eduKateSG · Why Science?
Open the panel, read its version and evidence colour, then ask whether that exact panel belongs to the biological and clinical question in front of you
Choose the closest route, then return to the full index whenever you need the wider evidence chain.
Science learning becomes powerful when students can tell a measured pattern from a model, a statistical decision from a mechanism, and a useful lead from a finished conclusion. Genomics England PanelApp is a publicly available knowledgebase for creating, storing, reviewing and querying virtual gene panels related to human disorders. Expert contributions and curation support a traffic-light evidence system, panel versioning and documented modes of inheritance. The official site also distinguishes the broad PanelApp knowledgebase from the signed-off NHS Genomic Medicine Service panels resource. A green entity on a versioned panel supports analysis within that panel’s scope; it does not mean every variant in that gene is pathogenic, every technology assays it equally well or the panel is automatically the current ordered NHS test. This guide supports Primary Science, PSLE Science, Secondary Science, O-Level Science 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: ClinGen gene–disease validity evidence framework; ClinVar variant classifications and evidence conflicts; Human Phenotype Ontology and structured clinical phenotypes; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.
Primary and official sources checked on 11 October 2026: Official Genomics England PanelApp; PanelApp Handbook V105; PanelApp consensus diagnostic gene panels study; NHS National Genomic Test Directory; 2026 Singapore–Cambridge O-Level Biology syllabus; MOE G2/G3 Lower Secondary Science teaching and learning syllabus; 2026 MOE G2 Computing syllabus. Resource interfaces and recommendations can change, so preserve the release, access date and documentation used.
| Stage | Question | Evidence | Responsible output |
|---|---|---|---|
| Frame | Which phenotype or clinical indication is the panel meant to cover? | Panel description, relevant disorders and version | A scoped panel question |
| Review | Why is each gene or genomic entity included? | Traffic-light rating, expert reviews, evidence and inheritance | A traceable inclusion decision |
| Apply | Which signed-off version and assay context are relevant? | Panel type, release, test directory and technical limits | A bounded analysis set |
| Validate | What case-level evidence supports a finding? | Variant, segregation, phenotype and independent review | An interpretation, not a panel-derived diagnosis |
Inside this guide
1–12 · Foundations and design
- 1. Start with a virtual gene panel
- 2. Understand what PanelApp is
- 3. Separate PanelApp from the NHS GMS panels resource
- 4. Read the panel description
- 5. Read panel type
- 6. Treat a super panel as a composition
- 7. Inspect panel status and version
- 8. Understand the traffic-light model
- 9. Read green carefully
- 10. Read amber as uncertainty
- 11. Read red as low evidence, not impossibility
- 12. Read mode of inheritance
13–24 · Models and estimation
- 13. Include STRs and CNVs when relevant
- 14. Read expert reviews, not colour alone
- 15. Count independent families correctly
- 16. Keep phenotype fit visible
- 17. Distinguish gene validity from variant pathogenicity
- 18. Respect assay limitations
- 19. Use dates and release notes
- 20. Audit additions and removals
- 21. Avoid panel shopping
- 22. Challenge a panel that is too narrow
- 23. Challenge a panel that is too broad
- 24. Use the API reproducibly
25–36 · Diagnostics and interpretation
- 25. Respect attribution and disclaimer
- 26. Protect patient privacy
- 27. Use bounded clinical language
- 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 result
- 35. Build a validation ladder
- 36. Finish with a reproducible bundle
Section 1 of 36
1. Start with a virtual gene panel
A virtual panel is a selected set of genes and other genomic entities used to focus analysis on a defined disorder area or indication.
Working checkpoint — virtual gene panel: Write the panel question and intended use before opening the gene list. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Section 2 of 36
2. Understand what PanelApp is
PanelApp stores public, versioned panels and invites expert review; the resource is a knowledgebase rather than a sequencing assay.
Working checkpoint — panel version: Separate the curated list from the laboratory technology that will interrogate it. Compare the displayed rating with its expert reviews, disease scope and current signed-off service version.
For future study, connect panel curation to biology, data science and ethics while checking current official pathways separately.
