reproducible
- 8. Separate curated human events from inferred species events
- 9. Follow literature references at the reaction level
Chapters 10–17: from analysis overlays to careful reporting
- 10. Understand what over-representation analysis asks
- 11. Check identifier mapping before trusting the ranking
- 12. Read an expression overlay as an overlay
- 13. Distinguish pathway knowledge from dataset interpretation
- 14. Compare pathways without ranking them by size
- 15. Write a pathway claim with its boundaries
- 16. Turn a pathway into a revision route
- 17. Connect pathway literacy to future work
CHAPTER 1 OF 17 · Read the pathway hierarchy
1. Start with the biological question, then choose the pathway level
A broad pathway name can cover many child pathways and individual reactions. Begin by writing the biological question in plain language. Are you asking how a signal starts, how a molecule is transported, or where a product is made? The question determines whether you need the top-level overview, a child pathway or one event. Zooming is not merely visual; it changes the scope of the statement you can support.
Try this
Imagine a fictional classroom question: “How does Signal P lead to Product Q?” A top-level map called “Cell Communication” is too broad. You follow the hierarchy to “P-dependent response” and then to an event converting Intermediate I into Q. Each step narrows the claim and makes the relevant participants easier to inspect.
Write three versions of the question: whole process, subpathway and reaction. Beside each, note what evidence object would answer it. This prevents a familiar pathway name from substituting for a precise explanation.
Evidence checkpoint
Write the question at pathway, subpathway and reaction scale. If the wording stays identical at all three levels, the question is probably still too broad.
CHAPTER 2 OF 17 · Read the pathway hierarchy
2. Read the hierarchy as containment, not a time line
Reactome’s hierarchy organises pathways and subpathways. A child pathway belongs within a broader process, but position in a list or ring does not automatically mean “happens earlier”. Temporal order is better examined through individual events and preceding or following relationships. Treat hierarchy as scope unless the event details support sequence.
Try this
In our fictional map, “Receptor activation” and “Nuclear response” appear as sibling pathways. The display tells you that both belong under the same broader response. It does not, by itself, prove that every receptor event occurs before every nuclear event. Some branches may run in parallel or feed back.
Choose two neighbouring nodes and write one safe sentence: “These are displayed as parts of the same parent pathway.” Then list the extra evidence needed to claim order. Safe sentences build confidence because they are exact, not timid.
Evidence checkpoint
Use hierarchy to describe containment. Claim timing only when event relationships or source evidence support an order, and allow for parallel or feedback branches.
CHAPTER 3 OF 17 · Read the pathway hierarchy
3. Open the event instead of interpreting the icon
The basic unit in the Reactome data model is a reaction or event. Open it and inspect its inputs, outputs, catalysts, regulators and cellular compartment. Icons summarise these objects, but a symbol can be misread when detached from the detail panel. Event-level reading is where a mechanistic statement gains its nouns and verbs.
Try this
Suppose the fictional diagram shows a circle becoming a square. The event details identify circle as `I in cytosol`, square as `Q in nucleus`, Enzyme E as catalyst and Factor R as positive regulator. Now you can describe the event precisely. Before opening the event, “circle becomes square” was only visual shorthand.
For one real event, copy the stable identifier and write a sentence with input, output and compartment. Add catalyst or regulator only if displayed. Do not infer a participant from colour or proximity alone.
Evidence checkpoint
Record the event identifier, input, output, compartment and any displayed catalyst or regulator. Do not translate an unexplained icon into mechanism.
CHAPTER 4 OF 17 · Read the pathway hierarchy
4. Keep physical entities and their forms distinct
Reactome physical entities can be simple molecules, proteins, sets or complexes. The same gene name can appear in different molecular forms or locations. A complex is not interchangeable with one subunit, and a modified protein may participate differently from its unmodified form. Read the entity label and details before collapsing everything to a gene symbol.
Try this
In the fictional event, Protein A is active only as a two-subunit complex in the membrane. Writing “A activates Q” hides the complex and location. A stronger sentence is “the A:B membrane complex participates as the catalyst in this event.” That sentence can later be shortened, but the detailed note protects the reasoning.
Make an entity card with name, type, components, modification and compartment. Leave any unavailable field blank. This card is particularly useful when linking pathway evidence to UniProtKB and reviewed protein records, where sequence and annotation identity become another layer.
