Why Science? | scSeqComm, Intercellular–Intracellular Signalling Paths and Functional Evidence

Section 1 of 36

1. Start with two evidence layers

scSeqComm joins intercellular ligand–receptor evidence between cell groups to intracellular signalling paths in receiver cells. This helps ask not only who might signal to whom, but whether the receiver contains a plausible route towards target genes and functional response.

Contents · Next section

Section 2 of 36

2. Keep intercellular and intracellular claims separate

Ligand–receptor compatibility and receptor-to-target paths answer different questions. One concerns exchange between cells; the other concerns possible processing inside the receiver. Publish both components so a strong score in one layer cannot hide weak evidence in the other.

Contents · Previous section · Next section

Section 3 of 36

3. Use the peer-reviewed method boundary

The scSeqComm study appeared in Bioinformatics in 2022, with an R package, official project page and archived submitted version. Record the software and network releases used. Later updates should be described separately from the workflow evaluated in the paper.

Contents · Previous section · Next section

Section 4 of 36

4. Understand the incoming route

The intercellular layer uses ligand expression in a source group and receptor expression in a receiving group together with a curated interaction resource. It prioritises compatible routes. RNA abundance does not directly measure ligand secretion, receptor protein or binding.

Contents · Previous section · Next section

Section 5 of 36

5. Understand the receiver path

The intracellular layer follows signalling knowledge from a receptor towards downstream genes in the receiver. A path is an eligible sequence in a prior network. It does not demonstrate that every intermediate was present, modified in the correct direction or active at sampling time.

Contents · Previous section · Next section

Section 6 of 36

6. Use functional characterization as a guide

Functional analysis organises target genes and pathways linked to inferred communication. Declare the gene universe, annotation source and statistical correction. Enrichment supports interpretation and experiment design; it does not prove the nominated function was caused by the ligand.

Contents · Previous section · Next section

Section 7 of 36

7. Treat combined evidence as prioritisation

A candidate supported by both intercellular and intracellular layers is more coherent than an isolated expression product. The combination still depends on cell labels, resources, expression thresholds and network topology. It narrows hypotheses rather than closing the case.

Contents · Previous section · Next section

Section 8 of 36

8. Create a specimen-first manifest

Record organism, tissue, condition, donor or animal, assay, batch, cells, annotations, preprocessing and experimental pairing. Preserve specimen identity throughout group summaries. Independent biological replication is judged across specimens, not across thousands of nested cells.

Contents · Previous section · Next section

Section 9 of 36

9. Audit cell annotations

Source and receiver labels determine the communication graph. Check markers, doublets, transitions and uncertain mappings. Repeat leading routes with a broader hierarchy. An over-split receiver can create a seemingly precise path supported by very few cells.

Contents · Previous section · Next section

Section 10 of 36

10. Freeze ligand–receptor resources

Save the exact eligible pair list, species mapping, direction and complex rules. Report how many pairs enter before and after filtering. Database changes can alter the network without any biological change in the specimen.

Contents · Previous section · Next section

Section 11 of 36

11. Freeze intracellular networks

Archive the signalling and regulatory networks used to connect receptors to targets. Record edge direction, sign when available and provenance. A shortest or highest-scoring path is a model choice, not an observation of molecular traffic.

Contents · Previous section · Next section

Section 12 of 36

12. Control expression and dropout

Sparse single-cell counts can hide receptors and downstream genes or exaggerate cell-group differences. Show detection fractions, distributions and specimen prevalence. Avoid aggressive imputation that creates smooth paths without corresponding measurement support.

Contents · Previous section · Next section

Section 13 of 36

13. Plan thresholds and multiple testing

Declare expression filters, score cut-offs and functional enrichment correction before interpreting results. A pipeline searches many source–receiver, ligand–receptor and pathway combinations. Selective reporting from this large space can manufacture certainty.

Contents · Previous section · Next section

Section 14 of 36

14. Archive a reproducible run

Retain the expression object, metadata, labels, databases, intracellular networks, parameters, seeds, score tables, path tables, enrichment results and plotting code. The final network must remain traceable to the cells, molecules and prior edges that produced it.

Contents · Previous section · Next section

Section 15 of 36

15. Practise with an invented path table

This fictional table is for reasoning practice, not scSeqComm output. Compare a candidate with strong intercellular and intracellular support, one with receptor but no downstream path, one with a path but weak ligand evidence and one that fails spatial validation.

Contents · Previous section · Next section

Section 16 of 36

16. Inspect ligand and receptor evidence

For every leading route, plot source ligand and receiver receptor expression by specimen and cell. Check protein form, essential complex subunits and annotation confidence. A group average can hide a receptor present in only a tiny or contaminated subset.

Contents · Previous section · Next section

Section 17 of 36

17. Inspect receptor-to-target intermediates

List the intermediate signalling and regulatory nodes connecting receptor to target genes. Test whether they are expressed and whether known direction agrees with the inferred response. A database path containing absent or contradictory intermediates should be downgraded.

