Is it possible to interpret HGVS without transcript version number?

Question

[curious]

I have:

"TranscriptRefSeq" = NM_000110

"DNAChange" = c.1432G>T

iiuc the HGVS is the union of these:

NM_000110:c.1432G>T

I want to find the the VCF description of this variant (eg GRCh38:1:97549652:C:A).

I plug NM_000110:c.1432G>T into variant validator and get:

"RefSeq variant accession numbers MUST include a version number"

If I add a version number it works: NM_000110.4:c.1432G>T

My question is, is it not even possible to map HGVS without the version number with variant validator or otherwise? My data does not provide this version number :/

[GenoMax]

You can lookup the version history of that accession here: https://www.ncbi.nlm.nih.gov/nuccore/NM_000110?report=girevhist

Since the changes in v.3 and v.4 are substantial it may be tricky to interpret the changes without version numbers.

If you simply search with NM_000110 you will always get the newest revision.

[curious]

Thanks, it goes beyond that as I have ~700 of these transcript accessions, none of which have versions. I would guess in the general case you need the version, please let me know if you think otherwise

[GenoMax]

If you need to lookup the version numbers of accessions, you could do it using EntrezDirect:

$ esearch -db nuccore -query NM_000110 | efetch -format acc
NM_000110.4

If you have 700 then provide a couple more examples and I can show you how to look those up using epost method.

Looks like you need version numbers for the Variant validator site.

[curious]

Below is a sample of what I have, I am a bit confused though you mentioned that the changes for NM_000110.4 and NM_000110.3 are substantial, how do I know then for example which version the author of the below list was referring to when they wrote NM_000110 (or are you suggesting I just guess it is the most recent?).

I am concerned that I will run into cases where where different cdna changes correspond to different VCF description changes ( eg GRCh38:1:97549652:C:A) depending on which transcript accession version they are associated with. Or is this not an issue?

NM_000110   c.2527A>G
NM_000038   c.4534G>A
NM_000321   c.733C>T
NM_015950   c.742dup
NM_013986   c.1943A>G
NM_015559   c.4001-3del
NM_012289   c.607G>A
NM_006218   c.3140A>T
NM_020975   c.2753T>C
NM_004119   c.972C>A
NM_000051   c.1661C>T
NM_005228   c.1787C>T
NM_001040108    c.2138del
NM_001015877    c.442G>T
NM_015338   c.1900_1922del
NM_001127208    c.238C>T
NM_003738   c.275G>A
NM_024675   c.2200A>T
NM_183416   c.1982C>T
NM_000368   c.1109C>G
NM_006180   c.1159+1G>A
NM_001080463    c.3482A>G
NM_001195427    c.289G>A
NM_018948   c.1028A>G
NM_001291415    c.1801A>T
NM_002015   c.266C>A
NM_020529   c.382G>A
NM_000400   c.713A>C
NM_016320   c.1618G>C
NM_016578   c.337G>C

[GenoMax]

If this analysis was done in the last year or so, you can (at some risk) assume that the latest version of the accession is being referenced (especially if you have no way to ask whoever generated this data). We can't really tell you one way or the other. If this is related to patient care you will want to triple check everything.

Here is how you can get the latest version numbers for accessions.

# Put accessions one per line in a file

$ more id_file
NM_000110
NM_000038
NM_000321
NM_015950
NM_013986
NM_015559
NM_012289

# Get the latest version numbers using Entrezdirect

$ epost -db nuccore -input id_file | efetch -format acc
NM_000321.3
NM_015950.5
NM_012289.4
NM_000038.6
NM_015559.3
NM_000110.4
NM_013986.4

[Zhenyu Zhang]

You need to ask whoever give your the data which gene model version they are using. If that's not available, the best you can do is to assume all 700 are from the same version of gene model, and then look up for all recent gene model versions to see which one has all 700 in it. If you find multiple gene model versions, you are out of luck.