Evidence checkpoint
Name whether the participant is a simple molecule, protein, set or complex. Preserve modification and compartment when those details distinguish the biological form.
CHAPTER 5 OF 17 · Open reactions and participants
5. Treat cellular compartment as part of the event
Location can change biological meaning. A molecule in the extracellular space, membrane, cytosol or nucleus is not necessarily playing the same role. Reactome event details can identify cellular compartments, often with links to controlled vocabulary terms. Keep that location in your notes and in any claim where it matters.
Try this
Our fictional signal has two events: P binds a receptor at the plasma membrane, and a regulator enters the nucleus. If you combine them into “P enters the nucleus”, you create a mechanism that the events do not state. The pathway connects consequences; it does not make every participant travel through every step.
Draw two boxes for compartments and place each participant where the event details locate it. Use arrows only for documented transitions. This small spatial check catches many confident but unsupported explanations.
Evidence checkpoint
Draw compartment boxes before arrows. An event chain can connect consequences without implying that every participant physically moves through every compartment.
CHAPTER 6 OF 17 · Open reactions and participants
6. Trace catalysts and regulators without merging their roles
A catalyst helps an event proceed; a regulator changes the event’s activity; an input is transformed or consumed in the event model. These roles are not synonyms. One entity can take different roles in different events, so read the selected reaction rather than assigning a permanent label to the molecule.
Try this
In a fictional example, Enzyme E catalyses I to Q, while R positively regulates E. Saying “R converts I to Q” skips a level. Saying “R is associated with the pathway” is safe but vague. The event structure supports the clearer statement that R regulates an E-catalysed conversion.
Create a four-column table: entity, event, role and source panel. Fill one row at a time. If the role is not displayed, write “participant—role not established here”. Precision makes later revision easier.
Evidence checkpoint
Place catalyst, regulator, input and output in separate columns. If a role is not displayed for the selected event, leave it unresolved.
CHAPTER 7 OF 17 · Open reactions and participants
7. Use stable identifiers to keep the route reproducible
Names can be similar, abbreviated or updated. A Reactome stable identifier gives another reader a durable route to the selected pathway or event. Copy it with the species and title. When a report discusses several levels, keep separate identifiers for the parent pathway and the event that directly supports the sentence.
Try this
Suppose your fictional notes contain `R-HSA-1000` for a pathway and `R-HSA-1001` for one event. Even though these invented identifiers resemble the format, they are not real records. The lesson is to distinguish levels: cite the event when discussing its participants and the pathway when discussing its broader context.
Practise a citation trail: page title, stable identifier, species, access date and one relevant literature reference. This is far more useful than a screenshot with no address.
Evidence checkpoint
Keep parent-pathway and reaction identifiers separate in citations. The reaction supports its participant statement; the pathway supplies the broader context.
CHAPTER 8 OF 17 · Open reactions and participants
8. Separate curated human events from inferred species events
Reactome is centred on curated human pathways and uses computational procedures to infer equivalent events in other species where appropriate. The interface can mark an event as inferred from another species. That status belongs in your interpretation. An inferred pathway is useful comparative evidence, but it is not the same as direct experimental demonstration in the selected species.
Try this
Imagine switching a pathway view from human to a model organism. Most shapes remain, so it is tempting to write “the pathway was experimentally shown in both species”. Instead inspect the species and inference label. A more accurate sentence is “Reactome displays a computationally inferred equivalent in the model organism.”
Write two headings: curated for this species and inferred from another species. Place each event under the appropriate heading. If the status is unclear, keep it out of a species-specific conclusion until checked.
Evidence checkpoint
Check the selected species and whether an event is curated for it or computationally inferred. Preserve that status in every species-specific sentence.
CHAPTER 9 OF 17 · Interpret analysis and species
9. Follow literature references at the reaction level
Reactome’s details can include literature references supporting pathways or reactions. Read the reference in relation to the selected event. A paper linked to a broad pathway may not test every arrow, and a review may summarise evidence differently from a primary experiment. The database gives you a route; the publication provides the study-level context.
Try this
For the fictional P-to-Q event, imagine two references: one reports binding and another reports a cellular response. Do not combine them into a single experiment. In your notes, assign each paper to the part of the event it supports and record organism, material and method.
Use a claim-evidence pair: “Claim: E catalyses I to Q in this event. Reactome object: identifier. Reference checked: citation. Scope observed: assay and organism.” This format teaches evidence without requiring advanced statistics.
