Stream: genomics
Topic: IG - Updated general page
Kevin Power (Sep 23 2021 at 21:29):
Arthur Hermann said:
I converted the page Kevin has been working onto a Word doc which is now on the CG's Google Drive (link follows).
Member's of the WG - PLEASE review this doc and make any comments and suggested edits that you feel should be applied. Why don't we give people two weeks for review (which means all suggested edits/changes should be in place by October 6th. Doc is at: https://docs.google.com/document/d/1xQ4cgvVoXUrKS_jVA4yP_Uy39KJLlu5J/edit?usp=sharing&ouid=112991217946311588811&rtpof=true&sd=true :work_in_progress:
I wanted to break this off into a new thread. If anyone has questions/thoughts/feedback, they are welcome on the document, or feel free to drop them on this thread as well.
Jamie Jones (Sep 23 2021 at 21:35):
I tried to graphically tease apart the derivedFrom and reasonReference references earlier and ran into some questions
Kevin Power (Sep 23 2021 at 21:35):
Fire when ready :smile:
Jamie Jones (Sep 23 2021 at 21:37):
Currently, we have define slices where implications can be delivered from variants, haplotypes, and genotypes. TMB and MSI seem like options as well (could add examples of a therapy match or prognosis based on variant+either)
Jamie Jones (Sep 23 2021 at 21:38):
We don't need to define a slice for every profile we want to permit (since Observation is still an option without validating a profile and slicing is open)
Jamie Jones (Sep 23 2021 at 21:39):
Other clinical observations included as supporting information may also help derive implications.
Kevin Power (Sep 23 2021 at 21:42):
so, do we remove the derivedFrom slices off of Implication but keep it as 1..X to indicate they must derive from something?
Jamie Jones (Sep 23 2021 at 21:43):
Yes I don't know how helpful defining a partial list of slices is.
Kevin Power (Sep 23 2021 at 21:44):
At one point, it was mostly saying 'it must derive from a Genomic Finding' but we have evolved past that now
Kevin Power (Sep 23 2021 at 21:45):
If others agree, we could probably make that part of the resolution to this JIRA?
https://jira.hl7.org/browse/FHIR-32089
Jamie Jones (Sep 23 2021 at 21:48):
Yes. For our recommended followup task, reasonReference needs correcting. In text it was intended to be either a finding or an implication but I think that may also merit being more general. Thoughts?
Kevin Power (Sep 23 2021 at 21:50):
I would agree with that - these things are cases of us trying to be helpful early on, but then evolving the IG without doing a full stack analysis
Kevin Power (Sep 23 2021 at 21:51):
I think our guidance can say findings or implications are the most likely targets for those references, but may not be the only ones.
Kevin Power (Sep 23 2021 at 21:51):
That change can go on:
https://jira.hl7.org/browse/FHIR-32090
Kevin Power (Sep 28 2021 at 22:05):
Hey all - with some great work from @Jamie Jones we have new diagrams to review on the re-vamped general page: Take a peak here:
http://build.fhir.org/ig/HL7/genomics-reporting/branches/kp-task-category-changes/general.html
Still welcome any other feedback. Would like some confirmation (say at least 2-3 other people) to comment here with either approval or suggestions before we go to a vote.
Jamie Jones (Sep 28 2021 at 22:10):
I'll take a pass at the Google doc text next. Would love folks to review if BodyStructure is worth including in the last diagram and if the presentation of DocumentRefence as a target for Related artifact's .resource field with data type= canonical(Any) feels right or should have more detail
Kevin Power (Sep 28 2021 at 22:34):
Jamie Jones said:
Would love folks to review if BodyStructure is worth including in the last diagram and if the presentation of DocumentRefence as a target for Related artifact's .resource field with data type= canonical(Any) feels right or should have more detail
No to BodyStructure.
And perhaps add 'Attachment, DocumentReference' to give a quick summary of the possible options?
Arthur Hermann (Sep 28 2021 at 23:03):
Thank you so much for the wonderful work on the images Jamie! I will review them over the next few days, but they look great. I will also be reviewing the text on the page.. thanks again to both you and Kevin for this great work! • https://media1.giphy.com/media/xT1XGCpi6pKLCQlSr6/giphy.gif?cid=c623cb35mr0y8cd92r8w8p551htnq1nfkxw2wxuor4qwrw3h&rid=giphy.gif&ct=g
Bob Dolin (Sep 29 2021 at 00:11):
@Jamie Jones @Kevin Power This is really nice. What do you think about adding some type of association line between implications and genomic findings/biomarker observations?
Kevin Power (Sep 29 2021 at 12:28):
In which diagram @Bob Dolin ?
Bob Dolin (Sep 29 2021 at 14:10):
Hi @Kevin Power - I was referring to the diagram 'Genomic Observations' in section 1.2.2. Although looking at it again, it might be hard to do, since biomarker observations are grouped with region-studied. But the idea was perhaps to try to convey that the IG conveys [1] genomic findings (variants, biomarkers, etc); [2] test characteristics (e.g. region-studied) that provide context to the findings; [3] implications, that are associated with genomic findings.
Bob Dolin (Sep 29 2021 at 14:10):
something like:
Bob Dolin (Sep 29 2021 at 14:10):
Kevin Power (Sep 29 2021 at 17:09):
@Bob Dolin - I pushed some tweaks. Maybe a little to subtle but take a look.
