eduKateSG · Why Science?
Read a BLAT-style alignment as 21 linked fields—then keep strands, blocks, coordinate spaces and uncertainty in view
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. PSL is UCSC’s line-oriented alignment format commonly produced by BLAT or psLayout. Each data line has 21 required fields: match and mismatch counts, repeat and N counts, query and target insertion summaries, strand, query and target names and sizes, overall start and end coordinates, block count, block sizes, query starts and target starts. A block contains no alignment gaps. UCSC documents special handling for negative-strand query coordinates and translated alignments. A well-formed PSL row is an algorithmic alignment description, not proof of unique origin, gene function, structural variation or biological mechanism. 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: FASTA sequence records and reference data; SAM and human-readable sequence alignments; BED and genomic feature coordinates; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.
Primary and official sources checked on 12 October 2026: UCSC Genome Browser PSL format specification, checked 12 October 2026; UCSC Genome Browser BLAT documentation, checked 12 October 2026; UCSC Genome Browser rearrangement display documentation, checked 12 October 2026; 2026 Singapore–Cambridge O-Level Biology syllabus; MOE G2/G3 Lower Secondary Science teaching and learning syllabus; 2026 MOE G2 Computing syllabus. Specifications, tools and recommendations can change, so preserve the dated documentation, exact source data, coordinate rules and software versions used.
A deeper reading route
PSL compresses an alignment into counts, outer bounds and ungapped blocks. The summary fields make quick comparisons possible, while the three block arrays reconstruct where aligned segments lie in query and target. Every layer should agree. A plausible percent identity calculated from summary counts can hide impossible block geometry if the arrays were truncated or reordered.
Query and target are distinct coordinate spaces. qName, qSize, qStart and qEnd describe the submitted sequence; tName, tSize, tStart and tEnd describe the database sequence. Treating one as the other can reverse an interpretation, especially when a short transcript is aligned to a whole chromosome.
Strand is more than a display symbol. UCSC explains that qStart and qEnd remain expressed on the forward query, while qStarts uses reverse-strand coordinates when the DNA query aligns on the minus strand. Translated searches can add a target-strand symbol and mix amino-acid block sizes with nucleotide target coordinates. A parser must follow the documented case rather than one universal shortcut.
Match counts are descriptive, not universal truth scores. repeats, Ns, mismatches and insertion counts depend on the aligned sequences and BLAT scoring choices. Two rows with similar matches can differ sharply in block arrangement, query coverage or alternative placements. Interpretation should include alignment coverage, competing hits and the biological question.
BLAT is designed for relatively similar sequences and fast searches. Its output can be excellent for locating transcripts, primers or assembled sequences, but it is not a replacement for every read aligner or homology method. Report the target database, assembly, server configuration and query type before comparing results across systems.
Validation reconstructs the alignment. Confirm 21 fields, array cardinalities and monotonic target blocks; ensure every block lies inside declared bounds; account for gap summaries; and fetch the relevant sequences when base-level confirmation matters. Include plus, minus, multi-block and translated fixtures.
A browser view can make a PSL record approachable, including rearrangement displays for compatible tracks. The view is not the entire evidence. Save the original row, query sequence, database identity, BLAT version and parameters so a later reader can reproduce both geometry and scoring.
For students, PSL turns a familiar matching exercise into structured evidence. The happy insight is that an alignment is not merely “similar” or “different”: it has direction, coverage, gaps, alternatives and a reference context, all of which can be inspected and questioned.
Identity percentages need a declared denominator. Some summaries use matches plus mismatches and repeat matches; others include query coverage, Ns or gap penalties. A single percentage can therefore differ between tools even for the same record. Publish the formula, raw fields and coverage so readers can reproduce the number and see which uncertainty was excluded.
Feature interpretation requires more than alignment geometry. A spliced transcript-like PSL can suggest exons separated by long target gaps, yet the record alone does not prove transcript expression, splice-junction support or gene identity. Combine it with annotation and assay evidence when those claims matter, and label an alignment-derived model as predicted until independently supported.
Round-trip visual checks can be made rigorous. Load a small PSL track, inspect plus and minus multi-block examples, export the underlying fields and compare them with the source row. Save region, assembly, display mode and screenshot only as a human-readable receipt; use field-level checks as the authoritative equivalence test.
Educational use is strongest when two candidates compete. Give learners one query and two invented target alignments with different coverage, mismatches and gaps. Ask them to choose what extra information they need before declaring a best explanation. That small uncertainty exercise mirrors real science better than a single perfect match.
