Bio-ontologies in medicine: From bench to bedside and back again - - PowerPoint PPT Presentation

Bio-ontologies in medicine: From bench to bedside and back again

XXIII International Conference of the European Federation for Medical Informatics, Oslo, 28–31.08.2011 Peter N. Robinson Charit´ e Universit¨ atsmedizin Berlin

Desiderata for biomedical ontologies, terminologies, and classifications

Bioinformatics for molecular biology ⋆ Terms for annotating and searching experimental data ⋆ Integrative semantic analysis of data ⋆ Overrepresentation analysis Clinical Informatics ⋆ Structured reporting of clinical data ⋆ Enabling decision support ⋆ Billing systems

◮ Bioinformatics and medical informatics have traditionally been

separate disciplines

◮ Need for bridge between them in the coming era of genomic

and personalized medicine.

Utility of clinical data for research

◮ Disease pathobiology is associated with breakdown of one or

more cellular networks

◮ Comorbidity analysis of 13 million medicare records. ◮

from the Barab´ asi lab: Lee et al. (2008) The implications of human metabolic network topology for disease comorbidity PNAS 105:9880–9885

Detecting Drug Interactions from EHR data

◮ Mine FDA’s Adverse Event

Reporting System for side-effect profiles involving glucose homeostasis

◮ strong signal for

comedication with pravastatin and paroxetine

◮ Validation by mining of EHR

data

◮

From the Altman lab, Tatonetti et al. (2011) Detecting Drug Interactions From Adverse-Event Reports: Interaction Between Paroxetine and Pravastatin Increases Blood Glucose Levels Clin Pharmacol Ther 90:133-42.

Computational Analysis of Human Phenotypes

Costello-Syndrom Neurofibromatosis Type 1 Noonan-Syndrome LEOPARD-Syndrome CFC-Syndrome

Unique challenges in the field of genetics and rare disease

(Un)controlled vocabularies

generalized amyotrophy generalized muscle atrophy muscular atrophy, generalized muscle atrophy, generalized

The Human Phenotype Ontology

◮ Ontologies represent a powerful tool for annotating,

extracting, and analyzing clinical data.

• rgan

abnormality cardiac malformation

• abn. of the

cardiac septa

• abn. of the

cardiac atria

• abn. of the

atrial septum atrial septal defect cardiovascular abnormality cardiac abnormality

general terms specific terms

◮ ∼ 10.000 terms, ∼ 55.000 annotations for 4804 monogenic

diseases

◮

http://www.human-phenotype-ontology.org

◮

Robinson PN et al. (2008) The Human Phenotype Ontology: a tool for annotating and analyzing human hereditary disease. Am J Hum Genet. 83:610–5.

Similarity Measures for the Human Phenome

c)

Disorder class

Ophtamalogical Developmental Dermatological Endocrine Cancer Cardiovascular Connective Tissue Ear, Nose, Throat Gastrointestinal Hematological Immunological Metabolic Multiple Psychiatric Muscular Neurological Nutritional Renal Respiratory Skeletal Bone Skeletal/Bone/ Connective tissue/ Development Cancer Ophth Heme/ Immuno CV Metab Derma Endo/Renal Neuro Muscular

sim(d1, d2) = 1 2 · |d1|

• s∈d1

max

t∈d2 sim(s, t) +

1 2 · |d2|

• s∈d2

max

t∈d1 sim(s, t)

Ontological Diagnostics in Human Genetics

• abn. of

the eye

• abn. of the
• cular region
• abn. of the

eyelid

• abn. of globe

localization or size

• abn. of the

palpebral fissures hypertelorism downward slanting palpebral fissures

Noonan Syndrome

a)

Syndrome term Query term Overlap between query and disease

• abn. of

the eye

• abn. of the
• cular region
• abn. of the

eyelid

• abn. of globe

localization or size

• abn. of the

palpebral fissures downward slanting palpebral fissures

Opitz Syndrome

b)

telecanthus hypertelorism

c)

Noonan Syndrome

downward slanting palpebral fissures hypertelorism

3.78 3.05 Opitz Syndrome

hypertelorism telecanthus

3.05 2.45

(IC of abn. of the eyelid) downward slanting palpebral fissures hypertelorism

Query (Q) sim(Q,Opitz) =2.45 + 3.05 2 = 2.75 sim(Q,Noonan) = 3.78 + 3.05 2 = 3.42

sim(Q → d) = 1 |Q|

• s∈Q

max

t∈d sim(s, t)

Q: Query terms d: Disease terms

The Phenomizer

◮ Search for diagnosis according to phenotypic features

The Phenomizer

◮

Now there is only one diagnosis with a significant P-value

◮

Sebastian K¨

• hler et al. (2009) Clinical Diagnostics with Semantic Similarity Searches

in Ontologies. Am J Hum Genet, 85:457–64. http://compbio.charite.de/Phenomizer

HPO and PATO: A Semantic Web of the Human Phenotype

density Pathological Bone

[Term] id: HP:0002796 ! Osteosclerosis intersection_of: PATO:0001788 ! increased density intersection_of: has_quality PATO:0001869 ! pathological intersection_of: inheres_in FMA:30317 ! Bone

Logical Definition of HPO Term Osteosclerosis ◮ Method for defining semantics of medical terms by linking

HPO to other bio-ontologies: GO, FMA, MPATH, Cell Ontology, ChEBI, PRO, etc.