Bob Dolin (Sep 29 2021 at 17:12):
thanks @Kevin Power . In Figure 'Genomics Implications', can we also show that implications are potentially also derivable from biomarker observations?
Kevin Power (Sep 29 2021 at 18:18):
@Bob Dolin - Take another peek.
Arthur Hermann (Sep 29 2021 at 19:50):
@Jamie Jones @Kevin Power - there is no longer an image 5 which is noted in the last diagram (see image).... correct? if so - needs to be removed or modified snippet.pdf
Kevin Power (Sep 29 2021 at 21:17):
Just pushed up a fix, should build and be ready to test in about 5 minutes.
Kevin Power (Sep 29 2021 at 21:36):
Ready for re-review now.
Arthur Hermann (Oct 04 2021 at 17:01):
A minor matter - but since this page is looking so fantastic! Can we link from this intro paragraph to the appropriate place on the page for these key areas? Overall Interpretations
Genomic Implications
Genomic Findings
Grouper
Region Studied
Other Observations
Recommended Actions
Contextual Resources
Kevin Power (Oct 04 2021 at 17:08):
oh, from that table? Yea, I can do that.
Kevin Power (Oct 04 2021 at 17:53):
@Arthur Hermann - links added.
Kevin Power (Oct 06 2021 at 15:16):
Reminder to #genomics - I would like to vote on this page by next Tuesday if possible, so any additional input would be great to have by this week.
Mullai Murugan (Oct 08 2021 at 17:15):
Apologies for the late response, hope it is not too late. @Kevin Power I tried reading this from the perspective of an unfamiliar new user and it was so very helpful! Thank you for doing this, Kevin! Great work! Here is my feedback on a few minor items -
- Use of grouper - with the concept of grouper now in question, I am not sure if it should be included here. Thoughts?
- The second para in section 1.2.1 references large or complex genomic reports - it could just be me, but I am not clear on what constitutes a large or complex report and how extends or summaryOf can be used. Maybe include an example in the GenomicReport profile and reference the example here?
- Region Studied - instead of saying mutation level, should we mention that chromosomal position?
Kevin Power (Oct 08 2021 at 18:13):
Thanks Mullai! Never too late :smile:
Mullai Murugan said:
- Use of grouper - with the concept of grouper now in question, I am not sure if it should be included here. Thoughts?
I considered this, but since we have Grouper in the IG still, and no plans to completely remove it, it seems like it should be included here. I could flag it with a 'hey this might change soon' comment?
- The second para in section 1.2.1 references large or complex genomic reports - it could just be me, but I am not clear on what constitutes a large or complex report and how extends or summaryOf can be used. Maybe include an example in the GenomicReport profile and reference the example here?
I don't know that I have a good example to include here I'm afraid, short of the short guidance we also provide:
This is particularly useful when different labs or services are performing later steps in the analysis, for example.
We mostly wanted to make sure implementers were aware of these extensions that are built into DiagnosticReport.
- Region Studied - instead of saying mutation level, should we mention that chromosomal position?
I think the 'mutation level' is referring to the gene-mutation component of the profile. I do wonder if the current guidance of It can provide this description as a list at the gene level, mutation level, or as a textual description. is OK, or if only providing the 3 options when there are others. Perhaps we should make that statement more generic, so something like It can provide this description in a number of ways, so review the components of the profile for your use case
Mullai Murugan (Oct 08 2021 at 18:36):
Kevin Power said:
Thanks Mullai! Never too late :)
Mullai Murugan said:
- Use of grouper - with the concept of grouper now in question, I am not sure if it should be included here. Thoughts?
I considered this, but since we have Grouper in the IG still, and no plans to completely remove it, it seems like it should be included here. I could flag it with a 'hey this might change soon' comment?
- MM - I think that is a good idea. What about also including a plug-in to composition if we decide that is going to be the new way?
- The second para in section 1.2.1 references large or complex genomic reports - it could just be me, but I am not clear on what constitutes a large or complex report and how extends or summaryOf can be used. Maybe include an example in the GenomicReport profile and reference the example here?
I don't know that I have a good example to include here I'm afraid, short of the short guidance we also provide:
This is particularly useful when different labs or services are performing later steps in the analysis, for example.
We mostly wanted to make sure implementers were aware of these extensions that are built into DiagnosticReport.
- Region Studied - instead of saying mutation level, should we mention that chromosomal position?
I think the 'mutation level' is referring to the gene-mutation component of the profile. .....```
Jamie Jones (Oct 08 2021 at 18:37):
Perhaps it can provide this description at the known level of detail using components
Mullai Murugan (Oct 08 2021 at 18:46):
Re item 3 - Generic, with some examples might be helpful. It could be just my perception, but don't want folks to think that these are the only options. As a reference lab, we always provide actual coverage at the position level, didn't want users to think that option is not available. @Jamie Jones you are referring to this item in your comment , below, yes?
Kevin Power (Oct 08 2021 at 18:56):
For Region Studied, I will use the wording from @Jamie Jones
@Mullai Murugan For Report/composition - I don't think we should yet, as that option is only added as part of R5, and we still build on R4. Anyone disagree with that?
Kevin Power (Oct 08 2021 at 19:09):
Change for Region Studied has built (same branch location).
Last updated: Apr 12 2022 at 19:14 UTC
