Research Infrastructures to boost R&D

in the field of rare Diseases

1

Ségolène AyméINSERM, Paris, France

Fundacion Ramon Areces29 Oct 2014

International Rare Disease Research Consortium (IRDiRC)

Cooperation at international level to stimulate, better coordinate and

maximize output of rare disease research efforts around the world

2

VOICE OFDATA

(EVIDENCE)

DIAGNOSIS• Technology

• devices, instruments, bioinformatics, systems

RARE DISEASE SECTOR

Clinical & AcademicIndustry & Manufactures Multiple Government DepartmentsPrivate Healthcare

THE CHALLENGE

Metabolomics

Natural History

Clinical expertise/experts

Genomics

Phenomics

Multiple Government Departments

Public healthcare system

Public healthcare system

Public healthcare system

Public healthcare and research system

Industry & Manufactures

Metabolomics

Clinical expertise/experts

Phenomics

Genomics

Private Healthcare

Interpretation and application

Interpretation and application

Clinical expertise/experts

Transcriptomics

Proteomics

Training

Training

Education Education

Education

Training

Proteomics

Proteomics

Multiple Government Departments

Policy

Policy

Clinical and disability services

Position statements

Rare Diseases PeculiaritiesDISADVANTAGE no or little evidence available small populations , scattered coding and classification poor no jurisdiction , or country with

sufficient data require collective data and case

finding for evidence not all rare diseases are the same in

terms of evidence: e.g. Cystic Fibrosis ≠ Progeria

orphan therapies fail the cost effectiveness threshold

ADVANTAGE Clarity in the extreme Phenotype:

– genotype atomise disease; – permit re-aggregation based on

pathways perturbed, not clinical presentation

New knowledge translation and the portal to Individualised medicine

4

0

20

40

60

80

100

120

140

160

180

200

0 5 10 15 20 25 30 35 40 45 50

Num

ber o

f dis

ease

s

Estimated prevalence (per 100 000)

Prevalence distribution of rare diseases

Motor Neurone DiseaseRetinoblastomaAngelman SyndromeNiemann-Pick diseaseNemaline myopathyMucopolysaccharidosis 1-3

Facioscapulohumeral dystrophyRett syndromeCongenital myopathy

70% of people living with a rare disease

75% of people living with a rare disease

Friedreich ataxiaAlport syndrome

NoonanIsolated Spina Bifida Cutaneous lupus erythematosus

Hereditary breast & ovarian cancer syndromeSystemic sclerosis

Neurofibromatosis type 1Charcot-Marie-Tooth disease

Diffuse large B-cell lymphoma Fragile X syndromeMarfan syndromeMyasthenia gravisTuberculosisTurner Syndrome

Familial long QT syndromeFetal cytomegalovirus syndromePartial chromosome Y deletion

Young adult-onset ParkinsonismSickle cell anemia

Williams syndromeCystic fibrosis

Duchenne muscular dystrophyHereditary spastic paraplegia

Malaria Mesothelioma

PhenylketonuriaFamilial adenomatous polyposis

Huntington diseaseHemophilia A

80% of People living with a rare disease

CURRENT STATUS OF RESEARCH IN THE FIELD OF RARE DISEASES BASED ON ORPHANET DATA

6

5707 ongoing research projects in Orphanet covering 2129 diseases, excluding clinical trials

European rare diseases research landscape (36 countries)

(February 2014)

513 Gene search 595 Mutations search 281 Gene expression profile 393 Genotype-phenotype correlation1048 In vitro functional study 509 Animal model creation / study 748 Human physiopathology study 179 Pre-clinical gene therapy 90 Pre-clinical cell therapy

31 Pre-clinical vaccine development452 Observational clinical study224 Epidemiological study

295 Diagnostic tool / protocol development

158 Biomarker development

25 Medical device / instrumentation development

79 Health sociology study 15 Health economics study

72 Public health / health services study

7

Percentage of clinical trials by category

International rare diseases clinical trial landscape

2476 ongoing national or international clinical trials for 629 diseases in 29 countries

(April 2014) 8

1%

78%

1%

2%

16%

1% 1%cell therapy clinical trialdrug clinical trialgene therapy clinical trialmedical device clin-ical trialprotocol clinical trialvaccine clinical trialother trial

Number of genes tested in each country in Europe by year

2010 2011

20122013

Possibility to diagnose Rare Diseases:over 2 362 genes tested to date

Number of genes tested by country Number of rare diseases tested by country

(April 2014)10

Medicinal products on the European market in 2013

68 orphan medicinal products 92 medicinal products without orphan designation with at least an indication for a rare disease or a group of rare disease

(January 2014)

Satisfaction for professionalsFrustration for patients

Anxiety for payors Slow translation from bench to bedside

Limited access to innovations Too few treatments compared to needs

Most patients feeling abandonned High cost of diagnostic tests and drugs

Not affordable Necessity to de-risk research

Cheaper R&D

12

How to speed-up research and de-risk it ?

