Why Science? | Seq-Scope, Sequencing-Array Barcodes and Submicrometre Spatial-Transcript Evidence

Section 1 of 36

1. Begin with a sequenced coordinate carpet

Seq-Scope begins with a dense lawn of DNA clusters on a sequencing surface. Each cluster carries a random high-definition map coordinate identifier, or HDMI. Sequencing the array first establishes which barcode occupies each position, turning sequence identity into a molecular map before tissue ever touches the surface.

Contents · Next section

Section 2 of 36

2. Separate the coordinate read from the tissue read

The first sequencing phase decodes spatial barcodes; the later molecular phase reads tissue-derived cDNA associated with those barcodes. Confusing these stages hides an important control. A transcript can be sequenced accurately yet assigned wrongly if the original coordinate map is incomplete or misregistered.

Contents · Previous section · Next section

Section 3 of 36

3. Capture RNA where tissue contacts the array

A tissue section is placed over the decoded surface and released RNA is captured by nearby primers. The resulting address describes the capture location, not necessarily the RNA molecule’s exact original nanometre position. Diffusion, section thickness, permeabilisation and physical contact set the real spatial boundary.

Contents · Previous section · Next section

Section 4 of 36

4. Make a second strand carrying the address

Captured RNA is copied so the cDNA library retains both molecular information and the HDMI address. Reverse transcription and second-strand synthesis are selective biochemical steps. RNA abundance, structure and accessibility can change recovery before sequencing depth or software enters the analysis.

Contents · Previous section · Next section

Section 5 of 36

5. Reconstruct a spatial transcript image

Sequence reads are grouped by their decoded HDMI coordinates and aligned to a reference transcriptome. The output resembles an image whose pixels are molecular capture sites. It is a reconstructed measurement, so reference build, mapping rules, duplicate handling and quality thresholds remain part of every visible pattern.

Contents · Previous section · Next section

Section 6 of 36

6. Distinguish coordinate density from biological resolution

Very dense barcode clusters create small address spacing. Biological resolution also depends on RNA movement, tissue preparation, molecule number and analysis. A submicrometre coordinate grid does not guarantee that every observed transcript stayed within a submicrometre of its origin or that every cell boundary is directly measured.

Contents · Previous section · Next section

Section 7 of 36

7. Keep the founding demonstration in scope

The 2021 Cell study showed microscopic and subcellular spatial-transcript patterns in mouse liver and colon using a repurposed sequencing platform. That supports feasibility in those tested samples. It does not make every tissue, laboratory preparation, disease claim or apparent organelle pattern equally reliable.

Contents · Previous section · Next section

Section 8 of 36

8. Decode the array before committing the specimen

An undecoded or poorly decoded cluster cannot provide a trustworthy location later. Measure usable HDMI fraction, cluster density, sequencing quality and duplicate coordinate behaviour before applying valuable tissue. An attractive tissue image cannot repair a weak address map after the experiment.

Contents · Previous section · Next section

Section 9 of 36

9. Treat flow-cell manufacture as an experimental variable

Cluster generation depends on library composition, amplification and instrument chemistry. Uneven density, optical crowding or sequence bias can make some regions easier to decode. Spatial quality maps should accompany the tissue data so manufacturing structure is not mistaken for anatomical structure.

Contents · Previous section · Next section

Section 10 of 36

10. Prepare a flat, intact tissue section

Folds, tears, variable thickness and incomplete contact alter capture. Record section orientation and morphology before permeabilisation, then register the post-capture data to a neighbouring or pre-imaged section when appropriate. A molecular gap may be a tissue gap, a contact gap or true biology.

Contents · Previous section · Next section

Section 11 of 36

11. Optimise permeabilisation without erasing place

Too little permeabilisation lowers recovery; too much allows RNA to travel farther before capture. Time, temperature and tissue composition matter. A pilot series should evaluate yield and spatial sharpness together because maximising counts alone can sacrifice the location information that makes the assay valuable.

