W3C

- DRAFT -

Semantic Web Health Care and Life Sciences Interest Group Teleconference

10 Nov 2015

See also: IRC log

Attendees

Present
EricP, David_Booth, Tony_mallia, Chris_Stoeckert, Davide_Sottara, Lloyd_McKenzie
Regrets
Chair
David Booth
Scribe
dbooth

Contents


<trackbot> Date: 10 November 2015

ValueSet

http://wiki.hl7.org/images/f/ff/ValueSet_approach.pdf

tony: Representing coding.system as an owl individual directly is a convenience that I'm using.

lloyd: It should have a value property.

(tony showed inference though bridging ont and snomed ont)

tony's owl: http://wiki.hl7.org/index.php?title=RDF_for_Semantic_Interoperability#High-level_concept_mapping_to_RDF_.28AR_typeCodes.2C_etc..29

The OWL links: http://wiki.hl7.org/index.php?title=File:FHIROntology.txt

http://wiki.hl7.org/index.php?title=File:FHIR_SNOMEDCT_Bridge.txt

http://wiki.hl7.org/index.php?title=File:FragmentOWLSNOMEDCT.txt

<ericP> FHIR Ont (open from URL in Protege)

<rhausam> can you send the link again, Eric?

<hsolbrig> No idea what is going on

<hsolbrig> with respect to audio, but I'm no longer seeing the screen share

-> http://wiki.hl7.org/images/1/16/FHIROntology.txt FHIR Ont (open from URL in Protege)

<hsolbrig> I'll just listen... apologies

lloyd: Would be good to start splitting the ontology into what will be generated by HL7 versus what will be in the bridging ontology.

eric: I see this being used as having a FHIR RDF deliverable, then use cases that motivate the design, and that includes SNOMED and bridging ont, maybe DL query, to give us tools to test to know if we broke something.
... So that when we're working on the XSLT implementation, we can hit the button, see the query and see if we get results.
... Make progress having a bunch of tests.
... Don't know how to run Protege or SNAP SPARQL queries from the command line.
... Could probably do it with OWL API and a reasoner.

rob: I have scripted the reasoner before.

eric: Need to get this in github first.

tony discussing lines 342-372

<hsolbrig> I intend to do this (audio compromised)

dbooth: We're at the point where it would be helpful if at least a couple of other people got your ont and tried it themselves.

<scribe> ACTION: Harold and Rob to try tony's FHIR ontology and provide feedback [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action01]

<trackbot> Created ACTION-34 - And rob to try tony's fhir ontology and provide feedback [on Harold Solbrig - due 2015-11-17].

How ORIM testing was done

eric: talked with lloyd about how validation was done with the ORIM. Easy to apply it to this situation.

<dbooth> Option A: owl-validation: the owl reasoner barfs if instance data specifies a code does not match the valueset.

<dbooth> Option B: external valueset validator: Owl reasoner does not barf. Instead, an extra closed world inference step is use (perhaps using SWRL or other) to detect the problem.

<dbooth> Proposal: We should do both, but separate the base FHIR ontology (which would NOT cause the reasoner to barf if a wrong code is used) from an additional validation ontology (which WOULD cause the reasoner to barf if a wrong code is used).

dbooth: Is it viable to do both?

eric: Yes, you just need to move certain assertions out.

lloyd: Remove the triple that asserts the binding to the element. E.g., administrative gender, a valueSet defines what it can be. The statement to withhold is the assertion that one of the codings for gender must come from this valueSet. If you yank that triple then you will no longer have a contradiction if the code is not from that VS.

rob: In the OWA, I dont' think that is enough to achieve option A validation.

dbooth: Remember that they're strings, not named individuals.

lloyd: It works.
... Need to decide which things you do not want OWL to choke on, for example illegal combinations of code and system.

eric: Do we want to use things like cd:administrativeGender?

lloyd: use the fhir: namespace.

<ericP> lloyd: nope, fhir:Observation.administrativeGender

<ericP> ... the RIM used 21090 where possible. FHIR has its own.

<ericP> ... in V3 we had attribute -> concept domain -> value set

lloyd: but we've simplified that in FHIR.

