CG WG Call Notes leading to 2016 January WGM

November 3, 2015 -- Weekly call

Agenda

• FHIR proposals on new resources and profiles
  • 5 min Updates from
    • ClinGen
    • GA4GH
    • IOM
    • FHIR Lead
    • other

• • FHIR Clinical Genomics Plan DSTU2.1+ - Gil
    • Amnon's comments

• • other?

Attendees

• Amnon Shabo (Shvo), Bob Milius, Perry Mar, Jonathan Holt, Jeremy Warner, Heming Yao, Mollie Ullman-Cullere, David Kreda, Eric Whitebay, Beth Pumo, Bret Heale, M'Lynda Owens, Bob Freimuth, Joey

Draft Minutes

• 5 min Updates from
  • ClinGen - Bob F. nothing new since last update
  • GA4GH - Gil
    • FDA - interested is hearing about standards for NGS, how to communicate with vendors
      • John - question, is FDA involvement part of a workgroup? Gil - no, Gil is representing the GA4GH in presenting to FDA. It will be webcast.
    • Gil will give a talk to GA4GH on use cases based on the HL7 CG DAM in early Dec.
    • Hackathon is planned for documentation for code that GA4GH is developing.

• • IOM - Grant
    • eHealth workgroup. White paper is being developed on current status on family history. Not yet published or know when, but Grant will forward it when it is.
    • JD Nolan and Sandy Sorenson asked Grant to work on pilot workgroup. Forming a pilot groups - each with lab, health care provider, ehr vendor, lis vendor. Developing material for work group. They haven't met yet.
    • 3rd annual policy conference in Wash DC in early Dec. Planning is slower this time. Theme is precision medicine. Grant will moderate a panel.
  • FHIR Lead - Gil
    • PSS approved for new resources. Sequence resource will be in the same group as specimen.
    • David Kreda: It makes sense. Also, just using Google to search into FHIR or the find on web page in the HL7 pages, you can find these things ASAP. When you get to 100+ resources, search is the way!
  • other
    • CPIC - Bob F - nothing new from CPIC, still working on LOINC code.

• FHIR Clinical Genomics Plan DSTU2.1+ - Gil
  • long discussion regarding Amnon's comments (please add your own recollections here, or make corrections)
  • Amnon email comments about this: "Here are a couple of initial comments on the proposal to have Sequence as a FHIR Resource:
    • If we can introduce a base resource to the next FHIR DSTU, it should be the current Genetic Observation Profile (with possible refinements) and then the proposed Sequence resource can be a profile over the base Genetic Observation Resource. The rationale is that Sequence is a type of genetic observation, so this is the natural modeling hierarchy. In this way, we can later introduce more profiles like structural variations, expression data, etc. If Sequence is the base resource, how would you represent other types of genetic data?
    • I like the approach stated in "Derived analysis and clinical context should not be encapsulated in the Sequence resource" (also described in other sections), which is aligned with my comments on all previous FHIR ballots, where I asked for separation of indication, interpretation, assessed condition, etc. from the genetic observation itself. By the way, I prefer to use the term "encapsulated" in the sense of encapsulating raw data (as done in previous HL7 specs we developed), while in here I would use "embedded" or "bundled" instead.
    • Referencing GA4GH repository should be an option if this is a Universal spec, because implementers might choose to refer to other repositories. It is important though to have another attribute that specifies the type of repository referenced, and more importantly - the structure of the data in that repository, or in other words, how the raw data is represented in that referenced repository.
    • You also propose to profile the FHIR DiagnosticReport Resource, in what is called "Standard Genetics Profile". I would call it "Genetic Report" so that "profile" refers to FHIR Profiles, and not necessarily genetic profiles. In fact, such a Genetic Report is somewhat similar to the specs we published as the CDA-based Genetic Testing Report the v2 testing results message, however, the FHIR base resource is limiting, at least comparing to what we achieved in the GTR, but in any case, it's good if we align this new profile with previous efforts.
    • On page 6, what's the difference between the two example profiles?
    • From reading the detailed description of the proposed Sequence resource, it seems that it serves two different goals - a single variant or a sequence (without variant info). What's the rationale of having the two goals achieved in one structure? What about single variants that were observed not through sequencing?
    • I really like the fact that the 'post-coordination' approach is reiterated in the subsection "Relationships" under the Sequence resource description.
    • You used the phrase "objective data" - could you explain what that means?
    • In summary, I propose to rename the "Sequence" resource as "Genetic Observation", and the "Genetics Profile" as "Genetic Report" (also consistent with the base resource name). I would also consider having a Sequence profile over the base Genetic Observation resource, where extensions could describe unique characteristic of sequences (including issues like reference sequences, NGS, re-sequencing, etc)."
  • There were some questions about the PSS that was approved by the DESD in the Paris meeting last may, regarding the scope and flexibility with what we can do. The PSS will be emailed to the listserv.
  • Gil will be presenting a tutorial on the Sequence Resource in Orlando. It's unclear whether he will be respresenting SMART Genomics or the Clinical Genomics Workgroup. We need clarification on what this tutorial will include.
  • Jonathon: I've made some comments to the google docs sequence resource, that Gil sent out. My major comments is that there are a lot of attributes that belong at the level of the interpretation and not at the level of sequence data.
  • David Kreda: By being dedicated to sequencing information specifically, Sequence functions as a clean wrapper for output from Sequencing systems and services. The reference via Observation allows a connection to be made in the clinical system - in many cases, several observations that apply to the same static (in case of germline) data. This is an excellent way to layer delivery of data access.