◮

Gkoutos GV, Mungall C, D¨

• lken S, Ashburner M, Lewis S, Hancock J, Schofield P, K¨
• hler S, and Robinson

PN (2009) Entity/Quality-Based Logical Definitions for the Human Skeletal Phenome using PATO. Proceedings of the 31st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2009)

PATO

[Term] id: HP:0001943 ! Hypoglycemia intersection_of: qualifier PATO:0000460 ! abnormal intersection_of: PATO:0001163 ! decreased concentration intersection_of: towards CHEBI:17234 ! glucose intersection_of: inheres_in FMA:9670 ! Blood

◮ Link HPO to other ontologies ⇒ Data integration to facilitate

analyzing, mining, and querying biological knowledge Class: Hypoglycemia EquivalentTo: ’decreased concentration’ and inheres_in some ’Blood’ and towards some ’glucose’ and qualifier some ’abnormal’

Computational Reasoning over the Human Phenotype

abnormal ion homeostasis intersection_of: PATO:0000001 ! quality intersection_of: qualifier PATO:0000460 ! abnormal intersection_of: inheres_in GO:0050801 ! ion homeostasis abnormal copper homeostasis intersection_of: PATO:0000001 ! quality intersection_of: qualifier PATO:0000460 ! abnormal intersection_of: inheres_in GO:0006878 ! copper ion homeostasis GO:0006878 copper ion homeostasis GO:0050801 ion homeostasis is_a

G e n e O n t

• l
• g

y

identical

is_a

◮ We can now exploit knowledge in ontologies such as GO,

FMA, MPATH, etc. to make sure our representation of knowledge about the human phenotype is accurate

◮ The definitions are also a basis for integrative computational

analysis

◮

K¨

• hler et al., Improving ontologies by automatically reasoning and evaluating logical definitions, 2011

Mouse Disease Models

◮ Semantic bridge between

human and model organism phenotypes:

◮ The International Knockout

Mouse Consortium (IKMC) ⇒ mutate all protein-coding genes in the mouse.

◮ Similar efforts for zebrafish ◮ How to harness this

information for human health?

Collaboration with Paul Schofield & George Gkoutos (U Cambridge), Damian Smedley (EBI), Cynthia Smith (JAX), Chris Mungall & Michael Ashburner & Suzi Lewis (GO), Monte Westefield & Barbara Ruef (ZFIN)

Interspecies comparisons

Phenotype ontologies

◮ HPO (Human Phenotype Ontology) ◮ MPO (Mammalian Phenotype Ontology) ◮ ZFIN (Zebrafish E/Q definitions) ◮ Semantic bridge between 3 species

Building a bridge between phenotype and genes

CNVs and Other Genomic Disorders

◮ Genomic Disorders: Deletions, Duplications, Rearrangements

affecting multiple genes

◮ Phenotype results from dosage imbalance or regulatory effects

• f one or more affected genes

◮ Diagnostic problem: Distinguish pathogenic from neutral

CNVs

◮ Scientific and medical problem: Decide which genes are

responsible for the phenotype

Williams-Beuren Syndrome

Semantic analysis of mouse phenotypes to find candidate genes

Williams-Beuren Syndrome

Numerous previously unknown “explanations” for the Williams Phenotype

Analysis of CNV Syndromes

1779 Human genes 6170 Mouse genes 1598 Zebrafish genes

53 426 4262 522 289 608 1011

◮ Analysis of 24 CNV disorders ◮ Phenotypic mapping

between disorders caused by mutations in orthologous genes

◮ 518 candidate genes for

individual phenotypic features,

◮ 348 not previously reported

in the literature.

◮ A basis for understanding

genotype-phenotype correlations in pathogenic CNVs

Doelken et al., manuscript submitted

Next Steps

◮ The HPO is still young (11/2008) but has

been adopted by many databases

◮ Wellcome Trust Sanger Institute: DECIPHER

and DDD databases

◮ NCBI: dbGAP, dbVAR (Whole exome

sequencing data)

◮ International Standards for Cytogenomic

Arrays (ISCA) Consortium

◮ GWAS Central ◮ Several research-oriented bioinformatics

databases

◮ Collaboration with OMIM & Orphanet on

rare disease classification ontology and links between the two projects

Thank you for your attention ....

Institut f¨ ur Medizinische Genetik und Human Genetik Charit´ e Universit¨ atsmedizin Berlin Sebastian Bauer Sandra D¨

• lken

Bego˜ na Garc´ ıa Mu˜ noz Johannes Gr¨ unhagen Gao Guo Peter Hansen Sebastian K¨

• hler

Syzmon Kie lbasa Christian R¨

• delsperger (alumnus)

Marten J¨ ager Peter Krawitz Claus-Eric Ott Angelika Pletschacher Peter N. Robinson

http://compbio.charite.de

K11656_Cover.indd 1 3/28/11 1:46 PM