Section 3 of 36
3. Separate PanelApp from the NHS GMS panels resource
The official site says PanelApp can show newer working versions, while the NHS GMS Panels Resource shows approved versions in current service use.
Working checkpoint — green evidence rating: Use the resource that matches the operational question. Preserve every panel tested and every excluded entity; a shortlist without selection history is not auditable.
For scientific writing, name the panel and version whenever a colour rating or gene list is reported.
Section 4 of 36
4. Read the panel description
Panel names can look similar while descriptions, relevant disorders and clinical indication codes define different scopes.
Working checkpoint — mode of inheritance: Save the description and indication with the result. Separate panel membership from variant pathogenicity, assay coverage and patient-level phenotype fit.
For students, ask: Which rule placed this entity in the set, and could a later evidence update change that decision?
Section 5 of 36
5. Read panel type
Rare Disease 100K, GMS, research, external-laboratory and super-panel labels describe different origins and uses.
Working checkpoint — expert review: Do not merge panels merely because their titles overlap. Treat entity type, inheritance label and version as scientific data: a silent mismatch can alter the candidate set.
For parents, a calm checkpoint is to ask what the panel can focus, what it can miss and what still needs independent review.
Section 6 of 36
6. Treat a super panel as a composition
A super panel is assembled from component panels and can change when those components change.
Working checkpoint — clinical indication: Record both super-panel and component versions. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Section 7 of 36
7. Inspect panel status and version
Version 0 panels remain open to review, while version 1 or higher signals that an initial expert review and curation cycle has been completed for defined uses.
Working checkpoint — virtual gene panel: Never strip the version from an exported panel. Compare the displayed rating with its expert reviews, disease scope and current signed-off service version.
For future study, connect panel curation to biology, data science and ethics while checking current official pathways separately.
Section 8 of 36
8. Understand the traffic-light model
PanelApp rates the evidence connecting an entity with the phenotypes covered by that panel. Colour is contextual, not a universal property of the gene.
Working checkpoint — panel version: Read the rating on the exact panel and version. Preserve every panel tested and every excluded entity; a shortlist without selection history is not auditable.
For scientific writing, name the panel and version whenever a colour rating or gene list is reported.
Section 9 of 36
9. Read green carefully
For rare disease panels, green denotes diagnostic-level evidence under the stated criteria and scope. It does not classify a patient variant.
Working checkpoint — green evidence rating: Open case-level variant evidence separately. Separate panel membership from variant pathogenicity, assay coverage and patient-level phenotype fit.
For students, ask: Which rule placed this entity in the set, and could a later evidence update change that decision?
Section 10 of 36
10. Read amber as uncertainty
Amber indicates borderline evidence that may merit research or review but is not treated like a green diagnostic entity in the described tiering context.
Working checkpoint — mode of inheritance: Keep amber results visibly distinct from diagnostic-grade inclusions. Treat entity type, inheritance label and version as scientific data: a silent mismatch can alter the candidate set.
For parents, a calm checkpoint is to ask what the panel can focus, what it can miss and what still needs independent review.
Section 11 of 36
11. Read red as low evidence, not impossibility
A red rating records limited evidence under the current panel criteria; future evidence can change a rating.
Working checkpoint — expert review: Preserve the review date and supporting comments. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Section 12 of 36
12. Read mode of inheritance
Inheritance labels help analysts interpret family data, but complex or multiple mechanisms require careful scope.
Working checkpoint — clinical indication: Compare the panel inheritance field with current disease evidence. Compare the displayed rating with its expert reviews, disease scope and current signed-off service version.
For future study, connect panel curation to biology, data science and ethics while checking current official pathways separately.
Section 13 of 36
13. Include STRs and CNVs when relevant
PanelApp can include genes, short tandem repeats and curated copy-number regions, so a panel is not always only a gene list.
Working checkpoint — virtual gene panel: Retain entity type in every export. Preserve every panel tested and every excluded entity; a shortlist without selection history is not auditable.
For scientific writing, name the panel and version whenever a colour rating or gene list is reported.
Section 14 of 36
14. Read expert reviews, not colour alone
Reviews can name publications, families, mechanisms, conflicting evidence and requested changes.