Contents · Previous section · Next section

Section 18 of 36

18. Read scores component by component

Publish the intercellular score, intracellular-path evidence and functional result separately before presenting a combined priority. Similar final rankings can arise from different weaknesses. Component reporting tells the laboratory which link should be tested first.

Contents · Previous section · Next section

Section 19 of 36

19. Compare specimens before conditions

Calculate route and path support within each donor or animal, then summarise recurrence. Pooling can make one specimen’s unusual source cells appear universally important. Show heterogeneity and avoid treating cells as independent replicates.

Contents · Previous section · Next section

Section 20 of 36

20. Use spatial data as orthogonal support

The paper used spatial transcriptomics as supporting validation in its case study. In a new dataset, verify source–receiver proximity or appropriate diffusion distance. The original spatial result does not transfer automatically to another tissue or experimental system.

Contents · Previous section · Next section

Section 21 of 36

21. Use time to test ordering

If the proposed route causes a receiver programme, receptor-proximal events should precede target-gene and functional changes. Sample appropriate intervals and preserve subject pairing. One endpoint cannot distinguish an incoming signal from a pre-existing receiver state.

Contents · Previous section · Next section

Section 22 of 36

22. Audit resource sensitivity

Repeat headline routes with a high-confidence ligand–receptor subset and an alternative compatible intracellular network. Mark which molecules, paths and functions persist. Resource disagreement is useful evidence about model dependence.

Contents · Previous section · Next section

Section 23 of 36

23. Audit score and threshold sensitivity

Vary one planned expression or score threshold and track ranking stability. A route that appears only at a permissive cut-off should be labelled threshold-sensitive. Do not search many settings and report only the version producing a desired pathway.

Contents · Previous section · Next section

Section 24 of 36

24. Compare transparent baselines

Compare scSeqComm with a simple expression-product score and a method using only intercellular evidence on the same inputs. The added intracellular layer earns trust when it yields reproducible, discriminating predictions—not merely a larger, more complicated network.

Contents · Previous section · Next section

Section 25 of 36

25. Publish broken paths and contradictions

Show strong ligand–receptor candidates lacking downstream support, coherent paths without source ligand, spatially incompatible routes and expected mechanisms not recovered. Negative examples calibrate what the combined workflow filters out.

Contents · Previous section · Next section

Section 26 of 36

26. Validate proteins and early responses

Confirm ligand and receptor proteins, physical opportunity and a receptor-proximal event before relying on later target genes. A downstream transcriptional programme may have many causes. Early measurements reduce ambiguity.

Contents · Previous section · Next section

Section 27 of 36

27. Design a path-discrimination experiment

Choose two incoming ligands that predict different receptors, intermediates or target programmes. Perturb them under matched timing and measure the earliest distinguishing events. Include a result that rejects both proposed paths.

Contents · Previous section · Next section

Section 28 of 36

28. Write a bounded path claim

Prefer: ‘scSeqComm prioritised a replicated ligand–receptor route with a declared receptor-to-target path and functional support.’ Avoid saying molecular information flowed along that path until intermediates and response have been observed.

Contents · Previous section · Next section

Section 29 of 36

29. Build a classroom relay model

Give students cards for sender, ligand, receptor, two intracellular steps and a target gene. Remove one card and ask what claim still survives. The activity shows why a complete-looking arrow depends on multiple distinct observations and prior-knowledge links.

Contents · Previous section · Next section

Section 30 of 36

30. Begin with Primary Science sequences

Primary Science teaches ordered processes, interactions and fair comparisons. scSeqComm adds a modern evidence ladder from an external signal to internal response. Students can practise marking which steps were measured and which were inferred.

Contents · Previous section · Next section

Section 31 of 36

31. Use PSLE Science to identify conclusions

PSLE Science asks learners to connect observations to mechanisms without overstating them. With an inferred path, ask what the assay measured, what the database supplied and which new experiment would separate competing explanations.

Contents · Previous section · Next section

Section 32 of 36

32. Connect Secondary and O-Level Biology

Secondary Science and O-Level Biology develop receptors, coordination, cells, genetics and experimental evaluation. scSeqComm links these topics through a computational network. The central lesson is that a plausible path needs molecular and functional validation at several levels.

Contents · Previous section · Next section

Section 33 of 36

33. Let Computing preserve provenance

Computing contributes graphs, path algorithms, scoring, databases, version control and reproducibility. Biology supplies direction, molecular context and experimental tests. A correct path search through an inappropriate network remains a wrong scientific model.

Contents · Previous section · Next section

Section 34 of 36

34. Choose opportunities by verified fit

For school choices and enrichment, verify current official programme information. Look for sustained work in biology, computing, data interpretation and practical investigation. Do not infer admissions, mentorship or career outcomes from fashionable labels.

Contents · Previous section · Next section

Section 35 of 36

35. See connected career families

The reasoning connects bioinformatics, systems biology, signalling research, biostatistics, data engineering and laboratory science. Different education routes emphasise different combinations. Requirements change, so confirm current prerequisites with the responsible institution.

