HL7 Clinical Genomics Weekly Call - 31 Mar 2020 11:00 AM (US Eastern)
Presiding Co-Chair (Patrick Werner - MOLIT Institut - firstname.lastname@example.org):
- Lloyd McKenzie - Gevity - email@example.com
- Rachel Kutner - Epic - firstname.lastname@example.org
- Joel Schneider - NMDP/CIBMTR - email@example.com
- Stephen Schwartz - Epic - firstname.lastname@example.org
- Kevin Power - Cerner - email@example.com
- Bob Milius - NMDP/CIBMTR - firstname.lastname@example.org
- JD Nolen - Children’s Mercy Hospital - email@example.com
- Liz Amos - NLM - firstname.lastname@example.org
- Scott Robertson - Kaiser Permanente - email@example.com
- Clem McDonald - NLM - firstname.lastname@example.org
- Bret Heale - Intermountain
- Jamie Jones - BCH - email@example.com
- Bob Freimuth - Mayo Clinic - firstname.lastname@example.org
- Bob Dolin - Elimu Informatics - email@example.com
- Dora Finkeisen - MOLIT Institut - firstname.lastname@example.org
Standing Informational Items
Agendas and Important Dates
CG Call Date
Topic 1: ‘find-variant’ operation - outstanding questions (Bob D / Patrick)
Topic 2: Update/Discussion on Implication Profiles (Jamie)
CG IG Lite: Liz & Clem
1: CMS/ONC rules
2: Vote on operations for IG
3: STU2 of IG
1: Reminder - CMS/ONC rules
2: STU2 of IG (cont...)
Tumor/normal examples to review
Working Group Meeting
Cancelled! May 16-22, 2020 • San Antonio, Texas
CG WG will be meeting Monday Q3 & Q4, and all-day Tuesday and Wednesday
- GA4GH Genomic Knowledge Standards (GKS) (leads: Bob Freimuth, Andy Yates)
- GA4GH news
- Having virtual connect meeting starting today,
- agenda has changed, still hope to include some working sessions
- Bob will report back after the meeting
- Variant Representation (formerly VMC) (lead by Larry Babb/Alex Wagner/Reece Hart)
- V1 RC was approved by GA4GH
- Variant Annotation
- Phenopackets on FHIR
- Pedigree Activity (Grant Wood)
- help GA4GH community to understand HL7 Pedigree standards
- Many researchers currently using PED format - family history statement, simple, not as robust as FHIR
- Grant is co-chairing that effort
- G2MC (Grant Wood)
- Looking to develop AI tool for literature review for poly-genetic risk, generate guidance
- Create KB on G2MC website
- HLA reporting (Bob)
- Driver project = GEM Japan
- Modeling and exchanging HLA results in the Japanese population
- HLA nomenclature vs discovery of new alleles
- Connecting to the HLA work done by NMDP within this WG
- ClinGen/ClinVar (Larry Babb, Bob Freimuth)
- Data Exchange/Platform WG, major driver project of GA4GH GKS
- will discuss further when Larry in on a call
- CDISC PGx (Dorina B.)
- CDISC involved in standardizing reporting of data to regulatory bodies. E.g. Pharma communication to FDA - looking to include genomics in research reporting.
- Clem: no specific insight - Transcelerate Biopharma (https://transceleratebiopharmainc.com) group fairly active in FHIR. FDA has active interest in FHIR.
- ONC Sync for Genes (Bob Freimuth)
- ONC website: https://www.healthit.gov/topic/sync-genes
- Final report for Sync for Genes Phase 2 has been finalized (but not posted yet)
- ACI manuscript is in press
- Phase 3 has started - more details to be announced soon
- ISO TC/215 Genomics Subcommittee (Liz, Clem, Bob)
- March 30 - April 3, 2020 in Arlington, VA
- Liz will check to see whether the meeting is open or ISO members only
- Scott R: must be a member of a national member organization to attend
- Clem - request from them to sign-off for transmission of big data (eg BAMs), may have problems with use case.
