Why Science? | CoTECH, Chromatin-Occupancy Barcoding and Transcriptome Evidence

Section 1 of 36

1. Start with an occupancy–expression question

CoTECH links an antibody-targeted chromatin-occupancy profile with RNA from the same cellular barcode. It can ask how a histone modification and transcription vary through a transition. The answer is strongest when direct observations, computational alignment and reconstructed pseudocells are kept as distinct evidence layers.

Contents · Next section

Section 2 of 36

2. Did you know an anchor can join experiments without becoming a cell?

If separate CoTECH experiments target H3K4me3 and H3K27me3, their RNA profiles can serve as anchors for matching similar states. The resulting pseudocell can approximate both marks, but no literal cell necessarily supplied both chromatin profiles. Reconstruction is useful precisely when labelled honestly.

Contents · Previous section · Next section

Section 3 of 36

3. Define chromatin occupancy operationally

Reads enriched near genomic regions targeted by the antibody provide an occupancy proxy under the protocol’s conditions. Signal depends on antibody specificity, epitope access, tagmentation or cleavage efficiency, background and coverage. It does not count every modified nucleosome or prove a functional effect.

Contents · Previous section · Next section

Section 4 of 36

4. Define the transcriptome as a sample

RNA molecules recovered from a cell are a sparse sample influenced by RNA integrity, priming, reverse transcription and sequencing. Nuclear RNA may emphasise newly transcribed or unspliced molecules. Report the counting model and avoid treating a zero as demonstrated absence.

Contents · Previous section · Next section

Section 5 of 36

5. Keep three entities separate

A measured cell contains one targeted chromatin channel and RNA. An aligned neighbour is another cell with a similar transcriptome. A pseudocell combines information across such neighbours or experiments. Confusing these units can turn computational correspondence into direct molecular co-occurrence.

Contents · Previous section · Next section

Section 6 of 36

6. Why transitions are scientifically interesting

During development, chromatin marks and RNA can change together, lead one another or temporarily decouple. Same-cell pairing helps map that relationship for the targeted mark. Pseudotime can organise cells, but observed stage, independent specimens and perturbation are still needed to test sequence and mechanism.

Contents · Previous section · Next section

Section 7 of 36

7. Plan separate mark experiments symmetrically

Use comparable specimens, stages, cell preparation, RNA chemistry, sequencing goals and biological replication for each antibody target. Randomise batches and share reference samples where possible. If one mark is measured in deeper or healthier cells, the pseudocell comparison may reflect technical asymmetry.

Contents · Previous section · Next section

Section 8 of 36

8. Preserve RNA and antibody epitopes

Fixation and permeabilisation must protect two very different analytes. Excess fixation may reduce enzyme access and RNA recovery; insufficient fixation can permit leakage or mixing. Record concentrations, temperatures, timings and washes, and inspect nuclear integrity before committing valuable developmental samples.

Contents · Previous section · Next section

Section 9 of 36

9. Validate each antibody independently

For H3K4me3, H3K27me3 or any other target, document specificity, lot and expected positive and negative loci. Compare with an orthogonal or bulk reference when appropriate. The transcriptome cannot rescue a chromatin channel generated by an unsuitable reagent.

Contents · Previous section · Next section

Section 10 of 36

10. Generate targeted chromatin fragments

Antibody-directed molecular chemistry marks or releases DNA near the chosen occupancy feature for sequencing. Reaction conditions influence fragment yield and background. Report enrichment, genomic distribution, fragment sizes and negative controls. The output is a targeted assay, not a complete chromatin map.

Contents · Previous section · Next section

Section 11 of 36

11. Capture RNA under the same identity

Reverse transcription creates RNA-derived cDNA that shares a cellular indexing history with chromatin fragments. Evaluate molecules, genes and mapping separately. A cell should not enter a cross-modal analysis merely because one channel is strong; the relevant claim requires adequate evidence in both.

Contents · Previous section · Next section

Section 12 of 36

12. Apply combinatorial barcodes

Cells or nuclei receive ordered barcode combinations through split-pool handling. The scheme expands throughput while sharing reactions. Accurate transfer records, balanced wells, unique barcode sequences and unused combinations are crucial. Estimate collisions and doublets because a mixed identity can create false occupancy–RNA coupling.

Contents · Previous section · Next section

Section 13 of 36

13. Construct modality-specific libraries

After barcoding, chromatin- and RNA-derived products are amplified, separated and sequenced. Publish read structures, index parsing, duplicate handling, mapping and molecule definitions. A reader should be able to follow raw sequence through to one accepted chromatin fragment or RNA count.