<hsolbrig> Let me know as soon as the directory is ready to go.

dbooth: the option A add-on ont will be the valueset-validation ontology

ADJOURNED

<ericP> git clone git@github.com:w3c/hcls-fhir-rdf.git

<hsolbrig> The github page presents othert options

<hsolbrig> He can clone directly from the web page...

<hsolbrig> Note "Clond in Desktop"

<hsolbrig> A way to clone!

<hsolbrig> Oops - forgot that you had to have the desktop available

<hsolbrig> Maybe the "git clone" would be easier...

<hsolbrig> Gotta sign off.

<ericP> http://www.w3.org/2013/12/FDA-TA/

<ericP> orange just hung up my call

<ericP> just a sec...

<Tony> OK

<Tony> Don't hear you yet

<scribe> Meeting: HL7/W3C FHIR Task Force

FHIR RDF ontology

* Tony (~Tony@public.cloak) has joined #HCLS

https://github.com/w3c/hcls-fhir-rdf

https://github.com/w3c/hcls-fhir-rdf/tree/master/ontology

* davide_sottara (~davide_sottara@public.cloak) has joined #hcls

Added readme: https://github.com/w3c/hcls-fhir-rdf/tree/master/ontology

dbooth: Please try the ontology!

eric: We should also think about test cases.

davide: Still uneasy about property names.

tony: also, I've been looking at round tripping of the schema, and I don't think we have a requirement for round-tripping the schema.

chris: I'd like to hear how this is intended to fit in with other ontologies. That's why I'm on the call. :) E.g., OBO Foundry. Want to avoid FHIR creating yet another ont that will need to be mapped to others.

tony: Still need to look at the gaps between this and OBO.

chris: The relations ont has objects and properties. Should see how these correspond.

davide: Still trying to figure out: is this an ont of the FHIR resources, or the things that FHIR models, i.e., patients, diseases, etc.
... for example, a SubstanceAdministration and an Immunization are disjoint: something cannot be a member of both.

lloyd: correct.
... the concept that represents the Immunization record and the SubstanceAdministration record are disjoint, but they can reference the same ont when they talk about the med being adminstered.

davide: To bridge into an upper I don't think we can do it directly.

lloyd: Right, FHIR is fundamentally about records -- not knowledge.

chris: Maybe a bridging ont can go from FHIR ont being about records and SNOMED and OBO being about real world things.

dbooth: I think that's correct.
... What is the best way for us to support that bridge between FHIR's record-oriented ont and real world ontologies?

chris: Would be good to do it in a way that causes minimal work for users. Barry Smith just published a book on OBO that is one recommended reading.
... There's also an older paper on OBO foundry.

http://www.amazon.com/Building-Ontologies-Basic-Formal-Ontol

http://www.amazon.com/Building-Ontologies-Basic-Formal-Ontology/dp/0262527812/ref=sr_1_1?ie=UTF8&qid=1447195449&sr=8-1&keywords=barry+smith

http://www.nature.com/nbt/journal/v25/n11/full/nbt1346.html

dbooth: What's the best way forward on this?

eric: What are the use cases? i worked hard to bridge FDA therapeutic ont to other stuff. But I didn't justify my time because I did not find any really motivating use cases. e.g., if we have ont X, we could answer clinical question Y.

I have a biobanking use case. Find cohorts of specimens across institutions. Need to know about specimens but also consent associated with those specimens and the medical record of the donor

davide: This is a discussion we had at Mayo. Thousands of local systems with their local semantics that we want to normalize, and FHIR is a candidate. Syntactic normalization is impossible. Want broader scope of meaning that relates them to other ont.

Guoqian Jiang championed use cases: http://wiki.hl7.org/index.php?title=RDF_for_Semantic_Interoperability#Use_Cases

davide: Not hard to identify use cases, but more work to document them.

chris: If we have some stake in some use cases, we can take them back to our folks to say "we can do this"

davide: If we assume that we will have that bridge ont (or mediation process) then that takes pressure off of the FHIR ont to fill that need.

<scribe> ACTION: Chris Stoeckert to provide additional resources about how to get into BFO, OBO, etc. [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action02]

<trackbot> Created ACTION-37 - Stoeckert to provide additional resources about how to get into bfo, obo, etc. [on Chris Pierce - due 2015-11-17].

