Stream: genomics
Topic: Validation issue for HGVS/HGNC code
Fei Yan (Nov 19 2019 at 19:38):
Hi,
I am validating a FHIR resource below:
{
"resourceType": "Observation",
"id": "15008",
"meta": {
"versionId": "1",
"lastUpdated": "2019-11-15T12:37:46.000-06:00",
"profile": [
"http://hl7.org/fhir/uv/genomics-reporting/StructureDefinition/obs-variant"
]
},
"component": [
{
"code": {
"coding": ...
},
"valueCodeableConcept": {
"coding": [
{
"system": "https://genenames.org",
"code": ".HGNC..",
"display": "..."
}
]
}
},
{
"code": {
"coding": ...
},
"valueCodeableConcept": {
"coding": [
{
"system": "https://varnomen.hgvs.org",
"code": "...",
"display": "..."
}
]
}
}
]
}
Once validation I got below errors:
{
"severity": "warning",
"code": "processing",
"diagnostics": "Code https://genenames.org/HGNC:... was not validated because the code system is not present",
"location": [
"Parameters.parameter[0].resource.component[0].value.ofType(CodeableConcept)"
]
},
{
"severity": "warning",
"code": "processing",
"diagnostics": "Code https://varnomen.hgvs.org/... was not validated because the code system is not present",
"location": [
"Parameters.parameter[0].resource.component[2].value.ofType(CodeableConcept)"
]
}
I have loaded the valueset HGNC/HGVS from the IG package (both valueset file size are 1kb). How to resolve this?
I am using HAPI-FHIR 4.0.1 jpa server. Any help will be appreciated.
Thanks,
Fei
Jamie Jones (Nov 19 2019 at 19:41):
Hi there!
I don't know any servers that currently perform validation of HGNC or HGVS codes. In particular, tx.fhir.org, the terminology server that runs in the java validator, isn't aware of the structure of those code systems, so you will always get those warnings using the standard validator.
Jamie Jones (Nov 19 2019 at 19:44):
I would also like to point out that 'obs-variant' is not the most updated version of our 'variant' profile, so it may be harder for us to help you troubleshoot. Our current structure definition for variant is defined at http://build.fhir.org/ig/HL7/genomics-reporting/variant.html, and you may want to check out some of the examples here.
Jamie Jones (Nov 19 2019 at 19:48):
With HAPI, you may be able to upload a CodeSystem resource with the right properties to permit validation on that server. @Patrick Werner may be able to point you in the right direction here, He's given a few tutorials on HAPI, and there is a separate zulip stream for that as well (#hapi ).
Patrick Werner (Nov 19 2019 at 19:52):
yes you could create a resource for these CS, even empty ones if you set the content to not-present
Patrick Werner (Nov 19 2019 at 19:52):
but this still will give you a warning i guess.
Patrick Werner (Nov 19 2019 at 19:53):
You could create a CS resource for HGNC, but HGVS as a grammar is more complicated. This would need a terminology server/tool which is called by hapi.
Patrick Werner (Nov 19 2019 at 19:55):
There are open-source HGVS validators, but none that i could find in java. Which would be easier to integrate. It is an item on my todo list to implement something to validate the more complex CS of our IG with hapi, but it is on there for quite some time already. Not sure when i'll find time for this.
Fei Yan (Nov 19 2019 at 20:03):
@James Jones @Patrick Werner Thanks for your response!
Bret H (Nov 21 2019 at 12:14):
@Fei Yan would you like to comment on https://chat.fhir.org/#narrow/stream/179197-genomics/topic/implementations about your project? Thank you very much for hitting zulip!
May Terry (Dec 08 2019 at 21:06):
Following up on this issue, I see that the FHIR validator on the Bundle-compound-heterozygote.json Variant example provided by the CG Reporting IG generates errors because it cannot validate the code to http://varnomen.hgvs.org and it has a binding strength of "required". Since I'm creating equivalent Variant examples in mCODE, should I just ignore these errors in the mCODE IG QAReport output as well? @Lloyd McKenzie - would this still be acceptable for mCODE STU publication given that it was accepted for the CG Reporting IG as well?
Lloyd McKenzie (Dec 09 2019 at 03:04):
Should be, yes
Last updated: Apr 12 2022 at 19:14 UTC