Contents · Previous section · Next section

Section 12 of 36

12. Measure capture and library complexity

Report unique molecules, genes, mitochondrial fraction and saturation across the surface. A dense coordinate map can still hold few captured molecules. Complexity should be compared by area, tissue region and specimen rather than reduced to one generous whole-slide average.

Contents · Previous section · Next section

Section 13 of 36

13. Register molecules to morphology

Spatial reads become interpretable when aligned to histology or other landmarks. Publish fiducials, transforms and local residuals, especially near tissue edges and distortions. Global alignment can look excellent while a small region is shifted enough to change cell or compartment assignment.

Contents · Previous section · Next section

Section 14 of 36

14. Choose binning for the question

Neighbouring HDMI coordinates may be combined into analysis bins to stabilise counts. Small bins preserve detail but can be sparse; large bins improve precision while mixing structures. Report the native coordinates and every bin size used so resolution is not created silently by presentation choices.

Contents · Previous section · Next section

Section 15 of 36

15. Practise with an invented Seq-Scope table

These fictional fields separate coordinate and chemistry problems.

All values are invented for teaching.

Contents · Previous section · Next section

Section 16 of 36

16. Map reads with ambiguity visible

Short reads can align to repeated genes, paralogues or multiple transcript isoforms. Keep uniquely mapped, multi-mapped and unresolved reads separate. A precise HDMI coordinate cannot remove uncertainty in molecular identity, and an exact gene match cannot remove uncertainty in spatial capture.

Contents · Previous section · Next section

Section 17 of 36

17. Segment cells only when evidence supports it

Subcellular coordinates tempt analysts to draw cell boundaries from transcript density. Validate segmentation with nuclei, membranes or morphology and retain unassigned molecules. Cells with unusual shape, low RNA or touching neighbours deserve uncertainty rather than forced polygons.

Contents · Previous section · Next section

Section 18 of 36

18. Interpret subcellular patterns cautiously

Mitochondrial, nuclear, cytoplasmic or zonated signals can be biologically informative. They can also reflect reference annotation, RNA stability, sectioning, diffusion or organelle abundance. Orthogonal imaging and replicated specimens should support claims that depend on a particular compartment.

Contents · Previous section · Next section

Section 19 of 36

19. Compare regions using equal measurement rules

Apply the same quality gates, binning and mapping policies across regions. Liver zones or colon layers can differ in RNA content and tissue permeability. A larger count difference may partly reflect assay efficiency unless positive controls and matched processing show otherwise.

Contents · Previous section · Next section

Section 20 of 36

20. Use independent tissue specimens as replicates

Millions of coordinates on one section share one animal, preparation and array. They are measurements, not independent biological replicates. Summarise effects by animal or independently sampled tissue, and treat adjacent sections or repeated bins as nested technical or spatial observations.

Contents · Previous section · Next section

Section 21 of 36

21. Audit the full evidence funnel

Count decoded clusters, tissue-covered clusters, captured molecules, quality reads, mapped reads, retained bins and assigned cells. Reporting each denominator reveals whether a striking final map represents broad coverage or a small, selected fraction of the original surface.

Contents · Previous section · Next section

Section 22 of 36

22. Challenge a headline resolution number

Ask what the number describes: centre-to-centre cluster spacing, chosen bin width, localisation precision, cell assignment or validated biological feature size. These are different quantities. Resolution should be reported as a measurement chain, not a single adjective detached from chemistry and analysis.

Contents · Previous section · Next section

Section 23 of 36

23. Challenge diffusion and molecular movement

RNA can move during permeabilisation or library preparation. Estimate spreading with sharp tissue boundaries, known localised transcripts or synthetic controls. When movement is appreciable, phrase conclusions at the supported region or compartment scale instead of the nominal cluster scale.

