Why a Variant Can Have More Than One Consequence
A variant is recorded at a genomic position, but labels such as missense, intronic, or splice donor describe its relationship to a particular transcript. A gene may have several transcripts with different exon structures, transcription start sites, or coding regions. Consequently, one genomic variant can fall in a coding exon of one transcript, an intron of another, and outside a third.
Consequence labels are useful summaries of where a variant lies and what sequence change it implies under a specified annotation. They are not, by themselves, measurements of the variant’s biological effect, evidence that it changes a phenotype, or a clinical classification. Interpreting a label requires knowing the reference genome assembly, the transcript used, and the annotation conventions.
Transcript Structure and Consequence Labels
A transcript is an RNA sequence produced from a gene. In a processed messenger RNA, exons are joined and introns are removed. Some exons contribute to the protein-coding sequence; other exon sequence may be in a 5′ or 3′ untranslated region (UTR), which is part of the transcript but is not translated into protein. A gene can produce multiple transcripts that include different exons or use different coding start and stop positions.
The transcript’s strand matters. On the forward strand, transcript sequence is read in the same orientation as increasing genomic coordinates; on the reverse strand, the transcript is read in the opposite orientation. Thus genomic “left” and “right” do not always correspond to a transcript’s 5′ and 3′ ends. Annotation tools use transcript orientation to determine coding changes, UTR location, and splice donor or acceptor positions.
| Label family | What the label indicates | Important qualification |
|---|---|---|
| Missense variant | A coding-sequence substitution changes an encoded amino acid. | It describes the predicted amino-acid change, not whether protein function changes. |
| Synonymous variant | A coding-sequence substitution does not change the encoded amino acid. | It does not establish that the variant has no biological effect. |
| Stop-gained variant | A coding change creates a premature termination codon. | The resulting transcript or protein effect depends on context, including transcript structure. |
| Frameshift variant | A coding insertion or deletion changes the reading frame. | This label applies when the net coding-sequence length change is not divisible by three. |
| In-frame insertion or deletion | A coding insertion or deletion changes sequence length by a multiple of three. | One or more amino acids may be added or removed without shifting the downstream reading frame. |
| Splice donor or acceptor variant | A variant overlaps a defined splice donor or acceptor region. | Exact region boundaries depend on the annotation convention; a label alone does not show the actual RNA product. |
| Splice-region variant | A variant lies in a designated region near a splice junction. | It is broader than a label for a core donor or acceptor position. |
| UTR, intronic, or noncoding-exon variant | A variant lies in the corresponding transcript region. | These location labels do not specify a protein amino-acid change. |
Other commonly encountered labels include stop-lost (a change removes a termination codon), start-lost (a change affects the annotated translation start codon), and intergenic (a position lies between annotated genes under the annotation’s rules). Labels and their precise definitions may differ among tools and annotation releases. In particular, splice-region windows and the prioritization of multiple consequences should be checked in the documentation for the annotation system being used.
Worked Example: One Variant, Two Transcript Consequences
Hypothetical data. In a specified genome assembly, a single-base substitution changes genomic base T to C at one position in a gene on the forward strand. Transcript A includes that position in a coding exon. Its coding-strand codon is TTA before the change and TCA afterward. Transcript B includes the same genomic position in an intron.
For Transcript A, TTA encodes leucine and TCA encodes serine. The substitution therefore changes the encoded amino acid from leucine to serine, so the consequence for Transcript A is missense.
For Transcript B, the position is in an intron rather than in its coding sequence. Its consequence for that transcript is therefore intronic under the stated transcript structure. The variant has not changed location in the genome; the transcript models differ.
The two labels do not conflict. They answer different questions: what does this position imply for Transcript A, and what does it imply for Transcript B? Neither label alone establishes whether either transcript is expressed in a relevant tissue or whether the variant changes a phenotype.
How Coding Labels Follow From the Reading Frame
Protein-coding sequence is read in successive groups of three nucleotides called codons. Each codon specifies an amino acid or a termination signal. A coding substitution can therefore leave the amino acid unchanged (synonymous), change it to another amino acid (missense), or create a termination codon (stop-gained). An insertion or deletion can alter the grouping of all downstream bases.
Worked Example: An In-Frame Deletion Versus a Frameshift
Hypothetical data. A transcript’s coding sequence begins GAA CCT GAC TAA, where spaces mark codon boundaries. Consider two separate deletions in this coding sequence. Assume for this illustration that the stated bases are deleted from the coding sequence and that the transcript annotation is otherwise correct.
Deletion of CCT. This removes three coding bases, so the net length change is \(3\), and \(3\) is divisible by \(3\). The sequence becomes GAA GAC TAA. The reading frame remains aligned: GAA and GAC are still complete codons. The deletion is in-frame; the deleted codon had encoded one amino acid. This label does not imply that the deletion is harmless.