- Looks to be virtual only
- eMERGE FHIR adoption (Larry Babb, Mullai Murugan, Kevin Power)
- Kevin P: mostly completed their work, have not broadly shared yet. Need to finalize some details.
- Clem: is there a document summarizing their FHIR work?
- Kevin: something presented in past meetings
- Mullai: still work in progress
- May posted STU1 Candidate 1 details to Zulip (Tues Dec 17)
- Information Modeling (IM) (Bob F)
- Minutes: https://docs.google.com/document/d/15kBa3HqxQfwwe0Uwgx3UNashTIanpGRiyRVbimE_j3A/edit#
- Draft model docs: https://docs.google.com/document/d/1Wys14HNJAEB_YJ-EeDPAKX50_oxiDqAKi3WD4wlfjbk/edit
- Export of model (http://gcds.mayo.edu/HL7CG/IM_190604/model.html)
- Currently looking at tooling to develop a logical model from our conceptual model
- Ideally would like to computationally derive it from the conceptual model, but will likely have to create most of it manually
- Scott is evaluating options
- Model-Driven Architecture (MDA)
- Conceptual model => logical model => physical model
- FHIR profiles/IGs would be expressed as a physical model
- Logical: language-neutral model for implementation, specifies relationships, cardinalities, data types that will map down to the physical layer
- Logical model provides a traceable link between our conceptual IM and the physical layer (e.g., FHIR)
- Arthur: does this mean people will have trouble implementing IG will have trouble because it isn't aligned with model?
- Bob F: No, just trying to illustrate more explicitly how our conceptual model maps into the FHIR space
- FHIR (Jamie, Gil)
- Identifying priority FHIR topics for next cycle
- Discussion split off into brainstorming doc for Implications https://docs.google.com/document/d/1SYdzxanCgkwzhhBJCBrguZ6H8HdjQFjG_l2nufAkCGU/edit#heading=h.n5eokak2o6ca
- Preparing proposals
- Reducing annotation/implication profiles Simplications IG STU2
- IG LITE companion - preparing to tighten guidance on Variant - better highlighting/comparing definitional components vs observational, showing which concepts may be derived from/related to others
- Looked at examples for Implications, compared old way vs new way.
- Implication migration spreadsheet of notes from migrating examples
WG projects and outreach
- nothing new
Topic -1 : Virtual Connectathon Track -24/7 Friday to Sunday
Would be done via Zoom, 24/7 Connectathon. Official Information to come
- Create and consume full reports using new structures
- Tumor-normal test report
- WES trio report
- eMerge-style report (with gap analysis)
- mCODE-style report (with gap analysis)
Bob M.: focus on IG STU2 themes for this connectathon, test already implemented STU2 profiles. Identify missing items/profiles in/for STU2.
Bob D.: has an implementation of $find-subject-variants on a VCF FHIR Facade server.Won’t be ready for the may connectathon. Might be ready in september.
Kevin.: even only the first 2 topics would be ok for the connectathon track.
Topic 0: Approval of Minutes from Last Meeting
- Motion: Approve minutes as written
- Move / 2nd: Bob M / Bret
- Vote: (Abstain / Opposed / In Favor): 0 / 0 / 15
- Result: motion passes
Topic 1: Tumor/normal examples for review (continued discussion) (Patrick)
FROM LAST WEEK:
- New extension on Observation (multi-specimen), multiple instances of the extension (references tumor and normal), and then Observation.specimen references tumor
- Could use .focus() ?
- Still need to know which specimen is which (tumor / normal)
- What about multiple Observations?
- Could do 3 (one of tumor, one for normal, one for ‘both’)?
- Typically only send 1 result (related to the tumor)
- Questions outstanding:
- What are these observations saying?
- Are they associated with the specimens correctly?
- Concerns about ‘true nature’ of how the specimens are linked versus some real world ‘how is it done today’
- Conversation can continue on Zulip
- Just link in DiagnosticReport?
JD: Specimen Resource is suitable to handle needed specimen concepts. Question is if DiagnosticReport is a Composition or stand alone.
Bob F.: could you add trio study to the “ultimate example report”. What is a specimen? If digital pathology is a specimen what is NGS data?