Evidence checkpoint
Match each literature reference to the event feature it supports. A pathway-level review does not automatically document every arrow in the diagram.
CHAPTER 10 OF 17 · Interpret analysis and species
10. Understand what over-representation analysis asks
Reactome analysis tools can test whether a submitted list contains more members of a pathway than expected under a reference background. That is a list-composition question, not proof that the pathway is active or causal. Results depend on identifier mapping, the submitted list, the background and statistical procedure. Inspect those inputs before interpreting a ranked pathway.
Try this
Take a fictional list of twenty genes. Five map to Pathway A, so 5/20, or 25%, of the submitted list is in that pathway. That fraction alone is not a significance test. You still need the background frequency and the tool’s statistical result. If mapping drops three identifiers, the effective denominator for mapped analysis is not automatically twenty.
Write the analysis question in one line: “Is this pathway represented more than expected in my mapped list?” This wording prevents the leap to “this pathway caused the phenotype.”
Evidence checkpoint
Phrase enrichment as over-representation in a mapped list. It does not directly measure pathway activity, flux, causality or clinical importance.
CHAPTER 11 OF 17 · Interpret analysis and species
11. Check identifier mapping before trusting the ranking
Analysis begins with mapping submitted identifiers to Reactome entities. Synonyms, obsolete identifiers, species mismatches and ambiguous mappings can change the usable list. Read the mapping summary and inspect unmatched items. A beautiful result chart cannot recover information that failed at the first step.
Try this
In a fictional upload, 20 identifiers produce 17 mapped genes, one ambiguous mapping and two unmatched strings. Record all three counts. Replacing the unmatched strings with guessed symbols would be a new analysis decision, not a harmless cleanup.
Build a mapping ledger with submitted value, mapped entity, mapping status and any correction. Keep the original list unchanged. This ledger also links naturally to identifier translation work elsewhere in the eduKate ecosystem.
Evidence checkpoint
Report submitted, mapped, ambiguous and unmatched identifiers. Corrections belong in a mapping ledger so the analysis can be repeated.
CHAPTER 12 OF 17 · Interpret analysis and species
12. Read an expression overlay as an overlay
The Pathway Browser can overlay user-supplied expression data onto pathways. The colours represent values you supplied or derived; they do not become curated pathway facts. Check the legend, scale, conditions and missing-value treatment. A coloured event may contain several entities, so inspect how the overlay is aggregated or displayed.
Try this
Imagine values of 2, 0 and -1 on three participants. A red event icon might summarise only part of that information. State “the submitted expression overlay colours this event under the chosen scale,” then inspect entity-level values. Do not write “Reactome proves the reaction increased.”
Save the input file, legend settings and exported result together. Reopening only the image later can erase the scale that gave the colours meaning.
Evidence checkpoint
Save the overlay input and legend with the image. Colours are properties of the submitted values and scale, not permanent Reactome annotations.
CHAPTER 13 OF 17 · Report pathway evidence
13. Distinguish pathway knowledge from dataset interpretation
Reactome’s curated knowledge describes events and relationships; your dataset supplies observations that may be mapped onto that knowledge. Keep these layers separate in writing. The pathway helps organise a hypothesis. The dataset provides measurements. An analysis connects them under explicit rules. None of the three should silently replace the others.
Try this
For our fictional case, Reactome describes an E-catalysed event. A class dataset shows higher RNA abundance for E after treatment. That does not directly measure catalytic activity. It may motivate a next experiment, but the conclusion should preserve the measurement type.
Use three labels in your report: database statement, dataset observation and interpretation. If a sentence contains material from two labels, cite both sources and explain the connection.
Evidence checkpoint
Label database knowledge, dataset observation and interpretation. A stronger report lets the reader see exactly where one layer ends and the next begins.
CHAPTER 14 OF 17 · Report pathway evidence
14. Compare pathways without ranking them by size
Large pathways contain more entities and may attract more submitted identifiers. Small pathways can be more specific. A fair comparison considers the statistical output, mapping coverage, hierarchy and biological question rather than raw hit count alone. Related pathways may also share genes, so several ranked results need not represent independent mechanisms.
Try this
Suppose Pathway A has 20 hits among 200 members and Pathway B has 6 hits among 12 members. Raw count favours A; submitted-list proportion within each pathway tells another story; formal analysis adds the reference background. Do not declare a winner from one number.
