HCLSIG/PharmaOntology/Meetings/2010-05-06 Conference Call

From W3C Wiki

Conference Details 

* Date of Call: Thursday May 6 2010 
* Time of Call: 12:00pm - 1:00pm ET 
* Dial-In #: +1.617.761.6200 (Cambridge, MA) 
* Dial-In #: +33.4.89.06.34.99 (Nice, France) 
* Dial-In #: +44.117.370.6152 (Bristol, UK) 
* Participant Access Code: 42572 ("HCLS2") 
* IRC Channel: irc.w3.org port 6665 channel #HCLS2 (see W3C IRC page for details, or see Web IRC) 
* Mibbit quick start: Click on mibbit for instant IRC access
* LiveMeeting URL: Presenter URL:  https://www.livemeeting.com/cc/jnjamericas/join?id=WF4H98&role=present&pw=P%3F%5B%3F_5q
* Duration: 1h 
* Convener: Susie


Agenda

  • Introduction to the Drug Ontology - Samson
  • TMO Ontology - Elgar, Michel, Trish
  • Demo Data - Michel, EricP
  • IO Informatics Data - Elgar
  • Artificial Claims Data - Janos
  • Interface/eMerge - Bosse, Chris
  • Outreach (IHI, Bio-IT World Europe, UPenn Translational Medicine) - Michel, Susie
  • AOB


Minutes

Attendees: Bob, Samson, EricP, Julia, Paul, Janos, Bosse, Chris, Dave, Michel, Joanne, Scott, Trish, Susie

Presentation:

Samson Tu on the Stanford Center for Biomedical Informatics Research Drug Ontology (supported by NIH at the Butte Lab)

Goals: · Synthesize info from multiple sources (core concepts + classification services + external links) · Easy to construct, minimal manual curation and version mgmt, provenance included Scope: · Drug/disease relationships · Drug properties · Drug/drug relationships

NOT for clinical decision support

Method: start with core, use a description logic reasoned to create drug hierarchies along different axes

Use Protégé Jython Tab to import data (UMLS, PharmGKB, Commercial Sources, Drug Bank) into a database

Types of info: adverse effects, mech. of action, indication information etc. from different sources

Concepts: ClinicalFinding, DrugTarget, IngredientDrug, MechanismOfAction

Query with OWL

Issues: users don’t want to use OWL, need to mash with other data sources, need to make information available externally

Solution: export as a mysql database, so can be queried in SQL (Issue: not as expressive as OWL)

Inferred relationships can be saved back to the db

Contributions: · integrated ontologies as a resource in a scalable automated manner, · can identify gaps and errors in sources

Status: · on hold due to resources · paper in the works · Prof. Butte to make decision on making Drug Ontology available publicly

Questions

Scott: Use of mysql: why not SparQL? Samson: haven’t tracked the usage yet, the internal group may be happy to use SQL

Scott: let’s think about how we can help make this public.

Susie: how well the TMO maps to the Drug Ontology. We can help each other promote our work

Paul: drug market info changes. How are changes managed? Samson: Drug ontology is meant to be frequently re-built (rather than incrementally changed over time) with automated scripts.

Susie: what other projects are you working on? Samson: modeling of clinical trials, mgmt of patient records, modeling of breast cancer protocols and related ontologies, rule-based systems..

Scott: is the ontology currently in use? Samson: not sure: Butte lab has access, but usage is not being tracked


Susie: next item is TMO updates.

Michel: not much done last week, people still need to look and report what needs work.

Paul: will send email to Michel with suggestions.

Janos: is there a class for encounters?

Susie: Michel sent links that may be relevant. We modeled patient data on a source that was not finished (Indivo). Now it’s open source and we can see what we are missing in comparison. Their code base is used by other companies (Microsoft and pharma consortiums). We need to see if we should move in that direction.

Susie: another source is blog by Chime Ogbuji (Cleveland Clinic) using SW technologies to help with clinical decision support. He will present his work for us on his Compubter-based Patient Record (CPR) ontology. We need to consider whether to use it.

Scott: there is also Ontology for General Medical Science (Rottenburg). All of those overlap and have more classes that we could use.

Susie: let’s get them all in and decide.

Susie: should we look at Samson’s paper on Ontology of Clincal Research?

Action: Trish will speak to Samson on Ontology of Clinical Research

Trish: it’s publicly available and new versions get uploaded into BioPortal, and it’s being used and is futher along.

Trish: will revisit possible connection

Susie: OMOP is an effort to take EHR data and mining it and combining it with payer data. This has a data model that Colin looked at it and didn’t like it, but OMOP is reaching out, so we should look at it more. Susie: will find a date for the Cleveland Clinic presentation.

Susie: any interest in OMOP?

Action: Joanne will look at the data model. Susie will look for more info.

Susie: some pharmas are working with OMOP to put data into a cloud and mine it.

Trish: how to structure comments on BioPortal to feed back into Protégé, bases on ‘Notes’ feature? What are needs?

Question: what is Notes feature?

Action: Trish to give overview on Notes feature.

Susie: Elgar hasn’t heard back from IO Informatics – if no response, will move on.

Scott: let’s look at some other data sources

Action: Scott to give overview of data in next week meeting (data from Duke, some from a Carolina warehouse)

Susie: interface

Chris: Luke at Marshfield Clinic in WI has EMR records for many years. Discussed their formats and interfaces. Connor, Eric and Chris will work helping Luke with ontologies, linking querying. Will get him to a SparQL endpoint, without looking at data. Part 2: ask simple questions across data and develop an interface for a clinician to put info into the system and get answers from the system. So, Part 1 is RDF technology, Part 2 is querying and UI.

Susie: the minutes are in the archive.

Susie: Eric – what about updating patient record information? Should we look at Indivo? And updating our links to RxNorm. Can you help us?

Eric: can go to Harvard and talk about Indivo. We do not want to maintain XML, but rather deal only with RDF.

Susie: if we go into Indivo schema, we need to get data into that format. Need to make sure all the source links are there.

Action: Eric to look into it and give us an update next week (maybe)

Eric: need to talk to Ben and Michel.

Action: Connor should also keep Indivo schema in mind for the Marshfield Clinic project

Susie: Outreach

Susie: will find out BioOntologies paper status (this Friday?)

Susie: IHI paper is next

Joanne: haven’t heard from BioIT Europe

Action: Joanne to follow up with Edel

Scott: I'd like to be a backup presenter

Susie: AOB?

Janos: won’t have artificial claims data before June.

Retrieved from "https://www.w3.org/wiki/index.php?title=HCLSIG/PharmaOntology/Meetings/2010-05-06_Conference_Call&oldid=44464"