Transcript Consequences and Coding-Variant Ambiguity

Why a Coding Variant Can Have More Than One Consequence

A variant is often summarized with a short label such as missense or loss of function. Those labels are not properties of a genomic position alone. They describe what the alternate allele is predicted to do in relation to a particular transcript and its encoded protein. At a locus with alternative transcripts, the same variant can fall in a coding exon of one transcript, an untranslated region of another, and an intron—or outside the transcript—of a third.

The relevant object is therefore not just “the gene.” It is a specified transcript, including its exon structure and coding sequence. A consequence annotation also depends on the variant representation and reference sequence used. In the examples below, assume each hypothetical variant has already been checked against the correct genome assembly, reference allele, and normalized representation. The remaining ambiguity comes from transcript choice.

Core Coding Consequence Labels

The CDS is the portion of a transcript translated into protein. Coding-DNA coordinates conventionally begin at the first base of the translation start codon, which is c.1; coordinates do not include upstream untranslated sequence. Each group of three coding bases forms a codon. The genetic code maps each codon to an amino acid or a translation stop signal.

• Synonymous variant: a coding substitution changes a codon but not the amino acid it encodes. “Synonymous” describes the protein sequence prediction; it does not guarantee that RNA processing or expression is unaffected.
• Missense variant: a coding substitution changes a codon so that it encodes a different amino acid.
• Stop-gained variant: a substitution changes an amino-acid codon into a stop codon, potentially ending translation early.
• Frameshift variant: an insertion or deletion changes the number of coding bases by a length that is not divisible by three. The reading frame downstream of the change is altered.
• In-frame insertion or deletion: a coding insertion or deletion changes sequence length by a multiple of three, so the reading frame is retained, although amino acids may be added or removed.
• Start-lost or stop-lost variant: a change disrupts the annotated translation start or stop, respectively. The biological outcome can depend on alternative initiation, transcript structure, or other sequence features.

Annotation tools may use labels such as high impact for predicted stop-gained, frameshift, or certain splice-site consequences. Such categories summarize a consequence model, not a measured amount of functional damage. A frameshift or premature stop can be strong evidence for a potentially disruptive coding effect, but whether it reduces functional protein depends on details such as transcript relevance and the location of the change. For example, a premature stop may or may not lead to degradation of the transcript; its position relative to exon boundaries is one factor, not a guarantee of that outcome.

Locating a Coding Change Within a Codon

For a coding-DNA position c.n, its codon number and position within that codon can be calculated directly. This is useful for checking whether a substitution changes the first, second, or third base of a codon before translating the reference and alternate codons.

The subtraction of one makes c.1 the first base of codon 1. Integer division groups bases 1–3 into codon 1, bases 4–6 into codon 2, and so on. The remainder identifies the base within the group. This calculation locates a base, but the consequence still requires the actual reference codon, alternate allele, and genetic-code translation.

Transcript Choice Changes the Interpretation

Transcripts from one gene can differ because they use alternative promoters, splice sites, or terminal exons. These differences can change the start of the CDS, the length and composition of coding exons, and the resulting protein. Consequently, the same genomic variant can map to different coding-DNA coordinates in different transcripts. It can also be coding in one transcript but untranslated, intronic, or absent from another.

A transcript selected as “canonical” or “principal” is a useful reference for consistent reporting, but that label does not make it universally most relevant. Transcript selection should be stated, and where the biological question specifies a tissue or a clinically established transcript, that context matters. Curated transcript sets such as MANE Select can provide standardized transcript choices, but they do not remove the need to consider relevant alternative transcripts.

The following example uses two transcripts in which the same genomic alternate allele maps to different coding positions and codons. The coding-DNA coordinates differ because the transcripts have different coding structures upstream of the variant.

Alternative Exons and Apparent Loss of Function

An exon can be included in one transcript and skipped in another. A variant in that exon might therefore disrupt the protein-coding sequence of one isoform while lying in an intron of another. “Isoform” is often used for a distinct transcript or protein product from the same gene; usage can vary, so the actual transcript accession and structure are more precise than the word alone.

A coding frameshift illustrates why the transcript context matters. If an insertion or deletion changes CDS length by one or two bases, the downstream codons are read in a different frame. This can change many amino acids and may introduce a premature stop. If the affected exon is excluded from a second transcript, that transcript will not have the same coding frameshift consequence. It is not valid to generalize a transcript-specific consequence to every product of the gene.

Resolving Conflicting Annotations

Different annotations are not necessarily errors. They may be correct descriptions of the same allele on different transcripts. Before treating an annotation set as contradictory, check the transcript accessions and versions, strand, coding coordinates, and whether each transcript contains the affected exon. A change in transcript version can also alter exon boundaries or the CDS, so transcript identifiers without versions may not fully specify the sequence being interpreted.

A practical comparison proceeds in a consistent order:

This comparison separates two questions: “What is the predicted consequence on each transcript?” and “Which transcript is most relevant to the interpretation?” Annotation establishes the first. Answering the second may require tissue-specific expression, clinical curation, or other functional evidence; a canonical label alone does not answer it.

Common Mistakes and Why the Distinction Matters

• Wrong reasoning: “The variant is in the gene, so it has one gene-wide coding consequence.” A gene can produce transcripts with different coding exons and reading frames. Correction: state the consequence per transcript, then identify the transcript chosen for a primary summary.
• Wrong reasoning: “Synonymous means harmless.” A synonymous substitution does not change the encoded amino acid on that transcript, but the label alone does not establish that RNA processing or expression is unchanged. Correction: describe the specific protein-sequence prediction and avoid inferring no biological effect from that label alone.
• Wrong reasoning: “A high-impact or loss-of-function annotation proves complete gene inactivation.” Such a label is a predicted consequence on an annotated transcript. Correction: distinguish the prediction from evidence that the transcript is relevant, that functional protein is lost, or that a phenotype results.
• Wrong reasoning: “Two tools disagree, so one must be wrong.” They may have used different transcript sets, transcript versions, or reporting conventions. Correction: compare the underlying transcript and sequence assumptions before comparing the labels.
• Wrong reasoning: “The canonical transcript is always the biologically correct one.” A canonical designation standardizes a choice but does not guarantee relevance in every tissue or application. Correction: report the selected transcript and consider relevant alternatives.

Transcript-aware interpretation matters because a coding label can influence which variants receive attention in a locus. If an annotation collapses all transcripts into one gene-level label, it can hide both uncertainty and the biological assumptions behind the prioritization. Clear reporting preserves that information without treating a predicted consequence as a measured functional result.