Samwald cshals2013
Upcoming SlideShare
Loading in...5
×
 

Samwald cshals2013

on

  • 258 Views

 

Statistics

Views

Total Views
258
Views on SlideShare
258
Embed Views
0

Actions

Likes
0
Downloads
0
Comments
0

0 Einbettungen 0

No embeds

Zugänglichkeit

Kategorien

Details hochladen

Uploaded via as Adobe PDF

Benutzerrechte

© Alle Rechte vorbehalten

Report content

Als unangemessen gemeldet Als unangemessen melden
Als unangemessen melden

Wählen Sie Ihren Grund, warum Sie diese Präsentation als unangemessen melden.

Löschen
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Ihre Nachricht erscheint hier
    Processing...
Kommentar posten
Kommentar bearbeiten

Samwald cshals2013 Samwald cshals2013 Presentation Transcript

  • Semantically EnablingGenetic MedicineFacilitating Patient - Guideline Matchingand Pharmacogenetic Clinical DecisionSupportMatthias SamwaldMedical University of Vienna & Vienna University of TechnologyW3C Semantic Web for Healthcare and Life Science Interest Group
  • Drug efficacy and toxicity can vary drastically betweenpatients with different genetic profiles
  • The drug warfarin is a popular example of clinicalpharmacogenomics
  • Such pharmacogenomic information isbecoming available for many other drugs
  • OWL 2 for pharmacogenomic knowledgerepresentation and clinical decision support
  • Allele definitions in the Human CytochromeP450 Allele Nomenclature Database
  • Allele definitions in PharmGKB
  • 1 Class: human with CYP2C9 *32 EquivalentTo:3 has some rs1057910_C4 SubClassOf:5 has some CYP2C9 *3,6 (has some rs1057910_C) and7 (has some rs1057911_A) and8 (has some rs1799853_C) and9 (has some rs2256871_A) and10 (has some rs28371685_C) and11 (has some rs72558188_AGAAATGGAA) and12 (has some rs72558189_G) and13 (has some rs9332239_C)...Allele definitions in OWL 2
  • 1 Class: human with CYP2C9 *182 EquivalentTo:3 (has some rs1057910_C) and4 (has some rs1057911_T) and5 (has some rs72558193_C)6 SubClassOf:7 has some CYP2C9 *18,8 (has some rs1057910_C) and9 (has some rs1057911_T) and10 (has some rs1799853_C) and11 (has some rs2256871_A) and12 (has some rs28371685_C) and13 (has some rs28371686_C) and14 (has some rs56165452_T) and15 (has some rs57505750_C) and16 (has some rs67807361_C) and...Allele definitions in OWL 2
  • Dosing guideline from an FDA drug label
  • Dosing guideline from an FDA drug label1 Class: human triggering CDS rule 92 Annotations:3 CDS_message "0.5-2 mg warfarin per day should be considered4 as a starting dose range for a patient with this genotype5 According to the warfarin drug label."6 EquivalentTo:7 (has some CYP2C9 *1) and8 (has some CYP2C9 *3) and9 (has exactly 2 rs9923231_T)(of course, entities in the ontology are mapped toother Semantic Web resources such as Bio2RDF)
  • Definitions can become quite complex –OWL reasoning helps identify inconsistencies and lacking definitions
  • Describing an individual patient in OWL1 Individual: example_patient2 Types:3 human,4 (has some rs1208_A) and (has some rs1208_G),5 (has some rs8192709_C) and (has some rs8192709_T),6 (has some rs9934438_A) and (has some rs9934438_G),7 has exactly 2 rs10264272_C,8 has exactly 2 rs9923231_T,9 has exactly 2 rs12720461_C,10 (has some ‘CYP2C9 *1’) and (has some ‘CYP2C9 *3’),11 has exactly 2 ‘CYP2C19 *1’,12 (has exactly 3 CYP2D6) and (has exactly 2 ‘CYP2D6 *1’)13 and (has exactly 1 ‘CYP2D6 *2’)...heterozygousSNP variantshomozygousSNP variantsallelic variantsand CNV
  • Describing an individual patient in OWL1 Individual: example_patient2 Types:3 human,4 (has some rs1208_A) and (has some rs1208_G),5 (has some rs8192709_C) and (has some rs8192709_T),6 (has some rs9934438_A) and (has some rs9934438_G),7 has exactly 2 rs10264272_C,8 has exactly 2 rs9923231_T,9 has exactly 2 rs12720461_C,10 (has some ‘CYP2C9 *1’) and (has some ‘CYP2C9 *3’),11 has exactly 2 ‘CYP2C19 *1’,12 (has exactly 3 CYP2D6) and (has exactly 2 CYP2D6_star_1)13 and (has exactly 1 CYP2D6_star_2)...heterozygousSNP variantshomozygousSNP variantsallelic variants"0.5 - 2 mg warfarin per dayshould be considered as astarting dose range for a patientwith this genotype according tothe warfarin drug label."OWL Reasoner
  • OWL 2 DL reasoning needs to be fast!!
  • TrOWL is massively more performant than the HermiT reasoner inclassifying our demo ontologyHermiT TrOWLGenomic CDS light(2150 classes, 9500 axioms)3 hours 48 minutes 18 secondsGenomic CDS(2300 classes, 11000 axioms)did not terminate within6 hours54 secondsOntologies have ALCQ expressivity.http://www.genomic-cds.org/http://trowl.eu/ (Pan et al.)http://hermit-reasoner.com/ (Horrocks, Motik et al.)
  • How can we put this into the hands of doctors andpatients?A prototype based on 2D barcodes
  • Making pharmacogenomics easier to understand for doctors and patientsMaking core set of pharmacogenomic data available for every patientMaking pharmacogenomic decision support available to every doctorProviding pharmaceutical manufacturers with a basic set of markers forpatient stratification that can be made cheaply available for large patientpopulations through prospective genotypingCurated set of 380+ markers, 50+ pharmacogenes and rule system:The Medicine Safety Code prototype
  • (data on 385 pharmacogenetic markers are encoded in there)A Medicine Safety Code can be represented as a QR code
  • Modern mobile devices can successfully decode very compactMedicine Safety Codes
  • The MSC provides a simple, barrier-free system for storing andinterpreting personal pharmacogenomic information(of course, the data from an MSC could also be embedded in an EHR where available)
  • Clinical decision support messages can beviewed quicklyWithin 15 seconds in best-case scenariosSee recent JAMIA paper for more informationUpcoming version will be backed by the ontology andreasoning system I showed you before
  • ConclusionsStill at an early stage(Semantic) Web technologies used for entire contiuum from basic knowledgerepresentation up to deployment in clinical routineOWL 2 ontologies and reasoning could help to make pharmacogenomicknowledge more transparent and error-freeOWL 2 could be established as a shared standard for representing(pharmaco-)genetic knowledge??
  • Future workPartnership with clinics, pharmaceutical companies, genetic testingproviders, payersCreate optimized reasoner for the specific OWL 2 fragment needed forgenomics? (ALCQ with arbitrary cardinalities)
  • ThanksW3C collaborators:Michel Dumontier (Carleton University)Robert R. Freimuth (Mayo Clinic)Richard Boyce (University of Pittsburgh)Robert L. Powers (Predictive Medicine, Inc.)Joanne S. Luciano (Rensselaer Polytechnic Institute)Eric Prud’hommeaux (W3C)M. Scott Marshall (MAASTRO Clinic)Simon Lin (Marhsfield Clinic)and others