Contents · Previous section · Next section

Section 14 of 36

14. Pilot before developmental inference

Measure usable cells, chromatin enrichment, RNA molecules, genes, duplicates, saturation and collision indicators. Check each antibody batch separately. A transition study needs balanced coverage across states and specimens; otherwise smoothing may bridge a technical gap and make it look like a biological continuum.

Contents · Previous section · Next section

Section 15 of 36

15. Practise with an invented CoTECH table

These fictional values teach evidence-layer discipline; they are not performance benchmarks.

Contents · Previous section · Next section

Section 16 of 36

16. Demultiplex before integrating

Set barcode whitelists and mismatch rules without seeing the desired biological grouping. Preserve invalid paths and unused combinations. Examine count outliers and incompatible markers. Re-run major findings under strict matching so barcode rescue does not become the hidden reason that two modalities appear coordinated.

Contents · Previous section · Next section

Section 17 of 36

17. Quality-control targeted chromatin

Report unique fragments, duplicate rate, target-to-background enrichment, genomic distribution, chromosome coverage and antibody-lot effects. Plot them by plate and specimen. A large fragment total with weak target enrichment can be less informative than a smaller, specific library.

Contents · Previous section · Next section

Section 18 of 36

18. Quality-control RNA separately

Report molecules, genes, mapping categories, stress or mitochondrial signal where appropriate, and cell-state markers. Plot RNA quality against chromatin quality. Avoid a single pass score that permits one channel to compensate for failure in the other. Cross-modal claims need both.

Contents · Previous section · Next section

Section 19 of 36

19. Use transcriptomes as anchors transparently

When aligning cells from different targeted-mark experiments, state the RNA features, neighbours, distance metric and uncertainty. Confirm that alignment is not driven by batch, depth or cell cycle. Show unmatched cells instead of forcing every profile into a partner. An honest gap may reveal missing biology.

Contents · Previous section · Next section

Section 20 of 36

20. Construct pseudocells with uncertainty

A pseudocell combines chromatin information across transcriptomically similar cells from separate experiments. Report how many source cells contribute, their specimens, distances and weights. Vary neighbour counts and anchoring features. Do not use phrases such as measured together when the second mark was computationally reconstructed.

Contents · Previous section · Next section

Section 21 of 36

21. Interpret bivalent-mark models cautiously

Active-associated H3K4me3 and repressive-associated H3K27me3 can both be discussed at developmental genes, but a pseudocell inference is not direct same-nucleosome or same-cell co-occupancy. Confirm with an assay designed for joint marks or purified populations before making a strong bivalency claim.

Contents · Previous section · Next section

Section 22 of 36

22. Build developmental trajectories with alternatives

Compare RNA-only, chromatin-only and joint or anchored trajectories. State root, branch rules, smoothing and confidence. Use observed stages and independent markers as checks. If a trajectory changes substantially when one specimen is removed, present it as provisional rather than a universal developmental path.

Contents · Previous section · Next section

Section 23 of 36

23. Challenge antibody-driven differences

A contrast between H3K4me3 and H3K27me3 experiments may reflect reagent efficiency or background. Use positive and negative loci, balanced sequencing, shared reference material and independent lots. Compare within each mark before comparing across marks. Technical asymmetry should not masquerade as biological opposition.

Contents · Previous section · Next section

Section 24 of 36

24. Challenge pseudocell certainty

Change the number of RNA neighbours, feature set and matching threshold. Propagate alignment uncertainty into downstream plots. Show results for directly measured channels beside reconstructed ones. A relationship that exists only under one anchoring choice is a hypothesis, not a stable property of the transition.

Contents · Previous section · Next section

Section 25 of 36

25. Challenge pseudotime direction

Computational order can reverse or bend when roots, branches or sampled states change. Compare with observed developmental stage and known lineage evidence. Repeat across independent specimens. Earlier pseudotime supports temporal reasoning only when the model’s direction is independently grounded.

Contents · Previous section · Next section

Section 26 of 36

26. Challenge enhancer–gene assignments

A chromatin feature near or correlated with a gene may share cell-state timing without regulating it. Genomic distance, coexpression and smoothing can inflate apparent links. Use matched nulls, multiple-testing control, independent accessibility or contact evidence and perturbation before calling an enhancer mechanism.

