Biology · Genomics, Omics & Bioinformatics
Genome annotation, genes, transcripts and functional uncertainty
One imagined twelve-position locus has two annotated transcript paths. Their retained sequence pieces differ even though they refer to the same locus. Beside them, a functional ledger distinguishes direct support, an inferred role, an empty entry and an explicit limited negative statement. Neither two transcript records nor four ledger rows means four genes.
Gene Expression: RNA & Proteins explains how sequence, RNA and protein are related. Phylogeny, Classification & the Tree of Life introduces ancestry and the interpretation of trees. Research Design, Data & Scientific Practice explains sampling, independent units and the limits of comparisons. Their English chapters are available in this working manuscript; publication remains pending. The examples here use letter strings, inclusive coordinates, averages, fractions and ordinary graph reading. Each convention is introduced before it is used.
Back to the Biology mapBy the end, you can
Read structural and functional annotations as evidence-backed interpretations, keeping loci, transcript variants and missing evidence distinct.
A feature needs a location and a represented entity
Structural annotation identifies features such as genes, exons and transcript models on a sequence representation. It connects a proposed biological structure to a coordinate system. Coordinates need an assembly or sequence version and a convention for endpoints; otherwise the same numbers can refer to different bases. A predicted feature is an interpretation of evidence, rather than a new piece of DNA created by writing its label.
Genes and transcripts have different identities. A locus can give rise to different transcripts, and not every transcript encodes a protein. Some transcript records represent alternative retained pieces; some include untranslated regions. Identifying a transcript's boundaries or sequence does not by itself determine its coding region, protein product or biological role.
The original locus uses one-based inclusive coordinates and a single stipulated orientation. Its two transcript models share positions one through three, then use either seven through nine or ten through twelve. The sequences are written as DNA-letter representations of retained transcript pieces. No coding frame, start site or protein function is assigned.
Two transcript paths belong to one stipulated locus
Read the diagram in words
ACG at positions one to three joins GCA at seven to nine for ACGGCA, or TTC at ten to twelve for ACGTTC. One locus has two assigned paths, each retaining six positions.
| Stipulated locus | ACGTTAGCATTC · + |
|---|---|
| Inclusive end position | 12 |
| Represented transcript | DNA-letter representation of retained pieces | Total retained positions |
|---|---|---|
| Transcript representation one | ACGGCA | 6 |
| Transcript representation two | ACGTTC | 6 |
| Represented transcript | Retained segment | Inclusive start position | Inclusive end position | DNA-letter representation of retained pieces |
|---|---|---|---|---|
| Transcript representation one | 1 | 1 | 3 | ACG |
| Transcript representation one | 2 | 7 | 9 | GCA |
| Transcript representation two | 1 | 1 | 3 | ACG |
| Transcript representation two | 2 | 10 | 12 | TTC |
| Stipulated locus | 1 |
|---|---|
| Represented transcript | 2 |
No protein-coding boundaries assigned
Original twelve-position locus, one-based inclusive endpoints and one supplied forward orientation. Retained transcript pieces are represented in DNA letters, not a claim of actual RNA sequencing. No coding frame, protein translation, splice motifs, real gene or procedure.
A functional label is a claim with a basis
Functional annotation associates an entity with an activity, process or location. A useful annotation preserves what supports the association: a direct observation, a computational comparison, a phylogenetic inference or a curatorial judgement, for example. These evidence types answer different questions. They are not interchangeable labels for certainty.
In the original evidence ledger, one product has a direct activity observation in the stated context. A second has a role inferred from a sequence comparison. A third has no recorded annotation for that role. The ledger deliberately uses local teaching categories rather than real database entries, so its simple contrasts can be read without mistaking them for assessed genes.
A direct observation may support a narrow activity while leaving other tissues, conditions or roles untested. An inference can be useful while depending on the comparison and its assumptions. An automated annotation can be correct; a manually reviewed annotation can later change. Evidence and context help assess a claim without converting the database's presentation into proof.
A statement and its evidence remain connected
Read the diagram in words
Two rows make positive role statements with different support. One has no captured statement; one gives a context-limited negative statement. Four rows are not four established positive functions.
| Represented product | What is claimed | Basis supplied in the record | Scope of the claim |
|---|---|---|---|
| Product one | Assigned activity supported | Direct observation stipulated | The stated observation context |
| Product two | Assigned activity proposed | Sequence comparison stipulated | Inference depends on the comparison |
| Product three | No recorded activity statement | No evidence captured in this ledger | Reason for the missing entry unspecified |
| Product four | Assigned activity absent in the stated context | Specific negative observation stipulated | Other roles and contexts remain open |
| What is claimed ✓ | 2 |
|---|---|
| No recorded statement | 1 |
| Explicit limited negative statement | 1 |
Evidence category is not a universal confidence rank
Four fictional products and local teaching evidence categories. No real GO identifier, experimental result, assay, automated confidence score or universal functional conclusion is assigned.
Missing information is not a negative biological result
A missing annotation means that the selected record contains no such statement. It need not mean that the product lacks the activity or that nobody knows anything about it. Information may be uncaptured, the evidence may concern a different entity, or the relevant work may not yet exist. The reason for the gap matters.
An explicit supported negative statement is different. In the ledger's fourth row, a specified activity is reported absent under a stated observation context. That statement is not a universal claim that the entity does nothing. Positive evidence, explicit negative evidence and no recorded evidence must remain separate when a table is summarised.
Annotations are revisable. New observations, corrected structures or improved relationships can change their interpretation. Recording the annotation release and the represented entity lets a later reader understand what claim was available when an analysis was done. A changed annotation does not necessarily imply that the underlying sequence changed.
Worked example
Trace two transcripts and four evidence states
The supplied locus string is ACGTTAGCATTC. Positions one through three are ACG, seven through nine are GCA, and ten through twelve are TTC. The first retained transcript representation is ACGGCA; the second is ACGTTC. Each has six retained positions.
There is one stipulated locus and two transcript models. Six retained positions do not identify a six-base protein-coding region: no coding annotation was supplied. The difference between the paths is a structural comparison, not a claim about their amounts or function.
The ledger contains two positive role statements with different evidence, one missing statement and one explicit negative statement in a limited context. Counting all four rows as four established positive functions would erase the distinctions that make the record interpretable.
Check your understanding
Do two transcript models in this example establish two separate genes?
Show reasoning
No. Both paths are assigned to one locus and share the same first retained segment. Distinct transcripts are different represented products of the locus, not automatically different genes.
Why can a six-position transcript representation not be translated into a claimed protein here?
Show reasoning
The model supplies retained sequence pieces but no coding frame or coding boundaries. Transcript sequence length and protein-coding length are different quantities, and some transcripts are noncoding.
What distinguishes a missing functional entry from the ledger's negative entry?
Show reasoning
The missing entry supplies no recorded statement. The negative entry supplies a particular context-limited statement with evidence. Neither can be silently treated as a universal absence of all function.
What to carry forward
Keep the entity, coordinates, evidence and context with an annotation; a blank field is not a biological verdict.
References
Resources supporting this lesson’s explanations and examples.
- Gene Ontology annotation documentation (Web documentation inspected 13 October 2026)Gene Ontology ConsortiumOriginal consortium documentation
- Stable IDs (Ensembl release 116, June 2026 archive)Ensembl/EMBL-EBIOfficial archived data documentation
- Gene (Official glossary, inspected 13 October 2026)National Human Genome Research Institute; Eric Green narrationAgency glossary