Stream: genomics
Topic: genomics "coding systems" and HTA
May Terry (Jun 04 2021 at 16:45):
Several JIRA negative ballot comments were issued against mCODE for genomics-related "vocab issues" - namely, a missing "coding system" request for the following: SequenceOntology, clinvar, and GTR. Since mCODE's genomics elements are aligning with the CG Reporting IG, I feel that this is an issue that really needs to be driven and resolved by the HL7 genomics working group rather than mCODE. Here are the HTA tickets that someone on our team started.
HTA-37. Sequence Ontology. 3/25/21
HTA-42. ClinVar 6/4/21.
HTA-43. NCBI Genetic Testing Registry identifier.
@Patrick Werner or @Joel Schneider - can either of you, or whoever is now handling genomics vocabulary, weigh in on these tickets? In the short-term, we need URLs to get past code librarians who are more interested in process than the actual issue or content at hand.
It does however bring a bigger issue of how our models and FHIR constructs can better distinguish knowledgebase content identifiers (e.g.: a GTR test id), nomenclature conventions (e.g.: HGVS), whether these should even be populated as a "coding system". Or just thinking aloud, it might need to even be a new data type.
Genomics is in the radar for most of this discussion because knowledgebases and variant representation are at core of many of our bindings and to treat them like a terminology coding system to me sounds pretty ridiculous.
cc: @Mark Kramer
Joel Schneider (Jun 04 2021 at 18:44):
Looks like these tickets are about external code systems which have not yet undergone formal review by the HTA.
I found Sequence Ontology and ClinVar Variant ID code systems listed on the FHIR code systems page, under Code Systems for Genetics, but did not see GTR listed there.
https://hl7.org/fhir/terminologies-systems.html
Joel Schneider (Jun 04 2021 at 18:52):
SO and ClinVar are also listed on the Genomics Reporting IG's external coding systems page.
http://build.fhir.org/ig/HL7/genomics-reporting/codings.html
May Terry (Jun 04 2021 at 19:51):
thanks @Joel Schneider ! So...thoughts on GTR? is this something that we should consider in CGWG?
Joel Schneider (Jun 04 2021 at 21:00):
CGWG seems like a reasonable starting point, but HTA owns the official process of engaging with external code system owners. Not sure if GTR should be classifed as a code system or identifier system.
https://confluence.hl7.org/display/TA/External+Code+System+Owner+Engagement+Process
https://confluence.hl7.org/display/TA/External+Terminologies+-+Information
May Terry (Jun 04 2021 at 22:27):
Registries like GTR does seem more like it should be an identifier system, but as you mentioned, I suppose we'll just see what HTA says once they pull that ticket out of the queue. Thanks Joel!
Joel Schneider said:
CGWG seems like a reasonable starting point, but HTA owns the official process of engaging with external code system owners. Not sure if GTR should be classifed as a code system or identifier system.
https://confluence.hl7.org/display/TA/External+Code+System+Owner+Engagement+Process
https://confluence.hl7.org/display/TA/External+Terminologies+-+Information
Patrick Werner (Jun 06 2021 at 15:42):
As Joel already pointed out: SO and Clinvar are already available.
Patrick Werner (Jun 06 2021 at 15:45):
Regarding the "what is a codesystem" discussion i think all of them are CodeSystems.
Patrick Werner (Jun 06 2021 at 15:45):
Everytime you can bind a list to a value, or check validity through a grammar. To me this is a CodeSystem.
Patrick Werner (Jun 06 2021 at 15:47):
Identifiers are used to identify Instances, and not abstract objects like a Test.
Patrick Werner (Jun 06 2021 at 15:48):
If you "convert" GTR into ObservationDefinitions, the GTR.id would go into ObsDef.code, and not identifier.
Patrick Werner (Jun 06 2021 at 16:02):
An Identifier System to me would be smth. like a passport n° or driver license n°.
Robert McClure (Jun 16 2021 at 16:39):
Just finding this thread. Specifically regarding GTR, I agree with Patrick that it is a code system, not an identifier system. As he notes, instance identifiers identify a single thing. not an idea that can be used in many places. Plus GTR tests can be versioned (see Hereditary Breast and Ovarian Cancer Panel) and that is consistent with code systems although rare at the concept level. Who is your terminology facilitator that can take the lead on all this? They need to reach out to HTA for some help and provide guidance on how to correctly structure this.
I think you all have some pretty challenging terminology representation issues so you need someone fully engaged.
Joel Schneider (Jun 18 2021 at 17:35):
I'm the nominal terminology facilitator for CG, and can facilitate moving these forward. At the moment, we appear to have 3 active mCODE-related HTA tickets in JIRA.
HTA-37 Sequence Ontology (SO)
HTA-42 ClinVar
HTA-43 GTR
The CG IG's External Coding Systems appendix lists over 2 dozen code systems. Some of the list items were identified during the V2 LRI work. It appears most of them have not yet gone through a formal HTA process. Any advice on how to proceed?
Bret H (Oct 11 2021 at 22:19):
@Robert McClure howdy. Did you get to connect with Joel?