Working checkpoint — panel version: Trace decisive ratings to their review evidence. Separate panel membership from variant pathogenicity, assay coverage and patient-level phenotype fit.
For students, ask: Which rule placed this entity in the set, and could a later evidence update change that decision?
Section 15 of 36
15. Count independent families correctly
The official criteria refer to unrelated families and may incorporate strong functional data; multiple reports of the same family are not independent replication.
Working checkpoint — green evidence rating: Deduplicate pedigrees before judging evidence quantity. Treat entity type, inheritance label and version as scientific data: a silent mismatch can alter the candidate set.
For parents, a calm checkpoint is to ask what the panel can focus, what it can miss and what still needs independent review.
Section 16 of 36
16. Keep phenotype fit visible
A green relationship for one disorder scope may not explain a case with a different phenotype or mechanism.
Working checkpoint — mode of inheritance: Compare the patient’s documented features with the panel scope. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Section 17 of 36
17. Distinguish gene validity from variant pathogenicity
A credible gene–disease relationship is necessary context, while one allele needs its own population, computational, functional and segregation evidence.
Working checkpoint — expert review: Keep gene-level and variant-level assertions in separate columns. Compare the displayed rating with its expert reviews, disease scope and current signed-off service version.
For future study, connect panel curation to biology, data science and ethics while checking current official pathways separately.
Section 18 of 36
18. Respect assay limitations
The official site warns that some non-coding, mitochondrial and technically difficult regions may not be accurately analysed by a given workflow.
Working checkpoint — clinical indication: Document callable regions and test limitations. Preserve every panel tested and every excluded entity; a shortlist without selection history is not auditable.
For scientific writing, name the panel and version whenever a colour rating or gene list is reported.
Section 19 of 36
19. Use dates and release notes
PanelApp records changes and periodically aligns GMS panels with updates to the National Genomic Test Directory.
Working checkpoint — virtual gene panel: Compare the release used with the decision date. Separate panel membership from variant pathogenicity, assay coverage and patient-level phenotype fit.
For students, ask: Which rule placed this entity in the set, and could a later evidence update change that decision?
Section 20 of 36
20. Audit additions and removals
Genes can be added, removed or rerated as evidence changes; a historical report should not silently inherit a later panel.
Working checkpoint — panel version: Archive the exact entity list used at analysis time. Treat entity type, inheritance label and version as scientific data: a silent mismatch can alter the candidate set.
For parents, a calm checkpoint is to ask what the panel can focus, what it can miss and what still needs independent review.
Section 21 of 36
21. Avoid panel shopping
Trying many panels and reporting only the one that produces an attractive candidate creates selection bias.
Working checkpoint — green evidence rating: Predefine panel choice or report every panel tested. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Section 22 of 36
22. Challenge a panel that is too narrow
A narrow panel can miss an unexpected phenotype, newly established gene or atypical presentation.
Working checkpoint — mode of inheritance: Plan when and how broader analysis will be considered. Compare the displayed rating with its expert reviews, disease scope and current signed-off service version.
For future study, connect panel curation to biology, data science and ethics while checking current official pathways separately.
Section 23 of 36
23. Challenge a panel that is too broad
A huge super panel increases review burden and incidental candidates without guaranteeing better evidence.
Working checkpoint — expert review: Match panel breadth to the question and validation capacity. Preserve every panel tested and every excluded entity; a shortlist without selection history is not auditable.
For scientific writing, name the panel and version whenever a colour rating or gene list is reported.
Section 24 of 36
24. Use the API reproducibly
PanelApp provides programmatic access, but queries should record endpoint, filters, panel identifier, version and retrieval date.
Working checkpoint — clinical indication: Save the raw response and checksum. Separate panel membership from variant pathogenicity, assay coverage and patient-level phenotype fit.
For students, ask: Which rule placed this entity in the set, and could a later evidence update change that decision?
Section 25 of 36
25. Respect attribution and disclaimer
The site asks users to verify accuracy before use and describes limits on responsibility for contributed content.
Working checkpoint — virtual gene panel: Cite PanelApp and the underlying literature. Treat entity type, inheritance label and version as scientific data: a silent mismatch can alter the candidate set.
