Protein nomenclature for insertions that encode a translation stop codon
David
Dear HGVS team,
I would like to submit a particular point of ambiguity in the HGVS nomenclature guidelines concerning insertions that encode a translation stop codon.
Example 1:
- NM_013976.5(GCDH):c.1074_1075insAGTTGAAGGAC
This insertion results in the addition of SerTer after Asp358. The translation stop codon is encoded in the insertion.
The Variant Validator, Mutalyzer and VEP tools represent this alteration as a frameshift:
p.(Gln359Serfs*2)
Example 2:
Modified from example 1 by adding one nucleotide:
- NM_013976.5(GCDH):c.1074_1075insAGTTGAAGGACT
Since the nucleotide is added after the encoded termination signal, it
does not impact the actual protein modification. However, the
insertion's total length is now a multiple of three, leading to its
classification as an insertion rather than a frameshift:
p.(Asp358_Gln359insSer*)
These two examples highlight how, for these prediction tools, the "multiple of three" rule plays a determinant role for the frameshift determination, while the presence of an encoded stop codon does not. In contrast, the HGVS guidelines do not involve the "multiple of three rule". Instead, they involve the open reading frame status of the inserted sequence.
The guidelines define frameshift as follows:
"Frameshift: a sequence change between the translation initiation
(start) and termination (stop) codon where, compared to a reference
sequence, translation shifts to another reading frame."
From the Frameshift, Insertions, Deletions-Insertions, and Duplications sections, it is stated that: inserted sequences on DNA or RNA level that:
- encode a translation stop codon in the inserted sequence are on the protein level described as an insertion of this sequence, not as a deletion-insertion removing the entire C-terminal amino acid sequence.
- encode an open reading frame which after the inserted sequence shifts to another reading frame, are described as a frameshift.
Notably, the emphasis on the term “after” (in the text) suggests that
it is crucial for the classification that the sequence after the
insertion continue to be translated.
The two insertions reported above share the following characteristics:
- they do not encode an open reading frame
- they do encode a translation stop codon in the inserted sequence.
- the translation of the affected protein (if it actually occurs in vivo) does not actually shift to another reading frame.
- they result in the same protein alteration
Without involving the "multiple of three" rule, It is unclear whether NM_013976.5(GCDH):c.1074_1075insAGTTGAAGGAC should be represented as a frameshift p.(Gln359Serfs*2) or simply as an insertion p.(Asp358_Gln359insSer*).
Since none of the examples provided in the guidelines directly address
this scenario, I suggest adding the "non-multiple-of-three" case as a
new example in the Frameshift section. This would help for the
determination of similar situations, including Duplications, and
Deletion-Insertions, reducing subjective interpretation and ensuring
consistent application of the HGVS nomenclature.
Kind Regards,
David Hernandez
Johan den Dunnen
Johan den Dunnen
HUGO HGVS Variant Nomenclature Committee (HVNC)
David
I hope this message finds you well.
I would like to follow up on the question I raised in May 2024 regarding the correct HGVS description of DNA insertions that encode a stop codon. At that time, you confirmed that such variants should be described as insertions, regardless of whether the inserted sequence length is a multiple of three.
However, this point remains a source of confusion within the community. Major tools such as Variant Validator, Mutalyzer, and VEP still appear to apply the "multiple of three" rule as the primary criterion for deciding between a frameshift and an insertion (or a delins), without considering whether the inserted sequence encodes a stop codon.
This suggests that the current HGVS guidelines are not sufficiently clear on this point. To help avoid inconsistent variant descriptions across tools, databases, and publications, I would like to kindly reiterate my request for clarifying the guidelines.
This could be addressed by adding an explicit example of a non-multiple-of-three insertion that encodes a stop codon and should nevertheless be described as an insertion.