Create a comparison card with pathway size, mapped hits, adjusted result if provided, parent pathway and shared identifiers. This makes the trade-offs visible and encourages calm interpretation.
Evidence checkpoint
Compare pathway size, mapped hits, hierarchy and shared identifiers before ranking results. Large hit counts can reflect large or overlapping pathways.
CHAPTER 15 OF 17 · Report pathway evidence
15. Write a pathway claim with its boundaries
A strong claim names the object, species, evidence layer and limit. For example: “In Reactome’s human pathway view, event X lists A and B as inputs and C as output; our expression overlay highlights two participants, but it does not measure event flux.” This sentence is longer than “the pathway is activated”, yet far more useful.
Try this
Take one draft conclusion and underline every verb. Ask whether the database, dataset or your interpretation supports each verb. Replace “proves” with a precise action such as lists, links, maps, enriches, overlaps or suggests. Precision can make a report sound more confident because the reader knows exactly what was done.
Add the stable identifier and access date near the claim. If the event was inferred for the selected species, say so. If you followed a linked paper, cite that paper separately.
Evidence checkpoint
Underline every causal verb and identify its source. Replace unsupported activation language with lists, maps, overlaps, enriches or suggests when those are the actual operations.
Reactome can support revision when students use it as a sequence of questions. What is the parent pathway? Which reaction am I viewing? What enters and leaves? What catalyses or regulates it? Where does it occur? Which species is shown? Which paper supports it? This route connects vocabulary, diagrams and evidence.
Try this
Choose a familiar topic such as signalling or metabolism. Open one child pathway and one event. Explain them to a partner without showing the diagram, then let the partner reconstruct a simple sketch from your words. Missing arrows or compartments reveal where the explanation needs work.
This is a productive use of science tuition in Singapore: the tutor helps the student move from memorising a pathway picture to interrogating a structured scientific record. The aim is transferable reasoning, not a database tour for its own sake.
Evidence checkpoint
Explain one event without the diagram, then let a partner reconstruct it. Missing participants or compartments show where your verbal model needs repair.
CHAPTER 17 OF 17 · Report pathway evidence
17. Connect pathway literacy to future work
Pathway reading combines biology, controlled vocabularies, data analysis, software interfaces and scientific writing. Those skills appear in bioinformatics, pharmacology, systems biology, research software, data curation and health research. Students do not need to choose a career now; they can notice which task they enjoy—mechanism, coding, evidence review or explanation.
Try this
Your next step can be small: select one Reactome pathway, open a single event and produce a one-page evidence card with identifier, species, participants, compartment and one reference. Then compare that card with your textbook diagram. The differences are often where learning becomes memorable.
Continue through the Science Learning Hub, explore wider education choices, and use the careers-by-subject guide to see how pathway literacy connects to future options.
Evidence checkpoint
Build one evidence card before exploring another pathway. Depth on a single event develops more transferable skill than a gallery of unexplained screenshots.
Contents · Previous chapter · Continue at the Science Learning Hub
A final event-level reading routine
Before closing the Pathway Browser, choose one event that directly supports your explanation. Record its stable identifier, species, inputs, outputs, compartment, catalyst or regulator when shown, and one supporting reference. If your result came from an uploaded gene list, attach the mapping summary and state whether the result was over-representation, topology-based analysis or an expression overlay. This short record separates curated pathway knowledge from the interpretation of your own data.
A useful final sentence follows a simple pattern: Reactome displays X for species Y; my dataset contributes Z under the stated mapping and analysis settings; together they suggest Q; they do not establish R. Filling those slots with exact objects produces a careful conclusion without draining the excitement from discovery.
| Reading object | What it can support | Question to keep visible |
|---|---|---|
| Pathway | An organised set or hierarchy of biological events | Am I reading a broad process or a specific child pathway? |
| Reaction or event | Inputs, outputs, catalysts, regulators and location | Which event supports the statement? |
| Physical entity | A molecule, set or complex participating in an event | Is the participant form and compartment the one I mean? |
| Analysis result | An interpretation of a submitted identifier list | What universe, mapping and statistical question produced this result? |
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
- Reactome: what is Reactome — accessed 24 October 2026.
- Reactome Pathway Browser user guide — accessed 24 October 2026.
- Reactome Details Panel user guide — accessed 24 October 2026.
- Reactome documentation on computationally inferred events — accessed 24 October 2026.