eduKateSG · Why Science?
Request the genomic slice you need, receive a retrieval ticket, and keep security and coordinate meaning attached to every data block
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. htsget is a GA4GH retrieval API for genomic read and variation data. A client requests an identifier and optional genomic region, then receives a JSON ticket whose URLs or inline data blocks are fetched and concatenated in order. GA4GH lists htsget as current and v1.3 as the last approved product version; the live technical specification identifies itself as v1.3.1. Partial retrieval can reduce unnecessary transfer, but it does not remove consent, access-control, privacy or scientific-context requirements. 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: gnomAD and population allele-frequency context; GA4GH VRS and computable variant representation; GA4GH Beacon v2 and federated genomic discovery; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.
Primary and official sources checked on 12 October 2026: Official GA4GH htsget product page, approved v1.3; Official htsget retrieval API specification v1.3.1; htsget protocol paper; 2026 Singapore–Cambridge O-Level Biology syllabus; MOE G2/G3 Lower Secondary Science teaching and learning syllabus; 2026 MOE G2 Computing syllabus. Standards, implementations and recommendations can change, so preserve the version, access date and documentation used.
| Stage | Question | Evidence | Responsible output |
|---|---|---|---|
| Request | Which reads or variants and genomic interval are needed? | Endpoint, ID, format, reference name, start and end | A precise bounded request |
| Ticket | Where are the authorised data blocks? | JSON ticket, URLs, headers and classes | An ordered retrieval plan |
| Assemble | Do the blocks form the promised output? | Block order, format, header and body | A usable stream or file |
| Govern | May this caller retrieve this slice? | TLS, bearer tokens, logs and purpose controls | A secure, auditable access |
Inside this guide
1–12 · Foundations and design
- 1. Start with the genomic question
- 2. Understand the product status
- 3. Use the reads and variants endpoints
- 4. Know the ticket pattern
- 5. Choose GET or POST deliberately
- 6. Use zero-based half-open coordinates
- 7. Name the reference sequence
- 8. Request only supported formats
- 9. Use header and body classes
- 10. Apply field filtering carefully
- 11. Understand server latitude
- 12. Read ordered data blocks
13–24 · Models and estimation
- 13. Handle HTTPS block URLs
- 14. Handle inline data URIs
- 15. Preserve format validity
- 16. Use service-info
- 17. Do not use htsget for discovery
- 18. Plan retries for transient errors
- 19. Read machine-readable errors
- 20. Protect traffic with TLS
- 21. Send bearer tokens correctly
- 22. Separate protocol from policy
- 23. Minimise privacy exposure
- 24. Audit region queries
25–36 · Diagnostics and interpretation
- 25. Account for CRAM references
- 26. Compare with whole-file transfer
- 27. Test client and server together
- 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 retrieval report
- 35. Build a validation ladder
- 36. Finish with a reproducible bundle
Section 1 of 36
1. Start with the genomic question
htsget can retrieve a selected region or a full dataset, so efficiency begins by asking for the smallest scientifically sufficient scope.
Working checkpoint — htsget: Write the identifier, coordinates, format and purpose before calling the API. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Section 2 of 36
2. Understand the product status
GA4GH lists htsget as a current API and v1.3 as its last approved version; the maintained specification page identifies protocol v1.3.1.
Working checkpoint — ticket: Record both product and protocol version context. Read each regional result beside its reference, coordinates, format and sample filters.
For future study, connect streaming to biology, networks and cybersecurity while checking official pathways separately.
Section 3 of 36
3. Use the reads and variants endpoints
The recommended paths separate read data from variant data, making the expected domain object explicit.
Working checkpoint — genomic interval: Do not send a variant identifier to a reads endpoint. Preserve missing blocks and errors; a reconstructed stream needs every required piece.
For scientific writing, name the interval convention, reference and protocol version.
Section 4 of 36
4. Know the ticket pattern
A successful request returns JSON under an htsget key; the ticket points to one or more data blocks rather than necessarily carrying one giant file.
Working checkpoint — data block: Validate the response prefix and schema before retrieval. Separate efficient retrieval from biological interpretation and permission to access.
For students, ask what region was requested and which context had to accompany it.
Section 5 of 36
5. Choose GET or POST deliberately
GET supports ordinary parameterised requests, while POST can express lists of regions and structured parameters without an oversized query string.
Working checkpoint — TLS: Archive the complete request body or URL. Treat coordinates, block order and bearer headers as scientific and security data.
For parents, ask whether partial genomic data remain sensitive and controlled.
Section 6 of 36
6. Use zero-based half-open coordinates
The specification uses zero-based, half-open intervals: start is included and end is excluded.
Working checkpoint — bearer token: Test boundary positions with a known toy interval. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Section 7 of 36
7. Name the reference sequence
A coordinate is ambiguous without the reference name or sequence context against which it is measured.
