Stream: genomics
Topic: obs-structural-genetic-findin
Patrick Werner (Jul 29 2019 at 15:58):
This topic came up during todays call: We are using the Variant profile for all kind of variants: SNV, smallInDels, structural variants. But we also have : http://build.fhir.org/ig/HL7/genomics-reporting/obs-structural-genetic-finding.html
Patrick Werner (Jul 29 2019 at 15:58):
Which is confusing.
Patrick Werner (Jul 29 2019 at 16:01):
The question came up: why are structural Variants sent using the Variant profile, but only structural Variants which are ISCN coded have their own (minimalistic) profile?
Patrick Werner (Jul 29 2019 at 16:02):
Our lab doesn't use ISCN, so for us structural variants are sent using the variant profile.
Bret H (Jul 29 2019 at 16:03):
just info for folks who need it: component.code can be LOINC and component.value can also be set. Much like observation.code and observation.value.
Jamie Jones (Jul 29 2019 at 16:22):
Would we want/need to set up any Invariants on Variant? If component for ISCN is populated, certain ones should not be, etc?
Jamie Jones (Jul 29 2019 at 16:23):
If we felt it would be helpful to do that it may be an argument against including it in the larger profile...
Kevin Power (Jul 29 2019 at 17:12):
If it is more straightforward for now, I would be OK moving the ISCN component to our Variant profile. Which sounds like what was discussed on the call today. It is already quite a catch-all, so might as well expand it a bit more until people start using it and help figure out our boundaries.
Patrick Werner (Jul 29 2019 at 18:49):
i can see us doing derived profiles from variant for specific sub types instead of doing too much invariant magic in the future.
Patrick Werner (Jul 29 2019 at 18:51):
which will bring up the variant type discussion again :smile:
Kevin Power (Jul 29 2019 at 19:30):
I would agree that sub-profiles probably make more sense. Besides, invariants on Variant sounds like we are approaching the event horizon of a black hole. :hole:
Jamie Jones (Jul 29 2019 at 21:25):
Updated inheritance diagram for genetic observations:
pasted image
Jamie Jones (Jul 29 2019 at 21:34):
our ISCN profile doesn't currently inherit from ComputableGeneticFinding so it's unclear if it officially counts as a "genetic finding" or not (for referencing by implications/interpretation/etc in terms of Figure 2 on General Reporting, e.g.).
Kevin Power (Jul 29 2019 at 21:37):
It is a finding (or so I would claim). It probably doesn't inherit from it because at least two of the components (gene and cytogenetic location) are not needed for our ISCN result. So, if we combine into Variant, that would be additional guidance we would need to provide.
Last updated: Apr 12 2022 at 19:14 UTC