Bret H (Oct 11 2021 at 22:22):
@Joel Schneider I think this page will help: https://confluence.hl7.org/display/TA
Joel Schneider (Oct 11 2021 at 22:29):
@Bret H Were you interested in a specific request?
Bret H (Oct 11 2021 at 22:31):
Not me. Thought you were looking for getting progress on the HTA jira tickets from Jun 16th. Are they closed?
Joel Schneider (Oct 11 2021 at 22:33):
Looks like two of them are still in "submitted" status and one is "ready for review".
Robert McClure (Oct 12 2021 at 16:37):
@Joel Schneider Best thing to do is attend the HTA call to clarify what you need help in getting done. Next call is Oct 20 3p ET. @Carol Macumber @Reuben Daniels It may be that HTA will become a part of TSMG but not yet.
Joel Schneider (Jan 20 2022 at 17:01):
HTA has published draft content for the Sequence Ontology (SO) external code system here:
Sequence Ontology (SO)
Please review and add any coments to HTA-37
Joel Schneider (Jan 20 2022 at 17:07):
HTA has published draft content for the ClinVar Variation ID external code system here:
ClinVar Variation ID
Please review and add any comments to HTA-42
There is also an open question regarding this one:
Joel Schneider the URI that was noted on the FHIR terminologies page is http://www.ncbi.nlm.nih.gov/clinvar which does not distinguish variant ID from allele ID. You mentioned that this might not be widely in use. Can you please query your genomic work group if it would be disruptive to update this URI to be specific for the Variant ID. For example, though not resolvable (which isn't required), it would be more meaningful (though also not required) to alter to something like: http://www.ncbi.nlm.nih.gov/clinvar/variantid and allele ID (if/when requested) would be http://www.ncbi.nlm.nih.gov/clinvar/alleleid
Note: Oid registry has two separate entries for variant 2.16.840.1.113883.6.319 and allele 2.16.840.1.113883.6.318
Joel Schneider (Jan 20 2022 at 17:09):
HTA has published draft content for the Genetic Testing Registry (GTR) external code system here:
Genetic Testing Registry (GTR)
Please review and add any comments to HTA-43
Bret H (Jan 20 2022 at 20:11):
regards ClinVar Variation ID
The Allele ID is an identifier for each individual nucleotide change in the set of variants that were interpreted. In other words, if the variation report is for interpretation of a haplotype defined by five SNPs, each SNP has its own Allele ID. The set of SNPs that make up the haplotype are represented by the Variation ID.
However they provide a search by the Variation ID. E.g. https://www.ncbi.nlm.nih.gov/clinvar/variation/1331421/?new_evidence=true
So "https://www.ncbi.nlm.nih.gov/clinvar/variation/" could be used for the variation id system. working on allele :working_on_it:
Bob Dolin (Jan 20 2022 at 20:15):
Here's a relevant comment from @Larry Babb https://chat.fhir.org/#narrow/stream/179197-genomics/topic/more.20examples/near/177298844
Bret H (Jan 20 2022 at 20:16):
so "Larry Babb5:43 PM
Please don't use the clinvar allele ID. " makes me think we need to update our use to indicate only the variation ID is allowed and use the url https://www.ncbi.nlm.nih.gov/clinvar/variation/
what do folks think?
Bret H (Jan 20 2022 at 20:24):
Most importantly, the clinvar variation id is what is referenced when people use clinvar ids. So, I think we're safe to not support allele id at this time.
Patrick Werner (Jan 25 2022 at 07:02):
Bret H said:
so "Larry Babb5:43 PM
Please don't use the clinvar allele ID. " makes me think we need to update our use to indicate only the variation ID is allowed and use the url https://www.ncbi.nlm.nih.gov/clinvar/variation/what do folks think?
i think it is good to be precise . +1 for the /variation
Joel Schneider (Feb 02 2022 at 19:51):
The HL7 Terminology Authority (HTA) group is interested in feedback from the Clinical Genomics workgroup regarding this draft code system info page for ClinVar Variation ID:
https://confluence.hl7.org/display/TA/ClinVar+Variation+ID
Please add any formal comments regarding this to the corresponding HTA ticket:
https://jira.hl7.org/browse/HTA-42
The HTA-42 ticket is on the agenda for the January 10 HTA meeting, so it would be best to submit any comments before then.
Joel Schneider (Feb 02 2022 at 19:54):
Correction: the next HTA meeting is on February 10 (not January 10).
Patrick Werner (Feb 07 2022 at 09:46):
I like the proposed new uris for variant and allele id
Bob Milius (Feb 07 2022 at 16:41):
Joel Schneider said:
The HL7 Terminology Authority (HTA) group is interested in feedback from the Clinical Genomics workgroup regarding this draft code system info page for ClinVar Variation ID:
https://confluence.hl7.org/display/TA/ClinVar+Variation+IDPlease add any formal comments regarding this to the corresponding HTA ticket:
https://jira.hl7.org/browse/HTA-42The HTA-42 ticket is on the agenda for the January 10 HTA meeting, so it would be best to submit any comments before then.
@Larry Babb @Larry Babb
Last updated: Apr 12 2022 at 19:14 UTC