Working checkpoint — htsget: Keep assembly and reference identifiers beside every interval. Read each regional result beside its reference, coordinates, format and sample filters.
For future study, connect streaming to biology, networks and cybersecurity while checking official pathways separately.
Section 8 of 36
8. Request only supported formats
Reads and variants can be delivered in formats a service declares and supports; an unsupported request should fail visibly.
Working checkpoint — ticket: Discover service capabilities before assuming BAM, CRAM, VCF or BCF. Preserve missing blocks and errors; a reconstructed stream needs every required piece.
For scientific writing, name the interval convention, reference and protocol version.
Section 9 of 36
9. Use header and body classes
A request can distinguish file header material from body data so clients can construct valid output for the selected use.
Working checkpoint — genomic interval: Preserve required headers when assembling regional data. Separate efficient retrieval from biological interpretation and permission to access.
For students, ask what region was requested and which context had to accompany it.
Section 10 of 36
10. Apply field filtering carefully
Variant requests can constrain returned fields or samples where supported, changing what the output contains.
Working checkpoint — data block: Record every field and sample filter. Treat coordinates, block order and bearer headers as scientific and security data.
For parents, ask whether partial genomic data remain sensitive and controlled.
Section 11 of 36
11. Understand server latitude
The design allows a server to return more data than requested but not less, because format blocks and indexing may require surrounding information.
Working checkpoint — TLS: Validate scientific scope after assembly rather than assuming perfect trimming. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Section 12 of 36
12. Read ordered data blocks
The ticket contains blocks that must be fetched and concatenated in the listed order. Reordering valid blocks can corrupt the result.
Working checkpoint — bearer token: Number and hash every fetched block. Read each regional result beside its reference, coordinates, format and sample filters.
For future study, connect streaming to biology, networks and cybersecurity while checking official pathways separately.
Section 13 of 36
13. Handle HTTPS block URLs
A block URL can point to another authorised endpoint and may carry headers needed for retrieval.
Working checkpoint — htsget: Treat URL and headers as sensitive capabilities. Preserve missing blocks and errors; a reconstructed stream needs every required piece.
For scientific writing, name the interval convention, reference and protocol version.
Section 14 of 36
14. Handle inline data URIs
Small data blocks can be supplied inline using a data URI rather than a second network request.
Working checkpoint — ticket: Decode with the declared media type and encoding. Separate efficient retrieval from biological interpretation and permission to access.
For students, ask what region was requested and which context had to accompany it.
Section 15 of 36
15. Preserve format validity
Concatenated blocks should create valid output in the promised format, but a successful HTTP transfer is not a bioinformatics validation.
Working checkpoint — genomic interval: Open the result with a trusted format-aware tool. Treat coordinates, block order and bearer headers as scientific and security data.
For parents, ask whether partial genomic data remain sensitive and controlled.
Section 16 of 36
16. Use service-info
The GA4GH service-info endpoint helps a client discover service identity, type, organisation and version before issuing retrieval requests.
Working checkpoint — data block: Save service-info with the query record. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Section 17 of 36
17. Do not use htsget for discovery
The specification does not discover valid dataset identifiers; clients learn them through another catalogue or workflow.
Working checkpoint — TLS: Record the source that supplied the ID. Read each regional result beside its reference, coordinates, format and sample filters.
For future study, connect streaming to biology, networks and cybersecurity while checking official pathways separately.
Section 18 of 36
18. Plan retries for transient errors
The specification recommends retry logic for transient 5xx responses, not blind repetition of invalid requests.
Working checkpoint — bearer token: Use bounded exponential backoff and preserve error details. Preserve missing blocks and errors; a reconstructed stream needs every required piece.
For scientific writing, name the interval convention, reference and protocol version.
Section 19 of 36
19. Read machine-readable errors
Defined errors distinguish invalid authentication, permission denial, missing resources, invalid input, unsupported format and unsatisfied ranges.
Working checkpoint — htsget: Route each error to a specific recovery step. Separate efficient retrieval from biological interpretation and permission to access.
For students, ask what region was requested and which context had to accompany it.
Section 20 of 36
20. Protect traffic with TLS
Sensitive tokens and human genomic data on public networks must use TLS 1.2 or later under the specification.
Working checkpoint — ticket: Reject insecure redirects and verify certificates. Treat coordinates, block order and bearer headers as scientific and security data.
For parents, ask whether partial genomic data remain sensitive and controlled.
Section 21 of 36
21. Send bearer tokens correctly
Where required, an OAuth 2.0 bearer access token travels in the Authorization header of the HTTPS request.
Working checkpoint — genomic interval: Never place the token in a query string or published log. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Section 22 of 36
22. Separate protocol from policy
htsget defines retrieval mechanics; institutions still decide identity proofing, authorisation, consent and acceptable use.
