Possible mismatch on genomic representation for NM_004260.4:c.2296C>G on GRCh37

[gcampof]

Describe the issue

I’ve encountered a discrepancy in the genomic representation of the transcript variant NM_004260.4:c.2296C>G when using the VEP GRCh37 online tool, and I would appreciate your input on whether this is expected behavior or a potential issue.
 
When I input this variant into VEP (GRCh37 web tool), I get the following genomic representation:
8-145738768-G-C
(job can be found here)
This appears to be a simple base substitution. However, when I use VariantValidator, the output for GRCh37 is:
8-145738768-GG-C
which is interpreted as a delins.
 
On further investigation, we observed that the GRCh37 reference genome contains an extra G at the position 145738768 compared to the transcript NM_004260.4. 
    • Is the current VEP output (8-145738768-G-C) considered accurate in this context, or is it a simplified representation that does not take such transcript-genome alignment differences into account?
    • Does VEP account for alignment gaps or discrepancies between transcript and genomic references when generating HGVSg output?
    • Would a representation like 8-145738768-GG-C be more appropriate in this case?
Any clarification on how VEP handles such cases, and whether this is expected behavior, would be greatly appreciated.

Best regards,
Guillem 

Additional information

Screenshot of the aforementioned region in hg19 visualized using IGV

System

• VEP version: GrCh37 web tool

[dglemos]

Hi @gcampof,

For input NM_004260.4:c.2296C>G, the vep output location is 145738769 (not 145738768).
Here are the HGVS notations outputed by VEP:

HGVSg=8:g.145738769G>C
HGVSc=NM_004260.4:c.2296C>G
HGVSp=NP_004251.4:p.Arg766Gly

You can check the variant representation in dbSNP: https://www.ncbi.nlm.nih.gov/snp/rs1586802958#hgvs_tab

In your screenshot there is a transcript represented - is that NM_004260?

Is the current VEP output (8-145738768-G-C) considered accurate in this context, or is it a simplified representation that does not take such transcript-genome alignment differences into account?
Does VEP account for alignment gaps or discrepancies between transcript and genomic references when generating HGVSg output?

The correct output is 8-145738769-G-C which takes into account any mismatch between RefSeq transcript and reference genome. In the output, this information can be found under BAM_EDIT. You can read more about correcting transcript models here: https://www.ensembl.org/info/docs/tools/vep/script/vep_other.html#refseq_bam

Would a representation like 8-145738768-GG-C be more appropriate in this case?

This representation does not look correct. You could contact the VariantValidator team to understand why it is represented as a delins.