1. Semantically Enabling
Genetic Medicine
Facilitating Patient - Guideline Matching
and Pharmacogenetic Clinical Decision
Support
Matthias Samwald
Medical University of Vienna & Vienna University of Technology
W3C Semantic Web for Healthcare and Life Science Interest Group

2. Drug efficacy and toxicity can vary drastically between
patients with different genetic profiles

3. The drug warfarin is a popular example of clinical
pharmacogenomics

4. Such pharmacogenomic information is
becoming available for many other drugs

5. OWL 2 for pharmacogenomic knowledge
representation and clinical decision support

7. Allele definitions in the Human Cytochrome
P450 Allele Nomenclature Database

8. Allele definitions in PharmGKB

9. 1 Class: 'human with CYP2C9 *3'
2 EquivalentTo:
3 has some rs1057910_C
4 SubClassOf:
5 has some 'CYP2C9 *3',
6 (has some rs1057910_C) and
7 (has some rs1057911_A) and
8 (has some rs1799853_C) and
9 (has some rs2256871_A) and
10 (has some rs28371685_C) and
11 (has some rs72558188_AGAAATGGAA) and
12 (has some rs72558189_G) and
13 (has some rs9332239_C)
...
Allele definitions in OWL 2

10. 1 Class: 'human with CYP2C9 *18'
2 EquivalentTo:
3 (has some rs1057910_C) and
4 (has some rs1057911_T) and
5 (has some rs72558193_C)
6 SubClassOf:
7 has some 'CYP2C9 *18',
8 (has some rs1057910_C) and
9 (has some rs1057911_T) and
10 (has some rs1799853_C) and
11 (has some rs2256871_A) and
12 (has some rs28371685_C) and
13 (has some rs28371686_C) and
14 (has some rs56165452_T) and
15 (has some rs57505750_C) and
16 (has some rs67807361_C) and
...
Allele definitions in OWL 2

11. Dosing guideline from an FDA drug label

12. Dosing guideline from an FDA drug label
1 Class: 'human triggering CDS rule 9'
2 Annotations:
3 CDS_message "0.5-2 mg warfarin per day should be considered
4 as a starting dose range for a patient with this genotype
5 According to the warfarin drug label."
6 EquivalentTo:
7 (has some 'CYP2C9 *1') and
8 (has some 'CYP2C9 *3') and
9 (has exactly 2 rs9923231_T)
(of course, entities in the ontology are mapped to
other Semantic Web resources such as Bio2RDF)

13. Definitions can become quite complex –
OWL reasoning helps identify inconsistencies and lacking definitions

14. Describing an individual patient in OWL
1 Individual: example_patient
2 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’)
...
heterozygous
SNP variants
homozygous
SNP variants
allelic variants
and CNV

15. Describing an individual patient in OWL
1 Individual: example_patient
2 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)
...
heterozygous
SNP variants
homozygous
SNP variants
allelic variants
"0.5 - 2 mg warfarin per day
should be considered as a
starting dose range for a patient
with this genotype according to
the warfarin drug label."
OWL Reasoner

16. OWL 2 DL reasoning needs to be fast!!

17. TrOWL is massively more performant than the HermiT reasoner in
classifying our demo ontology
HermiT TrOWL
Genomic CDS light
(2150 classes, 9500 axioms)
3 hours 48 minutes 18 seconds
Genomic CDS
(2300 classes, 11000 axioms)
did not terminate within
6 hours
54 seconds
Ontologies have ALCQ expressivity.
http://www.genomic-cds.org/
http://trowl.eu/ (Pan et al.)
http://hermit-reasoner.com/ (Horrocks, Motik et al.)

18. How can we put this into the hands of doctors and
patients?
A prototype based on 2D barcodes

19. Making pharmacogenomics easier to understand for doctors and patients
Making core set of pharmacogenomic data available for every patient
Making pharmacogenomic decision support available to every doctor
Providing pharmaceutical manufacturers with a basic set of markers for
patient stratification that can be made cheaply available for large patient
populations through prospective genotyping
Curated set of 380+ markers, 50+ pharmacogenes and rule system:
The Medicine Safety Code prototype

20. (data on 385 pharmacogenetic markers are encoded in there)
A Medicine Safety Code can be represented as a QR code

21. Modern mobile devices can successfully decode very compact
Medicine Safety Codes

22. The MSC provides a simple, barrier-free system for storing and
interpreting personal pharmacogenomic information
(of course, the data from an MSC could also be embedded in an EHR where available)

23. Clinical decision support messages can be
viewed quickly
Within 15 seconds in best-case scenarios
See recent JAMIA paper for more information
Upcoming version will be backed by the ontology and
reasoning system I showed you before

24. Conclusions
Still at an early stage
(Semantic) Web technologies used for entire contiuum from basic knowledge
representation up to deployment in clinical routine
OWL 2 ontologies and reasoning could help to make pharmacogenomic
knowledge more transparent and error-free
OWL 2 could be established as a shared standard for representing
(pharmaco-)genetic knowledge??

25. Future work
Partnership with clinics, pharmaceutical companies, genetic testing
providers, payers
Create optimized reasoner for the specific OWL 2 fragment needed for
genomics? (ALCQ with arbitrary cardinalities)

26. Thanks
W3C 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