Created ACTION-35 - Stoeckert to provide additional resources about how to get into bfo, obo, etc. [on Chris Pierce - due 2015-11-17].

tony: record of medical action, versus why it was done.

Ontology for Biomedical Investigations: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2903726/

Next topics

<dbooth> eric: We should also think about test cases.

<dbooth> davide: Still uneasy about property names.

<dbooth> tony: also, I've been looking at round tripping of the schema, and I don't think we have a requirement for round-tripping the schema.

<dbooth> eric: What are the use cases?

Use Cases

<ChrisS> I have a biobanking use case. Find cohorts of specimens across institutions. Need to know about specimens but also consent associated with those specimens and the medical record of the donor

https://www.w3.org/wiki/HCLS/ClinicalObservationsInteroperability#Use_Cases

eric: Goal is to see how we would make use of other ontologies with the FHIR RDF ont.
... Use cases should be written up that way.
... Need to show how it would be use: the FHIR data, the query to execute, the bits of the ont that enable the query, etc.

<scribe> ACTION: Chris Stoeckert to show SPARQL queries used in biobanking and show our use case [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action03]

<trackbot> Created ACTION-38 - Stoeckert to show sparql queries used in biobanking and show our use case [on Chris Pierce - due 2015-11-17].

* trackbot is creating a new ACTION.

Created ACTION-36 - Stoeckert to show sparql queries used in biobanking and show our use case [on Chris Pierce - due 2015-11-17].

<scribe> ACTION: Davide to get uses cases from Mayo [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action04]

<trackbot> Error finding 'Davide'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.

* trackbot is creating a new ACTION.

Error finding 'Davide'. You can review and register nicknames at <http://www.w3.org/2014/HCLS/track/users>.

Round tripping of FHIR schema

* TallTed has quit ("")

lloyd: We'll put whatever we find useful into the OWL from the XML schema, but it does not need to be round trippable.
... We need to round trip the instances -- not the schema.

* ChrisS (~ChrisS@public.cloak) has left #hcls

* Tony has quit ("Page closed")

dbooth: agreed.

<scribe> Chair: David Booth

Summary of Action Items

[NEW] ACTION: Chris Stoeckert to provide additional resources about how to get into BFO, OBO, etc. [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action02]
[NEW] ACTION: Chris Stoeckert to show SPARQL queries used in biobanking and show our use case [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action03]
[NEW] ACTION: Davide to get uses cases from Mayo [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action04]
[NEW] ACTION: Harold and Rob to try tony's FHIR ontology and provide feedback [recorded in http://www.w3.org/2015/11/10-hcls-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/11/10 23:38:08 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)


WARNING: Possible internal error: join/leave lines remaining: 
        <dbooth> * Tony (~Tony@public.cloak) has joined #HCLS
        <dbooth> * davide_sottara (~davide_sottara@public.cloak) has joined #hcls


No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth

WARNING: Replacing previous Present list. (Old list: Tony_Mallia, Thomas_Lukasik, Rob_Hausam, Lloyd_McKenzie, Harold_Solbrig, EricP, David_Booth)
Use 'Present+ ... ' if you meant to add people without replacing the list,
such as: <dbooth> Present+ EricP, David_Booth, Tony_mallia, Chris_Stoeckert, Davide_Sottara, Lloyd_McKenzie

Present: EricP David_Booth Tony_mallia Chris_Stoeckert Davide_Sottara Lloyd_McKenzie
Found Date: 10 Nov 2015
Guessing minutes URL: http://www.w3.org/2015/11/10-hcls-minutes.html
People with action items: chris davide harold rob stoeckert

WARNING: Possible internal error: join/leave lines remaining: 
        <scribe> * Tony (~Tony@public.cloak) has joined #HCLS
        <scribe> * davide_sottara (~davide_sottara@public.cloak) has joined #hcls



WARNING: Possible internal error: join/leave lines remaining: 
        <scribe> * Tony (~Tony@public.cloak) has joined #HCLS
        <scribe> * davide_sottara (~davide_sottara@public.cloak) has joined #hcls



[End of scribe.perl diagnostic output]