Contents · Previous section · Next section

Section 36 of 36

36. Finish with a bounded scientific claim

A defensible conclusion reports the ligand–receptor evidence, receiver path, target functions, network versions, specimen recurrence, spatial feasibility and perturbation results. scSeqComm joins intercellular and intracellular hypotheses coherently; it does not directly observe molecular transfer or causal information flow.

scSeqComm asks a useful two-layer question. First, is there evidence for a ligand–receptor route between cell groups? Second, does the receiving cell contain a plausible intracellular path from that receptor towards genes and functions of interest? Joining the layers can sharpen a hypothesis, but it also joins their uncertainties.

Did you know? A plausible receptor-to-target path can exist in a curated network even when none of its intermediate proteins were measured as active. Network reachability is a map of possible routes, not a video of molecules moving through the receiver cell.

Preserve the layers separately before combining them. Report the extracellular ligand–receptor evidence, the intracellular path and the functional annotation in distinct columns. Then readers can tell whether a strong result is supported throughout or driven mostly by one layer.

Cell-group labels must be auditable. Ambient RNA, doublets and transitional states can create spurious ligand or receptor evidence. Repeat key routes under conservative quality filters and a broader cell hierarchy. If a finding disappears when one fragile subcluster changes name, describe it as label-sensitive.

The intracellular network is a versioned prior. It favours well-studied pathways and may omit context-specific regulation. Store every edge’s source and confidence, then rerun leading findings with high-confidence or alternative networks. Network-sensitive paths are still hypotheses, but they need clearer labelling.

Direction matters at both layers. The source cell supplies the ligand, the receiving cell supplies the receptor, and the internal path runs towards targets in that receiver. A reversed cell label or reversed network edge can create a biologically polished but incorrect story. Hand-audit a small set end to end.

RNA detection is only one gate. Ligand processing and release, receptor surface localisation, phosphorylation and transcription-factor activity may not track transcript abundance. Select orthogonal assays for the weakest unmeasured step instead of repeating another RNA measurement.

Functional characterisation should not outrun the gene list. Publish which target genes support each function, the reference universe and the correction method. Broad terms can appear significant because of overlapping annotations. A specific function with transparent membership is more useful than a long decorative enrichment list.

Specimen-aware analysis remains essential. Compute route or pathway support within biological specimens and report prevalence and uncertainty. Pooling cells first can let one donor or tumour dominate the signal while making the result appear universally replicated.

Spatial transcriptomics can support anatomical plausibility, as used in the published study, but spatial RNA is not direct molecular transfer. Match resolution and mediator class to the claim. Coarse spots containing several cell types require deconvolution and careful uncertainty.

Timing can discriminate a path. After stimulating or blocking the source route, measure a proximal receptor event, an intermediate regulator and later target expression. If the supposed target changes before receptor activation, the proposed direction needs revision.

Thresholds at either layer can change the joined path. Show continuous component scores and repeat the analysis across planned expression and network cut-offs. A conclusion should not depend on the exact point where two noisy values become binary.

Compare the joined model with simpler baselines: ligand–receptor score alone, target-expression change alone and direct receptor–target association. scSeqComm adds value when the integrated path produces a stable, discriminating prediction that those components do not provide separately.

Negative controls should preserve realistic structure. Use incompatible source–receiver compartments, abundance-matched non-ligands, receptor paths ending in irrelevant targets or cell-label permutations within specimens. Each control tests a different way the joined model could appear coherent by chance.

Perturbation can be staged. Block the ligand or receptor, then assay the earliest predicted intracellular event before examining broad functional output. A rescue or alternative-ligand test can distinguish route specificity from a general stress response.

The tumour-microenvironment application in the original paper demonstrates workflow feasibility, not universal transfer. Recheck disease type, cell states, specimens, tissue architecture and network coverage. Published validation belongs first to the published dataset.

Archive expression matrices, cell labels, ligand–receptor resource, intracellular network, functional annotation, thresholds, scores and software versions. The final path diagram is not enough to reproduce the analysis or understand why a candidate appeared.

A balanced report includes one route supported at both layers, one extracellular-only signal, one network-sensitive intracellular path and one negative control. This makes the method’s added value visible without hiding disagreement.

For learners, scSeqComm connects transport and response in biological systems with graphs and algorithms. Observation, variables, fair testing and source evaluation begin in school Science; Biology and Computing extend them into molecular pathways and reproducible analysis.

The bounded conclusion is that scSeqComm can link candidate intercellular exchange to curated intracellular paths and functions, producing more coherent hypotheses for follow-up. It does not directly observe secretion, receptor activation, path traversal or causation. Those claims require matched protein, spatial, temporal and perturbation evidence.

Keep an explicit evidence ledger for every headline route so that new protein, spatial or perturbation results can strengthen or reject one layer without silently rewriting the others.

Contents · Previous section · Continue to the Science Learning Hub