• Motion:
  • The Orlando FHIR connectathon in Jan 2016 will include a FHIR resource and/or profile that will include the Sequence Resource being developed by Gil.
  • discussion: The question of whether we are ready for a FHIR connectathon was raised. Would it more prudent to delay it to the May meeting in Montreal? Should there be a "bake-off" between alternative versions/architectures of the resource? David argued that the connectathon was the place to vette the resource and even if we had design/architectural issues with it, that shouldn’t be used an argument to delay. He argued that others should be able bring their solutions to it as well. Amnon disagreed and argued against it using examples of previous efforts with the V3, V2, and CDA work of needing to get the semantics and modeling done right early.
  • abstain: 0
  • nay: Amnon
  • yea: Gil, Grant, Bret, David, Heming, Jeremy, Joey, Jon, M'Lynda, Mollie
  • The motion passed.
  • It was pointed out that the motion does not imply whether other versions/architectures will be included or evaluated then. The deadline of the supplying materials for the connectathon is not certain. Gil heard Nov 22 as a deadline during the Atlanta meeting. This needs to be verified, as well as what kinds of materials are needed.

October 27, 2015 -- Weekly call

Agenda

• None

Attendees

• Amnon Shabo (Shvo), Siew Lam, Jonathan Holt, Anwar, Bret Heale, Perry Mar, M'Lynda Owens, David Kreda, Bob Freimuth

Draft Minutes

• Amnon commented on the FHIR development that if we can introduce a base resource to next FHIR DSTU, it should be the current Genetic Observation Profile (with possible refinements) and then Sequence can be a profile over the base Genetic Observation Resource. The rationale is that Sequence is a type of genetic observation so that's the natural hierarchy. We can then introduce more profiles like structural variations, expression data, etc.
  • Discussion was held and David suggested to reiterate the issue when Gil can present the current plan
• Brett presented the genomic info button and asked to have joint discussions with the HL7 Clinical Decision Support Work Group
  • During December - a joint conf. call
  • During the Orlando WGM - a joint meeting
  • Motion accepted

October 20, 2015 -- Weekly call

Agenda

• 5 min Updates from
  • ClinGen
  • GA4GH
  • IOM
  • other?

• Atlanta WGM review
  • minutes

• "HL7 v2 genomics report lite"

• Orlando WGM - room reservation
• Oct 23 deadline

• FHIR review (Gil)

• Upcoming deadlines:
  • 2015-10-23: Deadline to complete the Post-WGM Evaluation (WG Health metric)
  • 2015-10-23: Deadline to complete WGM Room Request form for the January WGM (WG Health metric)
  • 2015-10-23: Deadline to post your October WGM minutes on the website or wiki (WG Health metric)
  • 2015-10-23: Deadline to notify HQ of additions/changes/corrections to Co-Chair openings
  • 2015-10-25: Notification of Intent to Ballot (NIB) Deadline
  • 2015-10-26: Initial Harmonization proposals due
  • 2015-10-28: Co-Chair Nominations Open

Attendees

• Amnon Shabo (Shvo), Siew Lam, Jon Holt, Gil Alterovitz, Grant Wood, Anwar, Bob Freimuth, Bret Heale, Clem Mcdonald, David Kreda, Eric Whitebay, JD Nolen, Kevin Power, Mollie (joined after 30 minutes)

Draft Minutes

• ClinGen
  • Bob F gave an update on structured variants (opy number, large indels, etc) in ClinGen
  • Jon asked if using ISCN syntax would help (Cytogenetics)
  • Bob F - good question, issue about fuzzy borders that need to be considered. Need more info about what ISCN can do.
  • question about whether anyone has looked at CYDAS - a system for parsing ISCN. This is an open source project that hasn't been updated in a long time.
• GA4GH
  • Grant gave an update on GA4GH, in particular on the effort related to family health history, and how FHH can be incorporated to the EHR. There is a short document (about 4-5 pages) that is being developed. It includes as section on HL7 standards on family history (Grant wrote this). The document is not a history of the domain, but it captures how family history is used today. It hasn't been published it yet, but expect it to be published in a couple weeks.
  • Gil - GA4GH is looking at different use cases. Gil will present to GA4GH use cases based on the DAM in December. Gil will share the date/time.
• IOM
  • Grant and JD gave an update on the EHR and genomics group
  • Grant - Sandy Aronson, and JD Nolan, asked Grant to co-chair another workgroup for moving to pilot phase.
  • JD - also working with PMI
• Gil gave an update on FHIR activities
• CPIC
  • Bob F. - CPIC informatics workgroup met yesterday, terminolgy/semantics work, proposed observation codes to use, they are slightly different from IOMs codes. Will work on harmonizing and send to Clem for LOINC. Analagous to v2 gentetic report LOINC codes. Harmonized with Delphi method of harmonizing phenotypic terms using experts digesting many rounds of surveys.
• Atlanta WGM review
  • Bob Milius gave an overview of the Atlanta WGM discussions
  • Amnon commented on the FHIR discussion that if we can introduce a base resource to FHIR it should be the Genetic Observation Profile (with possible refinements) and then Sequence becomes a profile over the base Genetic Observation Resource. The rationale is that Sequence is a type of genetic observation so that's the natural hierarchy.
• Gil gave an update on FHIR activities
  • FHIR is asking each Work Group to designate a single point of contact that will act as group lead for FHIR communications.
  • Mollie made a motion to have Gil designated as the CG rep. to FHIR ('group lead')
  • Jon asked about the exact role of FHIR rep.
  • Bob M. suggested we should have full transparency in the relations with FHIR, and that in any case the group can designate others and it's not an election process
  • Bob F. suggested that an update from the FHIR group lead be put in the agenda for every weekly call.
  • Amnon said we need to get it to the DMP
  • Motion passed (0 abstains, 0 nays)