Stream: genomics
Topic: HGNC code
Bob Milius (Oct 18 2018 at 15:32):
When looking up a gene name in HGNC, it displays the code for a gene as prefix of HGNC: followed by a number. Is the prefix an integral part of the code? e.g. https://www.genenames.org/cgi-bin/gene_symbol_report?hgnc_id=4931 shows HGNC id = HGNC:4931. But the url shows hgnc=4931
What should be captured as the valueCodeableConcept.coding.code for Gene Studied? HGNC:4931 or 4931?
Just want to establish best practices...
Kevin Power (Oct 18 2018 at 16:07):
I think this is the sort of thing @Patrick Werner and @Julian Sass want to provide guidance on in the ValueSets document they are working on. This might be a good comment to make on their Google Doc
Julian Sass (Oct 18 2018 at 16:24):
yes I think we talked about HGNC during the Baltimore WGM, that we want to use the HGNC ID with prefix e.g. <code value=”HGNC:4931”/> and use the symbol as display value
Bob Milius (Oct 18 2018 at 16:37):
I asked HGNC for guidance. Here's their response.
I would use the ID with a prefix as it makes it easier to identify that the ID is for a HGNC gene report. The prefix is also fine to use in the URL (ie https://www.genenames.org/cgi-bin/gene_symbol_report?hgnc_id=4931 and https://www.genenames.org/cgi-bin/gene_symbol_report?hgnc_id=HGNC:4931) both work)
Bret H (Oct 19 2018 at 12:55):
Totally agree with HGNC, copying my commemts from email thread,
The code is the best thing to use. The code 4931 will always point to the same gene. The approved symbol changes from time to time and there are examples where different genes share a symbol as a synonym. Even with the potential for indicating the wrong gene, people will still often use the gene symbol.
But for a true identifier that is guaranteed to point to one specific gene, use the code (such as HGNC: 4931).
If needed, I am sure the folks at HGNC would be willing to comment and affirm the above. (Which you and they did)
Bret H (Oct 19 2018 at 13:00):
Added comment to spreadsheet
Bob Milius (Oct 19 2018 at 14:35):
I asked HGNC if they genenames.org supported other namespaces besides HGNC. Here's their response:
As previously written officially the gene ID should have HGNC: as a prefix however it is not necessary to use the prefix to link to the page in the URL. I urge you to use the HGNC: however as this does identify the ID as a HGNC gene ID without having to associate it with the URL. Our beta site will be replacing the current site in November and as default the URL will have the prefix HGNC: (e.g https://beta.genenames.org/data/gene-symbol-report/#!/hgnc_id/HGNC:1097) as standard (however again it will work if the HGNC: is dropped).
We do have other numerical IDs within the HGNC site that are unique identifiers for the gene family/group records (https://www.genenames.org/cgi-bin/genefamilies/set/654). Here the ID does not have a prefix so using HGNC: for the gene records helps to clearly identify one as a gene ID and the other as a family/group. We also have a sister site called the VGNC that names vertebrate genes. If in future we decide to merge the two sites into one, the prefix will be important to distinguish a human gene ID from a vertebrate gene ID.
By using the prefix we are also following on with the tradition within nomenclature committees in adding a prefix for the resource as part of the ID as the same style of ID can be seen within the Mouse (http://www.informatics.jax.org/marker/MGI:88190) and Rat (https://rgd.mcw.edu/rgdweb/report/gene/main.html?id=RGD:619908) nomenclature resources.
Bob Dolin (Oct 19 2018 at 15:05):
I kinda think it's up to the code system to decide what their codes are, but I do see this as differing from, say, SNOMED or LOINC. I wonder if the folks at genenames.org know that in HL7, a code is ALWAYS coupled with a code system? On the other hand, it sounds like the "HGNC:" prefix may be necessary to disambiguate similar numeric strings that don't represent genes.
Grahame Grieve (Oct 23 2018 at 08:17):
yes it's not clear quite what signficance of it is. The important thing is :
Grahame Grieve (Oct 23 2018 at 08:18):
. I urge you to use the HGNC: however as this does identify the ID as a HGNC gene ID
as compared to what else?
Andrew Patterson (Oct 23 2018 at 10:53):
gene families are also assigned integer codes in the hgnc data. So the AGPAT family of genes is given the family code 46. Which would make an ideal little terminology grouping construct - but the integer space used for the family ids and the gene ids have not been kept distinct.
So just did a quick search and HGNC:1625 = XCR1.. there is also a gene family 1625 = AGC family kinases
Last updated: Apr 12 2022 at 19:14 UTC