Contents · Previous section · Next section

Section 24 of 36

24. Challenge barcode-decoding gaps

Missing or low-quality HDMI calls produce spatial holes and biased coverage. Plot their location and sequence composition. Interpolation may aid visualisation but must never be presented as observed molecules; masks should distinguish measured, missing and inferred positions.

Contents · Previous section · Next section

Section 25 of 36

25. Challenge batch and surface effects

Arrays, reagent lots, sequencing runs and operators can imprint patterns. Randomise groups, include reference tissue and model batch at the specimen level. If condition and array are confounded, a beautiful spatial difference cannot be separated from production history.

Contents · Previous section · Next section

Section 26 of 36

26. Challenge reference and annotation choices

Transcript assignment depends on genome build, gene models and multimapping rules. Reanalyse key findings with a current, versioned reference and record changes. Spatial precision does not compensate for an outdated or incomplete molecular dictionary.

Contents · Previous section · Next section

Section 27 of 36

27. Keep medical interpretations inside research scope

Spatial transcript maps can reveal hypotheses about tissue organisation or disease. They are not clinical tests. Human use requires consent, governance, protected metadata, locked analytical procedures, independent cohorts and evidence that an assay is accurate and useful for the intended decision.

Contents · Previous section · Next section

Section 28 of 36

28. Compare Seq-Scope with Slide-seq

Both use a decoded spatial barcode surface, but Seq-Scope uses dense sequencing clusters while Slide-seq uses mapped barcoded beads. Their capture chemistry, address geometry, density, tissue contact and analysis differ. Choose by the validated scale and biological question, not by whichever headline resolution sounds smallest.

Contents · Previous section · Next section

Section 29 of 36

29. Build a safe coordinate-grid activity

Students place coloured paper transcripts over a fine numbered grid, then allow some tokens to drift before recording them. They compare native squares with larger bins. The exercise shows how address density, molecule movement and aggregation jointly determine the resolution of a claim.

Contents · Previous section · Next section

Section 30 of 36

30. Connect to Primary Science observation

Children can distinguish an object from the grid square used to record it. A finer grid gives more possible addresses, yet an object crossing a boundary remains uncertain. This builds careful observation without requiring laboratory samples.

Contents · Previous section · Next section

Section 31 of 36

31. Strengthen PSLE Science process skills

Pupils can identify controlled variables, compare the fictional fields and explain why the highest molecule count is not automatically the best spatial result. They practise matching conclusions to the actual measurement and recognising when a fair test needs more than one indicator.

Contents · Previous section · Next section

Section 32 of 36

32. Extend through Secondary and O-Level Science

Biology supplies cells, tissues and gene expression; Chemistry supplies hybridisation and enzymes; Physics supplies optics and diffusion; Mathematics supplies coordinates and uncertainty; Computing supplies mapping and segmentation. Seq-Scope turns school subjects into one connected evidence system.

Contents · Previous section · Next section

Section 33 of 36

33. Use the topic for school-choice questions

Families can check official school information for inquiry, biology, computing, microscopy and data projects. A paper grid or public dataset supports safe exploration. Do not infer a school’s equipment, programme access, admissions rules or outcomes from a general science topic.

Contents · Previous section · Next section

Section 34 of 36

34. See the career network

Spatial transcriptomics brings together molecular biology, surface chemistry, sequencing engineering, histology, bioinformatics, imaging, statistics and research ethics. Each role has different current training and safety requirements; no article can promise admission, employment or a single guaranteed pathway.

Contents · Previous section · Next section

Section 35 of 36

35. Did You Know? The sequencer becomes a microscope-like map

Seq-Scope gives a sequencing surface a second job: clusters first provide known molecular addresses, then help locate tissue-derived reads. The cleverness lies in connecting two readings of the same surface. The final map is still reconstructed evidence, not a direct photograph of every RNA molecule.