A careful PSL review also checks arithmetic across every block. The blockCount value must agree with the number of entries in blockSizes, qStarts and tStarts; each interval must stay within qSize or tSize; and successive blocks should respect the declared strand and order. Reviewers can then reconstruct query and target coverage from the arrays and compare those totals with qStart, qEnd, tStart and tEnd. These checks catch truncated rows, shifted columns and unit mistakes before a visually plausible alignment reaches annotation or reporting. Keep the original row beside the reconstructed intervals so the audit remains reversible.
Database choice is part of the result, not background scenery. Searching one transcript collection, one genome assembly or one set of repeat-masked targets can change the candidates a query is able to find. Record the exact target release, sequence identifiers and masking policy, along with whether the query was DNA, protein or translated sequence. When a later database release changes a top hit, the right explanation may be improved reference coverage rather than a changed biological specimen. That distinction is essential in reproducible bioinformatics and a useful reminder that evidence always comes from a defined comparison. A compact results table should therefore name the database beside every reported alignment and preserve enough provenance for the same target collection to be rebuilt or retrieved.
| Stage | Question | Evidence | Responsible output |
|---|---|---|---|
| Summarise | How many aligned bases match or differ? | matches, misMatches, repMatches and nCount | A bounded similarity summary |
| Locate | Where are outer query and target bounds? | Names, sizes, strands, starts and ends | A two-coordinate-space placement |
| Reconstruct | How do ungapped segments connect? | blockCount, blockSizes, qStarts and tStarts | An explicit gapped alignment geometry |
| Judge | What can the record responsibly support? | Database, alternatives, parameters and sequence checks | A qualified alignment claim |
Inside this guide
1–12 · Foundations and design
- 1. Start with the alignment question
- 2. Recognise a 21-field record
- 3. Know the typical producers
- 4. Separate algorithm from evidence
- 5. Keep query and target distinct
- 6. Read matches
- 7. Read mismatches
- 8. Read repeat matches
- 9. Read N counts
- 10. Read query insert counts
- 11. Read target insert counts
- 12. Read strand precisely
13–24 · Models and estimation
- 13. Read query identity and size
- 14. Read query bounds
- 15. Read target identity and size
- 16. Read target bounds
- 17. Count ungapped blocks
- 18. Read blockSizes
- 19. Read qStarts
- 20. Read tStarts
- 21. Validate array punctuation
- 22. Reconstruct block ends
- 23. Check outer bounds
- 24. Handle negative-strand queries
25–36 · Diagnostics and interpretation
- 25. Handle translated alignments
- 26. Calculate coverage responsibly
- 27. Compare alternative hits
- 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 alignment note
- 35. Build a validation ladder
- 36. Finish with a reproducible bundle
Section 1 of 36
1. Start with the alignment question
A PSL row should be read in light of why a query was aligned and which target database was searched.
Working checkpoint — PSL record: Name query type, target assembly and intended inference. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Section 2 of 36
2. Recognise a 21-field record
UCSC requires 21 ordered fields on every PSL data line. Missing or shifted fields change all later meanings.
Working checkpoint — BLAT alignment: Count fields before calculating any statistic. Read every count beside query coverage, block geometry, strand and target assembly.
For future study, connect genetics to intervals and algorithms while checking pathways separately.
Section 3 of 36
3. Know the typical producers
PSL lines are commonly generated by BLAT or psLayout and can be displayed as tracks.
Working checkpoint — query and target: Record producer, version and command or server settings. Preserve secondary placements and failed validations; one attractive hit is not the complete search.
For scientific writing, report query, database, geometry, alternatives and limits.
Section 4 of 36
4. Separate algorithm from evidence
BLAT finds alignments under its search and scoring strategy. A returned row is not automatically the only or correct biological explanation.
Working checkpoint — blockSizes: Inspect competing alignments and query coverage. Separate algorithmic alignment, sequence similarity, annotation and mechanism.
For students, ask which pieces align, in which direction and with what gaps.
Section 5 of 36
5. Keep query and target distinct
The query is the submitted sequence and the target is the searched sequence collection. Their names, sizes and coordinates answer different questions.
Working checkpoint — strand: Label both axes in every diagram. Treat coordinate space, units, array counts and orientation as scientific inputs.
For parents, ask whether a top match was compared with alternatives.
Section 6 of 36
6. Read matches
matches counts aligned bases that match and are not repeats under the PSL convention.
Working checkpoint — alignment evidence: Report it beside coverage and other count fields. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Section 7 of 36
7. Read mismatches
misMatches counts aligned bases that do not match. Sequencing errors, polymorphism or wrong placement can contribute.
Working checkpoint — PSL record: Do not assign a biological cause from the count alone. Read every count beside query coverage, block geometry, strand and target assembly.
For future study, connect genetics to intervals and algorithms while checking pathways separately.