For parents, a calm checkpoint is to ask what the panel can focus, what it can miss and what still needs independent review.
Section 26 of 36
26. Protect patient privacy
Panel selection may reveal sensitive diagnostic hypotheses even when no genomic data are shared.
Working checkpoint — panel version: Minimise identifiers and follow approved governance. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Section 27 of 36
27. Use bounded clinical language
A panel prioritises review; it does not diagnose, prescribe or replace local expert assessment.
Working checkpoint — green evidence rating: Say “included for analysis” rather than “causal”. Compare the displayed rating with its expert reviews, disease scope and current signed-off service version.
For future study, connect panel curation to biology, data science and ethics while checking current official pathways separately.
Section 28 of 36
28. Connect to Primary Science
A carefully labelled sorting activity shows how one object can enter a group only under stated rules.
Working checkpoint — mode of inheritance: Use fictional organisms rather than health data. Preserve every panel tested and every excluded entity; a shortlist without selection history is not auditable.
For scientific writing, name the panel and version whenever a colour rating or gene list is reported.
Section 29 of 36
29. Connect to Secondary Science
Inheritance, variation and evidence strength explain why gene lists evolve as new families and experiments are reported.
Working checkpoint — expert review: Separate genetic association from molecular mechanism. Separate panel membership from variant pathogenicity, assay coverage and patient-level phenotype fit.
For students, ask: Which rule placed this entity in the set, and could a later evidence update change that decision?
Section 30 of 36
30. Connect to mathematics
Sets, overlaps, versions and false-positive burden make panel design a quantitative choice.
Working checkpoint — clinical indication: Calculate overlap without treating union size as evidence quality. Treat entity type, inheritance label and version as scientific data: a silent mismatch can alter the candidate set.
For parents, a calm checkpoint is to ask what the panel can focus, what it can miss and what still needs independent review.
Section 31 of 36
31. Connect to computing
APIs, identifiers, inheritance codes and release snapshots turn expert curation into reusable data.
Working checkpoint — virtual gene panel: Validate entity type and version at import. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Section 32 of 36
32. Connect to pathways
Panel curation links genetics, medicine, bioinformatics and data governance and can illuminate STEM education and career pathways.
Working checkpoint — panel version: Check current official training requirements separately. Compare the displayed rating with its expert reviews, disease scope and current signed-off service version.
For future study, connect panel curation to biology, data science and ethics while checking current official pathways separately.
Section 33 of 36
33. Did you know?
The same gene can be green on one panel and a different colour on another because evidence is judged against the phenotype scope of each panel.
Working checkpoint — green evidence rating: Always name the panel with the rating. Preserve every panel tested and every excluded entity; a shortlist without selection history is not auditable.
For scientific writing, name the panel and version whenever a colour rating or gene list is reported.
Section 34 of 36
34. Write a reader-ready result
A strong summary names panel identifier, title, type, version, indication, entity rating, inheritance, decisive evidence and assay limits.
Working checkpoint — mode of inheritance: Make panel context visible beside every candidate. Separate panel membership from variant pathogenicity, assay coverage and patient-level phenotype fit.
For students, ask: Which rule placed this entity in the set, and could a later evidence update change that decision?
Section 35 of 36
35. Build a validation ladder
Confirm panel choice, version and rating; inspect reviews; verify assay coverage; then assess variant, segregation and phenotype evidence independently.
Working checkpoint — expert review: Strengthen the weakest rung before concluding. Treat entity type, inheritance label and version as scientific data: a silent mismatch can alter the candidate set.
For parents, a calm checkpoint is to ask what the panel can focus, what it can miss and what still needs independent review.
Section 36 of 36
36. Finish with a reproducible bundle
Deliver panel metadata, entity export, version, release notes, review citations, query parameters, assay scope, candidate table and approval notes.
Working checkpoint — clinical indication: Make the entire panel decision reconstructable. Keep the panel identifier, type, version and retrieval date beside every exported entity.
For a classroom model, use fictional genes and versioned evidence cards so nobody mistakes the activity for clinical testing.
Contents · Previous section · Continue to the Science Learning Hub