Contents · Previous section · Next section

Section 36 of 36

36. Close with a bounded Seq-Scope conclusion

A defensible statement names array decoding, cluster density, tissue preparation, permeabilisation, capture, sequencing, mapping, binning, registration, segmentation and independent specimens. It may report spatially resolved transcript patterns at validated scales; it should not equate address spacing with universal biological resolution or claim mechanism from location alone.

Pre-registration can name specimen, array, decoded-HDMI threshold, permeabilisation condition, minimum unique molecules, reference build, mapping policy, native coordinate size, planned bins, registration ceiling, primary segmentation and biological replicate. Exploratory features may follow in a labelled analysis. This prevents resolution, filtering and binning from drifting toward the most attractive microscopic pattern.

The barcode surface needs its own quality-control release. Publish cluster density, usable HDMI fraction, spatial holes, duplicate sequences, per-cycle quality and surface-wide gradients before tissue capture. If array quality is known only after biological results appear, it becomes too easy to dismiss bad regions selectively. A tissue-free control can separate manufacturing pattern from specimen pattern.

The molecular funnel should track tissue-covered coordinates, captured molecules, cDNA, sequenced fragments, unique molecules, mapped reads, retained bins and cell-assigned transcripts. Denominators must remain visible. A final figure containing millions of points can still exclude large areas or molecule classes, and those exclusions may correlate with anatomy.

Permeabilisation should be evaluated with both yield and sharpness. Plot molecule recovery against distance from a known boundary or localised control across a time series. The best condition is not automatically the one with the largest library. It is the condition that provides enough molecules while preserving the spatial distinction needed for the question.

Binning is a statistical decision. Report results at several justified scales and show how genes, compartments and neighbourhoods change. Native HDMIs are observations; a binned count is a summary. Smoothing, interpolation and imputation can aid interpretation, but their outputs must not be presented as newly measured molecules.

Mapping uncertainty should remain attached to coordinates. Release reference versions, aligner settings, multimapping decisions and transcript annotations. Repeat central results with an alternative plausible policy when gene families or isoforms matter. A coordinate may be extremely precise while the sequence identity remains only family-level, and the conclusion should reflect the weaker link.

Segmentation needs orthogonal morphology. Use nuclei, membranes or tissue structure and publish masks, unassigned molecules and border confidence. Compare a cell-based analysis with a segmentation-free neighbourhood or density analysis. If a finding survives both, it is less likely to be an artefact of one boundary model.

Spatial statistics should preserve tissue geometry. Molecules within one field are autocorrelated, bins overlap biological structures and cells share specimens. Use animal or donor as the biological unit, define neighbourhood scales before testing and correct exploratory multiplicity. Vast coordinate counts increase descriptive detail, not the number of independent organisms.

Protocol transfer requires fresh validation in every tissue. Liver, colon, brain, tumour and archival human sections differ in extracellular matrix, RNA integrity, autofluorescence and diffusion. Re-establish capture, spreading, mapping and morphology alignment rather than importing a resolution number from the founding study.

A bounded result might read: ‘Across independent mouse tissue sections, decoded Seq-Scope arrays recovered replicated spatial transcript patterns after array, permeabilisation, mapping, binning, registration and segmentation controls.’ It should report native cluster spacing and validated biological feature scale separately. It should not imply that every RNA is localised with the nominal address spacing.

Reproducibility requires HDMI sequences and coordinates, array quality maps, tissue images, orientation, permeabilisation metadata, raw sequence data, reference files, molecule tables, transforms, bins, masks, code, environments and checksums. Failed arrays and excluded fields belong in the record. The next experiment should target the dominant address, capture or assignment uncertainty.

Figures should connect the microscopic map to its evidence chain: decoded clusters, tissue boundary, unique molecules, mapping classes, raw and binned views, registration residuals, masks and specimen-level estimates. A reader should be able to follow one reported local feature back through coordinate, sequence and morphology without relying on a polished composite alone.

Contents · Previous section · Continue to the Science Learning Hub