Section 8 of 36
8. Read repeat matches
repMatches separates matching bases identified as repeats. Repeat context can create multiple plausible placements.
Working checkpoint — BLAT alignment: Inspect masking and alternative hits. Preserve secondary placements and failed validations; one attractive hit is not the complete search.
For scientific writing, report query, database, geometry, alternatives and limits.
Section 9 of 36
9. Read N counts
nCount records aligned N bases, which represent unknown sequence rather than a confirmed canonical match.
Working checkpoint — query and target: Keep uncertainty visible in identity summaries. Separate algorithmic alignment, sequence similarity, annotation and mechanism.
For students, ask which pieces align, in which direction and with what gaps.
Section 10 of 36
10. Read query insert counts
qNumInsert and qBaseInsert summarise gaps inserted in the query side of the alignment.
Working checkpoint — blockSizes: Compare the summary with adjacent blocks. Treat coordinate space, units, array counts and orientation as scientific inputs.
For parents, ask whether a top match was compared with alternatives.
Section 11 of 36
11. Read target insert counts
tNumInsert and tBaseInsert summarise gaps on the target side.
Working checkpoint — strand: Check whether the pattern fits introns, indels or another hypothesis. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Section 12 of 36
12. Read strand precisely
The strand field identifies query orientation and can contain a second symbol for translated target orientation.
Working checkpoint — alignment evidence: Parse one or two symbols according to query type. Read every count beside query coverage, block geometry, strand and target assembly.
For future study, connect genetics to intervals and algorithms while checking pathways separately.
Section 13 of 36
13. Read query identity and size
qName and qSize define the coordinate system for qStart, qEnd and qStarts.
Working checkpoint — PSL record: Verify qSize against the actual query sequence. Preserve secondary placements and failed validations; one attractive hit is not the complete search.
For scientific writing, report query, database, geometry, alternatives and limits.
Section 14 of 36
14. Read query bounds
qStart and qEnd give the overall aligned query interval and use forward-strand coordinates even for a negative DNA query.
Working checkpoint — BLAT alignment: Test the documented reverse-coordinate formulas. Separate algorithmic alignment, sequence similarity, annotation and mechanism.
For students, ask which pieces align, in which direction and with what gaps.
Section 15 of 36
15. Read target identity and size
tName and tSize define the target coordinate system, commonly a chromosome in a named assembly.
Working checkpoint — query and target: Resolve the target against an exact reference release. Treat coordinate space, units, array counts and orientation as scientific inputs.
For parents, ask whether a top match was compared with alternatives.
Section 16 of 36
16. Read target bounds
tStart and tEnd bound the aligned region on the target and use zero-based half-open browser-style coordinates.
Working checkpoint — blockSizes: Confirm length and boundary conversions explicitly. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Section 17 of 36
17. Count ungapped blocks
blockCount states how many gap-free alignment blocks appear in the three lists.
Working checkpoint — strand: Require every list to have exactly this many values. Read every count beside query coverage, block geometry, strand and target assembly.
For future study, connect genetics to intervals and algorithms while checking pathways separately.
Section 18 of 36
18. Read blockSizes
blockSizes gives the size of each ungapped block; protein-query PSLs use amino-acid sizes while target coordinates remain nucleotides.
Working checkpoint — alignment evidence: Identify the unit before summing lengths. Preserve secondary placements and failed validations; one attractive hit is not the complete search.
For scientific writing, report query, database, geometry, alternatives and limits.
Section 19 of 36
19. Read qStarts
qStarts gives each block start on the query, with special reverse-coordinate handling for negative-strand DNA queries.
Working checkpoint — PSL record: Transform coordinates only with qSize and strand in view. Separate algorithmic alignment, sequence similarity, annotation and mechanism.
For students, ask which pieces align, in which direction and with what gaps.
Section 20 of 36
20. Read tStarts
tStarts gives each block start on the target. These starts should reconstruct ordered geometry appropriate to the record.
Working checkpoint — BLAT alignment: Check every block stays inside target bounds. Treat coordinate space, units, array counts and orientation as scientific inputs.
For parents, ask whether a top match was compared with alternatives.
Section 21 of 36
21. Validate array punctuation
Block arrays are comma-separated and often end with a trailing comma. Naive splitting can create a false empty element.
Working checkpoint — query and target: Use a PSL-aware parser and test cardinality. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Section 22 of 36
22. Reconstruct block ends
Each end equals start plus block size in its coordinate space, subject to translated-unit rules.
Working checkpoint — blockSizes: Draw every block before interpreting gaps. Read every count beside query coverage, block geometry, strand and target assembly.
For future study, connect genetics to intervals and algorithms while checking pathways separately.