Working checkpoint — data block: Cite the governing access decision. Read each regional result beside its reference, coordinates, format and sample filters.
For future study, connect streaming to biology, networks and cybersecurity while checking official pathways separately.
Section 23 of 36
23. Minimise privacy exposure
A narrow genomic slice can still be identifying or reveal a sensitive variant. Smaller transfer does not mean harmless data.
Working checkpoint — TLS: Apply the same controlled-access policy to partial results. Preserve missing blocks and errors; a reconstructed stream needs every required piece.
For scientific writing, name the interval convention, reference and protocol version.
Section 24 of 36
24. Audit region queries
Query logs support security and reproducibility but can themselves reveal genes, loci or cases of interest.
Working checkpoint — bearer token: Protect logs and set justified retention periods. Separate efficient retrieval from biological interpretation and permission to access.
For students, ask what region was requested and which context had to accompany it.
Section 25 of 36
25. Account for CRAM references
CRAM decoding can depend on exact reference sequences, so transport success can still lead to interpretation failure when references differ.
Working checkpoint — htsget: Resolve and verify reference sequence digests. Treat coordinates, block order and bearer headers as scientific and security data.
For parents, ask whether partial genomic data remain sensitive and controlled.
Section 26 of 36
26. Compare with whole-file transfer
Full transfer may be appropriate for genome-wide analysis; regional streaming is valuable for browsing or targeted questions.
Working checkpoint — ticket: Choose by scientific scope, not fashion. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Section 27 of 36
27. Test client and server together
Different implementations can expose the same protocol yet differ in supported formats, limits and authentication.
Working checkpoint — genomic interval: Run conformance and end-to-end tests with realistic permissions. Read each regional result beside its reference, coordinates, format and sample filters.
For future study, connect streaming to biology, networks and cybersecurity while checking official pathways separately.
Section 28 of 36
28. Connect to Primary Science
A librarian sending only requested pages plus the book’s contents page illustrates targeted retrieval and required context.
Working checkpoint — data block: Use fictional books and check page order. Preserve missing blocks and errors; a reconstructed stream needs every required piece.
For scientific writing, name the interval convention, reference and protocol version.
Section 29 of 36
29. Connect to Secondary Science
Genomic intervals connect DNA, variation and evidence to the importance of reference frames and exact coordinates.
Working checkpoint — TLS: Translate one safe toy interval using half-open notation. Separate efficient retrieval from biological interpretation and permission to access.
For students, ask what region was requested and which context had to accompany it.
Section 30 of 36
30. Connect to mathematics
Intervals, byte ranges, ordering and transfer size show how boundary conventions change counts.
Working checkpoint — bearer token: Compute length as end minus start for a half-open interval. Treat coordinates, block order and bearer headers as scientific and security data.
For parents, ask whether partial genomic data remain sensitive and controlled.
Section 31 of 36
31. Connect to computing
HTTP, JSON, authentication, redirects, range selection and streaming make htsget a rich API case study.
Working checkpoint — htsget: Test expired tokens, partial block failure and order corruption. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Section 32 of 36
32. Connect to pathways
Secure genomic streaming links biology, software, networks, cloud systems and cybersecurity and can illuminate STEM pathways.
Working checkpoint — ticket: Check current official course and professional requirements separately. Read each regional result beside its reference, coordinates, format and sample filters.
For future study, connect streaming to biology, networks and cybersecurity while checking official pathways separately.
Section 33 of 36
33. Did you know?
The ticket can mix remotely fetched blocks and inline data while the client assembles one ordered output.
Working checkpoint — genomic interval: Inspect every block source and header before concatenation. Preserve missing blocks and errors; a reconstructed stream needs every required piece.
For scientific writing, name the interval convention, reference and protocol version.
Section 34 of 36
34. Write a reader-ready retrieval report
Name the protocol version, service, dataset ID, interval convention, format, filters, authentication tier, block count and validation outcome.
Working checkpoint — data block: Redact tokens and signed URLs. Separate efficient retrieval from biological interpretation and permission to access.
For students, ask what region was requested and which context had to accompany it.
Section 35 of 36
35. Build a validation ladder
Validate the identifier and coordinates, inspect service-info, obtain a ticket, fetch blocks securely, assemble in order, then verify format and biological context.
Working checkpoint — TLS: Stop on missing or altered blocks. Treat coordinates, block order and bearer headers as scientific and security data.
For parents, ask whether partial genomic data remain sensitive and controlled.
Section 36 of 36
36. Finish with a reproducible bundle
Deliver the permitted request, service metadata, redacted ticket manifest, block checksums, client version, errors and interpretation limits.
Working checkpoint — bearer token: Do not redistribute controlled genomic bytes without approval. Keep the protocol version, service, request, ticket manifest and assembled result together.
For a classroom model, use fictional page ranges rather than genomic records.
Contents · Previous section · Continue to the Science Learning Hub