Contents · Previous section · Next section

Section 27 of 36

27. Validate coupling and uncoupling

If a mark changes before RNA, after RNA or independently, choose an orthogonal assay at the decisive genes and stages. Add imaging, targeted chromatin profiling, expression assays or perturbation. Validation is especially important for reported uncoupling because missing data can imitate temporal separation.

Contents · Previous section · Next section

Section 28 of 36

28. Teach direct and reconstructed evidence early

Primary learners can compare a photograph, a neighbour’s description and a collage made from both. Ask which is directly observed and which is assembled. This playful distinction prepares students to label evidence sources and avoid treating a useful model as the original object.

Contents · Previous section · Next section

Section 29 of 36

29. Practise PSLE Science sequencing

Give students three observed time points and ask whether a change happened between them or was directly seen. Add an inferred intermediate and label it clearly. The exercise strengthens data reading, interpolation and the difference between an observation and a reasonable prediction.

Contents · Previous section · Next section

Section 30 of 36

30. Deepen Secondary Science modelling

Students can draw three layers: measured chromatin, measured RNA and computational alignment. They then list an assumption at every connecting arrow and propose a control. This makes sophisticated bioinformatics approachable through familiar concepts of variables, reliability, models and testable explanations.

Contents · Previous section · Next section

Section 31 of 36

31. Connect to O-Level inquiry

The official 2026 Singapore–Cambridge Biology and Chemistry syllabuses value planning, analysis and evaluation. CoTECH supplies contexts for antibodies, enzymes, nucleic acids, chemical conditions, graphs and uncertainty. The educational aim is to evaluate an evidence chain, not memorise abbreviations.

Contents · Previous section · Next section

Section 32 of 36

32. Make evidence-based school choices

Use official school sources to verify present subject offerings, science programmes, research partnerships, selection criteria and student support. Ask how much access a typical student receives and whether the learning culture suits them. Avoid outdated rankings or assumptions that one named programme guarantees a career outcome.

Contents · Previous section · Next section

Section 33 of 36

33. See careers in connected teams

Chromatin–transcriptome research involves molecular biologists, developmental scientists, computational biologists, statisticians, software engineers, reagent developers, laboratory technologists, data stewards and communicators. Different post-secondary routes can lead into these teams. Reliable habits and transferable skills matter more than one narrow label.

Contents · Previous section · Next section

Section 34 of 36

34. Use tuition to strengthen evaluation

Science tuition is most helpful when it identifies a precise need and measures progress: interpreting graphs, separating observations from inferences, designing controls or writing complete explanations. Use unfamiliar datasets and ask learners to state assumptions. Extra lessons should complement school teaching and healthy routines.

Contents · Previous section · Next section

Section 35 of 36

35. Did you know a pseudocell is useful because it is not pretending?

A pseudocell openly reconstructs complementary chromatin information from transcriptomically similar cells. Its value comes from making an otherwise impossible comparison testable. Its limitation is equally productive: because it is reconstructed, researchers know exactly which direct joint-mark experiment would strengthen the claim.

Contents · Previous section · Next section

Section 36 of 36

36. Finish with labelled evidence layers

CoTECH becomes most persuasive when every figure distinguishes directly measured occupancy, directly measured RNA, aligned neighbours, pseudocell reconstruction, trajectory inference and functional validation. The method matters because it connects chromatin and expression across transitions while giving careful scientists a map of what still needs testing.

A careful CoTECH study should preregister two linked designs: the direct same-cell experiment for each targeted mark and the cross-experiment reconstruction plan. List independent specimens, developmental stages, antibodies, plate balance, target cells, minimum chromatin and RNA information, and primary transitions. State which conclusions use direct occupancy–RNA pairs and which use transcriptome-anchored pseudocells. That distinction is not a weakness. It is the map that tells readers how close each conclusion sits to the original molecule.

Symmetry across targeted-mark experiments is essential. Match tissue preparation, fixation, RNA chemistry, indexing, sequencing goals and biological replication as closely as possible. Include shared reference material or overlapping specimens when feasible. If the H3K4me3 experiment has deeper RNA and the H3K27me3 experiment has stronger chromatin enrichment, an anchor can align technical quality instead of biological state. Report per-mark quality distributions and downsample or model the asymmetry before constructing a combined picture.

Every antibody needs direct validation in the material used. Record clone or catalogue, lot, concentration, incubation and expected positive and negative loci. Quantify enrichment and background, and compare with independent bulk or targeted data where appropriate. H3K4me3 and H3K27me3 are often discussed as active- and repressive-associated marks, but their biological interpretation depends on genomic context. Measure first, describe the local pattern second, and use functional shorthand only with care.