Section 23 of 36
23. Check outer bounds
The minimum and maximum block coordinates should agree with qStart/qEnd and tStart/tEnd under the strand convention.
Working checkpoint — strand: Reject unexplained disagreements. Preserve secondary placements and failed validations; one attractive hit is not the complete search.
For scientific writing, report query, database, geometry, alternatives and limits.
Section 24 of 36
24. Handle negative-strand queries
UCSC documents that qStarts is reversed while qStart and qEnd stay forward-oriented for negative DNA-query PSLs.
Working checkpoint — alignment evidence: Keep raw and transformed coordinates in the audit. Separate algorithmic alignment, sequence similarity, annotation and mechanism.
For students, ask which pieces align, in which direction and with what gaps.
Section 25 of 36
25. Handle translated alignments
Protein or translated queries introduce amino-acid units and six-frame target logic that ordinary DNA assumptions cannot decode safely.
Working checkpoint — PSL record: Route translated records through dedicated tests. Treat coordinate space, units, array counts and orientation as scientific inputs.
For parents, ask whether a top match was compared with alternatives.
Section 26 of 36
26. Calculate coverage responsibly
Query coverage relates aligned query span or summed blocks to qSize, but clipping, gaps and translated units affect the denominator.
Working checkpoint — BLAT alignment: State the formula and units used. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Section 27 of 36
27. Compare alternative hits
Repeated or paralogous sequence can produce several good placements. One top score does not establish uniqueness.
Working checkpoint — query and target: Report score separation and retained alternatives. Read every count beside query coverage, block geometry, strand and target assembly.
For future study, connect genetics to intervals and algorithms while checking pathways separately.
Section 28 of 36
28. Connect to Primary Science
Matching patterned strips can introduce aligned pieces, gaps and direction without biological data.
Working checkpoint — blockSizes: Ask pupils to explain why one match may not be unique. Preserve secondary placements and failed validations; one attractive hit is not the complete search.
For scientific writing, report query, database, geometry, alternatives and limits.
Section 29 of 36
29. Connect to Secondary Science
Sequence alignment connects genetics to evidence evaluation: similarity supports relatedness questions but does not alone prove function.
Working checkpoint — strand: Separate observation, algorithm and interpretation. Separate algorithmic alignment, sequence similarity, annotation and mechanism.
For students, ask which pieces align, in which direction and with what gaps.
Section 30 of 36
30. Connect to mathematics
Intervals, ratios, array cardinality and coordinate transformations make PSL a useful discrete-mathematics case.
Working checkpoint — alignment evidence: Reconstruct a two-block example by hand. Treat coordinate space, units, array counts and orientation as scientific inputs.
For parents, ask whether a top match was compared with alternatives.
Section 31 of 36
31. Connect to computing
Parsers, ordered fields, typed units and visualisation show how compact records support interoperable tools.
Working checkpoint — PSL record: Test malformed counts and reverse strands. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Section 32 of 36
32. Connect to pathways
Alignment work connects biology, genomics, software and data stewardship and can illuminate STEM pathways.
Working checkpoint — BLAT alignment: Check current official qualifications separately. Read every count beside query coverage, block geometry, strand and target assembly.
For future study, connect genetics to intervals and algorithms while checking pathways separately.
Section 33 of 36
33. Did you know?
A PSL negative-strand DNA query stores overall qStart and qEnd in forward coordinates while the qStarts block list is reversed.
Working checkpoint — query and target: Use the official formula before comparing coordinates. Preserve secondary placements and failed validations; one attractive hit is not the complete search.
For scientific writing, report query, database, geometry, alternatives and limits.
Section 34 of 36
34. Write a reader-ready alignment note
Name query, target database, assembly, BLAT version, parameters, PSL row, alternatives and validation.
Working checkpoint — blockSizes: Keep the original query sequence and record hash. Separate algorithmic alignment, sequence similarity, annotation and mechanism.
For students, ask which pieces align, in which direction and with what gaps.
Section 35 of 36
35. Build a validation ladder
Check 21 fields, counts, names, bounds, strands, arrays, units, sequences and alternative placements.
Working checkpoint — strand: Stop at the first inconsistent coordinate space. Treat coordinate space, units, array counts and orientation as scientific inputs.
For parents, ask whether a top match was compared with alternatives.
Section 36 of 36
36. Finish with a reproducible bundle
Deliver queries, target manifest, PSL, commands, software, fixtures, visual receipts and bounded conclusions.
Working checkpoint — alignment evidence: Make every placement traceable to exact inputs. Keep query sequence, target manifest, BLAT version, parameters, PSL record and competing hits together.
For a classroom model, align invented symbol strips and record all plausible placements.
Contents · Previous section · Continue to the Science Learning Hub