Deletion of one C from CCT. This removes one coding base, so the net length change is \(1\), which is not divisible by \(3\). After that position, the sequence is regrouped from the changed sequence rather than in the original triplets. The annotation is therefore a frameshift. The downstream amino-acid sequence may change, and a new stop codon may occur, but its position cannot be determined from the deleted base alone; the downstream sequence is needed.
The comparison uses the coding sequence, not merely the number of genomic bases between two coordinates. For complex variants, an annotation must account for the net sequence change in the transcript.
Splice and Noncoding Consequences
Splicing joins selected exons and removes intervening introns. The splice donor is at the exon-to-intron boundary, and the splice acceptor is at the intron-to-exon boundary, described in transcript orientation. A variant annotated as affecting a donor or acceptor overlaps a region defined by the annotation convention. A splice-region label generally refers to a broader nearby region. These labels identify a position relative to splice machinery; they do not prove that splicing is altered or specify which alternative RNA product, if any, is produced.
A variant may instead be labelled as lying in a 5′ UTR, 3′ UTR, intron, or noncoding-transcript exon. These are descriptions of transcript location, not synonymous with “no consequence.” For example, UTR sequence is transcribed but not translated as part of the annotated protein. A variant there has no direct coding amino-acid substitution under that transcript’s annotation, though the location label alone does not settle whether there is a regulatory or other effect.
Worked Example: Interpreting a Splice-Site Label
Hypothetical data. In a specified annotation release, a variant falls at the first intronic base immediately after an exon in a transcript’s orientation. The annotation convention defines the first two intronic bases at that boundary as the splice donor region. The transcript has a coding sequence in the exon before the boundary.
Under these stated rules, the variant receives a splice donor consequence for this transcript because it overlaps the defined donor region. The location is sufficient to explain the label; it is not sufficient to conclude that the exon will be skipped, that an intron will be retained, or that the mature RNA will be abnormal.
To describe the consequence more specifically, one would need evidence about the resulting transcript, not just the positional label. A different annotation convention could use different region boundaries, so the convention and release should be reported when the distinction matters.
Transcript Choice and Annotation Practice
A consequence report should be read as a set of transcript-specific statements, not as one universal property of a genomic position. A transcript choice may change whether the variant is in an exon or intron, whether an exon is coding, which codon is affected, or whether the affected sequence is in a UTR. A gene can therefore have a high-impact coding label for one transcript and a noncoding or intronic label for another.
Annotation depends on more than transcript choice. The genome assembly determines the coordinate system; the reference and alternate alleles determine the change being described; and transcript records and software versions determine the sequence model and consequence rules. For reproducibility, a report should identify the genomic assembly and, when relevant, the transcript accession and version, along with the annotation source or release. A transcript described as “canonical” or “representative” is a selected reference for reporting, not proof that it is the only biologically relevant transcript.
Worked Example: A UTR Label on One Transcript and Coding on Another
Hypothetical data. A genomic substitution lies in an exon shared by two transcripts of the same gene. Transcript C has a longer annotated coding sequence that includes the position. Transcript D uses a different coding start, so the shared exon sequence at this position is in its 5′ UTR. Assume the annotation records both transcripts and the genomic alleles are correctly matched to the reference assembly.
For Transcript C, the position is in coding sequence. Its label must be determined from the specific codon and alternate allele; if the substitution changes the encoded amino acid, the transcript-specific label is missense. For Transcript D, the position is in the 5′ UTR, so its label describes a UTR location rather than a coding amino-acid substitution.
Thus “exonic” alone is not enough to infer “coding,” and “coding” alone is not enough to infer “missense.” The transcript’s coding boundaries and the codon change are both needed. If only one transcript is reported, the selected transcript should be made explicit rather than allowing its label to appear universal.
Common Mistake: Treating the Label as a Functional Verdict
A common error is to reason, “This variant is annotated as missense, so it damages the protein,” or, “It is synonymous, so it has no effect.” The correction is that these are sequence consequence labels. Missense says that the annotated codon changes its amino-acid specification; synonymous says it does not. Neither statement alone measures protein activity, RNA abundance, disease risk, or clinical significance. Likewise, a splice-site label identifies a position relative to a splice boundary, not an observed splice defect.
A second error is to report one transcript’s consequence without naming or documenting the transcript choice, especially when another transcript receives a different label. Preserve the transcript context, check the annotation convention, and distinguish the genomic variant from each transcript-specific description. Assessment of in silico functional prediction is a separate question from understanding these consequence labels.
Check Your Understanding
Answer each question using transcript-specific reasoning.
- Why can the same genomic variant be labelled missense for one transcript and intronic for another?
- A coding deletion removes six bases. What can be concluded about the reading frame, and what cannot be concluded from the length change alone?
- How does a splice donor label differ from evidence that the transcript is actually mis-spliced?
- Why does an exonic label not necessarily mean that a variant changes a protein sequence?
- Name two pieces of annotation context that should accompany a transcript-specific consequence report.