Barcode identity and the two modality funnels should be reproducible. Publish split-pool indexes, read structures, whitelists, mismatch rules, unused combinations, collision estimates and doublet diagnostics. For chromatin, show trimming, mapping, duplicates, target enrichment and accepted fragments. For RNA, show mapping, molecule counting and detected genes. Then show the intersection. A cell with one failed channel cannot support direct occupancy–expression coupling even if it remains useful for a separate analysis.

Transcriptome anchoring requires a declared model. Report selected genes, normalisation, dimensionality reduction, distance metric, neighbour count, mutual-neighbour rules, matching threshold and unmatched cells. Colour anchors by specimen, batch, cell cycle and depth. Use held-out markers or stages to evaluate whether biological state rather than technical similarity drives the pairing. Repeat the alignment under several reasonable feature sets. An anchor that moves dramatically is uncertain and should pass that uncertainty downstream.

Pseudocells need provenance. For each reconstructed unit, retain the source cells, their specimens, targeted marks, distances and weights. State whether the pseudocell averages several neighbours or selects a pair. Never call its two chromatin channels simultaneously measured. Provide direct single-mark cells beside the reconstructed view and change neighbour numbers in a sensitivity analysis. If a reported bivalent pattern disappears when one nearby cell is removed, it is a fragile computational hypothesis, not stable co-occupancy.

Developmental trajectories should be checked against observed time. Build RNA-only and mark-specific orderings before the combined model. State roots, branches, smoothing and confidence, and compare with known stages or lineage evidence. Repeat within independent specimens and leave each specimen out in turn. Naïve-to-primed stem-cell change and endothelial-to-haematopoietic transition both contain branching and timing questions that a single pooled curve can oversimplify. Show disagreement rather than hiding it.

Coupling and uncoupling claims deserve matched evidence. If chromatin occupancy and RNA move together, test whether depth, cell state or smoothing explains the correlation. If they separate in time, ask whether one channel has more dropout or poorer stage coverage. Estimate change points with uncertainty and compare across specimens. Use orthogonal targeted assays at decisive genes and stages. Temporal separation can motivate a mechanistic model, while perturbation is needed to test direction and causality.

Enhancer–gene and bivalency language should be conservative. A nearby or correlated chromatin feature is not automatically an enhancer, and two marks reconstructed in a pseudocell are not proof of the same nucleosome or even the same direct cell. Use contact data, accessibility, known annotations or targeted perturbation as complementary evidence. For direct joint-mark claims, choose an assay designed to observe both features in the same material. A clear limitation makes the next experiment obvious and useful.

Figures should present direct and reconstructed units with different visual encodings. Begin with sample balance, antibody controls, barcode purity and two-channel quality. Continue with direct mark–RNA profiles and RNA-anchor validation. Only then show pseudocells, trajectories and proposed mechanisms. Label observed stage versus pseudotime, direct fragments versus imputed values and cell versus pseudocell. Include source-cell and specimen counts beneath every reconstruction so the reader can see its evidential support.

Reproducibility and stewardship include approvals or consent, specimen provenance, cell culture or tissue conditions, fixation, antibody and enzyme lots, oligonucleotides, plate maps, raw reads, index assignments, direct matrices, anchor graphs, pseudocell provenance, filters, code and environments. Human data need suitable genomic privacy controls. Safety covers biological materials, fixatives, enzymes, heat, sharps and amplified DNA under local procedures. Developmental and animal studies should report reduction, refinement and replicate logic transparently.

CoTECH gives students a particularly useful lesson about models. Primary learners can distinguish an observation from a collage; PSLE learners can label an inferred point between observations; Secondary learners can identify assumptions in an alignment; O-Level learners can evaluate whether evidence supports association, sequence or mechanism. The same ladder helps families discuss school choices and careers without hype. Modern science values people who can connect laboratory work, computing, statistics and communication while keeping every evidence layer labelled.

Close with a robustness grid varying barcode rescue, doublet filters, chromatin and RNA cut-offs, antibody-background thresholds, anchor genes, neighbour counts, matching thresholds, pseudocell weights, trajectory roots, smoothing and specimen inclusion. Mark which cell states, occupancy–RNA relationships, bivalent models and coupling or uncoupling claims persist. Direct findings that recur across specimens deserve confidence. Reconstructed findings that survive several models earn priority for a genuinely joint or perturbational test, which is exactly how responsible science moves forward.

Contents · Previous section · Continue to the Science Learning Hub