Bob F.: would it be helpful that someone from O&O will present their ideas about specimen, DR, Observation, etc.
JD: should we do a meeting in may to exchange ideas O&O < -- > CG? (keep in touch in zulip, do a doodle poll)
Bob D.: JD could describe what O&O is currently doing in the lab space, and this then can be applied/transferred to the IM group
Bret: you could use tumor-normal testing, but then lose the ability to type the Specimen (Blood/tissue type)
Bob D.: specimen is important, but methodology is also important.
Bret: There are some older slides explaining this profile.
Bob F.: Topic is also coming up in G4GH at the moment.
Bret: is trio analysis comparable to tumor-normal testing?
Bob F.: could us relations between specimen or Observations
Will ask May and other Stakeholder for Feedback, continue the discussion after consolidating feedback.
From May via zulip:
Topic 2: Implication Profiles: Merge components?
Current Zulip discussion: https://chat.fhir.org/#narrow/stream/179197-genomics/topic/Associated.20Phenotype.20Nomenclature
Questions from the discussion:
- Can we merge associated-phenotype and associated-cancer?
- In both profiles? (Diagnostic & Therapeutic)
Before we can answer the question if these can be merged, we should define what this component states, provide examples, what is inside and outside of scope.
10:08:30 From Clinical Genomics Work Group : https://docs.google.com/document/d/12-uBrMmav71a3_c9h_FXQteJo_I5Kt72NEBYXZuwhFg/edit#
10:20:22 From Kevin Power : https://confluence.hl7.org/display/FHIR/2020-05+Connectathon+24
10:42:33 From Bret Heale : awesome. what's data and what's style
10:42:38 From Bret Heale : +1
10:45:12 From Bret Heale : please let's record that session
10:45:21 From Bret Heale : and make it cannon
LOINC changes for Level of Evidence / Clinical Significance
See these notes from Swapna: LOINC Significance vs Evidence and TMB code proposal.pdf
Level of evidence
See these previous call notes for earlier discussion: CG-2019-08-27
Need to consider the following two new motions:
- Motion A: Use 53037-8 for both germline and somatic variant clinical significance reporting, and add information to the Term description about the different guidelines for somatic and germline variants; Keep the Answer list the same, but update the type from Preferred to Example
o 1st/2nd -
o Discussion -
o Abstain/Nay/Yea - / /
o Result -
(Notes from Jamie:)
Want to separate clinical significance from level of evidence.
Sites want to continue to use the tier system as well.
Ask: a therapeutic somatic variant:
(note, this example uses placeholder value “AMP Guidelines” instead of a tier)
Code: somatic-predictive (TBD LOINC)
Value: E.g. Resistant, Responsive, Not-Responsive, Sensitive, Reduced-Sensitivity, Adverse Response
Component: Level of evidence -
Value: from LOINC Answer List LL5356-2 (preferred) CAN USE TIER SYSTEM HERE
- Motion B: Consider creating new codes for diagnostic, therapeutic, and prognostic significance (see Quest screenshot, LabCorp report) and/or type of evidence (see Baylor report)
o 1st/2nd -
o Discussion -
o Abstain/Nay/Yea - / /
Need to create an example to understand the meaning of this change/concepts.
Need a caretaker for this topic.
FHIR-25296 Uncallable subregions in a region studied
FHIR-24598 How to reference a region studied observation from genotype, haplotype, variant observations
FHIR-19844 PGx High Risk Allele Medication Impact is confusing
Clinical Genomics Reference Docs
- Mission & Charter
- Currently displayed here:
- Edit here:
- Approved in Sep 2019 WGM
- Review complete as of Aug 1, 2017
- Approved in Sep 2019 WGM
- Decision Making Process
- New DMP: http://www.hl7.org/documentcenter/public/procedures/Default_HL7_WG_DMP_2018.pdf
- DMP Addendum template: http://www.hl7.org/documentcenter/public/procedures/DMP_Modification_Template_2018.docx
- We will review/edit/approve by Sep 2019 WGM