Improve coordination and synergies of research at world levelTo increase the research volume and the quantity of data

Support in-silico research to make optimal use of available data

Find new business-model for R&DTo reduce the cost and et profide affordable treatments

13

TO BOOST COORDINATION AT WORLD LEVEL

14

IRDiRC policy and guidelinesPrinciples applying to Research activities

15

Sharing and collaborative work in RD research Sharing of data and resources Rapid release of data Interoperability and harmonization of data Data in open access databases

Scientific standards, requirements and regulations in RD research Projects should adhere to IRDiRC standards Develop ontologies, biomarkers and patient-centered outcome data Cite use of databases and biobanks in publications

IRDiRC policy and guidelines

16

Participation by patients and / or their representatives in research Act in the best interest of patients Involve patients in all aspects of research Involve patients in design and governance of registries Involve patients in the design, conduct and analysis of clinical trials Acknowledge patients contribution in articles

IRDiRC policy and guidelinesPrinciples applying to Funding Bodies

17

Promote the discovery of genes Promote the development of therapies Fund pre-clinical studies for proof of concept Promote harmonization, interoperability, sharing, open access

data Promote coordination between human and animal models Promote active exchanges between stakeholders through

information dissemination of ongoing projects and events

IRDiRC policy and guidelines Endorsement of standards and tools

Endorsement of standards and tools contributing to IRDiRC objectivesOntologies: HPO, ORDO…Standards: BRIF…Data sharing: PhenomeCentral, DECIPHER…Ouctome measures: NINDS, PROMIS…

18

IRDiRC Recommended Label to be used in highlighting tools, standards and guidelines,

which contributes directly to IRDiRC objectives Application for ‘IRDiRC Recommended’ label is open to all, including

non-IRDiRC members ‘IRDiRC Recommended’ may be awarded to similar tools, standards

and guidelines Submission of 1-2 pages application Evaluation of the application by a review panel Approval/rejection of the application by the Executive Committee

19

INITIATIVES TO SPEED UP DATA SHARING

20

Rational

Research produces an enormous amount of data If shared, will facilitate the development of diagnostics and

treatments while ensuring efficient utilization of scarce resources

Resources include patient and family material (extracted DNA, cell lines, pathological samples), technical protocols, informatics infrastructure, and analysis tools

Datasets include phenotypes, genomic variants, other ‘omic’ data, natural histories, and clinical trial data…

21

Barriers to Data Sharing Technical and Financial issues

Storing terabytes…Securing dataProviding the logistics for sharing dataStatistical and algorithmic issues to combine datasets

Ethical and Legal issuesData across public and private networksPricacy protection at national level

Cultural issuesReluctance to share data from researchers/

Institutions/Regulatory bodies

22

A ClearingHouse of Data Standardsis in development at IRDiRC

Five main fields of application Standards in Genomics and other OMICS Standards in Phenotyping Standards in Outcome Measures for clinical trials Standards in Human Data Registration Open-access Data Repositories to store data

Alignment with other efforts to ensure interconnection and shareability between data RD-Connect PCORI, Comete ELIXIR, BD2K, Data FAIRport

23

PhenoTips and PhenomeCentral

Repository of data Hub for data sharing CareforRare,

RDConnect NIH undiagnosed

patients

24

Open Acess Data Repositories

ClinVar and ICCG Public archive of

variants and assertions about significance

NCBI resource

Decipher Database of Chromosome imbalances and phenotypes

Using Ensembl resources

Sanger Institute Wellcome Trust

INITIATIVES TO SPEED UP DATA MINING

25

Rational

Make the most of remarkable advances in the molecular basis of human diseasesdissect the physiological pathwaysimprove diagnosisdevelop treatments

Make rare diseases visible in health information systems to gain insight into themto access real life data already collected

Improve coding of RD whichever coding system used Cross-reference coding systems: Orpha nomenclature, ICD10, MeSH, SnoMed-CT, MedDRA

What is the problem ? Computers are not smart enough….

The following descriptions mean the same thing to you: generalized amyotrophygeneralized muscle atrophymuscular atrophy, generalized

But your computer thinks they're completely unrelated

27

28

Phenomes: a continuum

Group of phenomes

« Disorder » level

Subtypes

• Top of classification = System disorder

• Group

• Clinical criterion• Disease, syndrome,

condition,anomaly…

• Etiological• Clinical• Histopathological…

• No type: waiting to have a type attributed

• Disease• Malformation syndrome• Morphological anomaly• Biological anomaly• Clinical syndrome• Particular clinical situation

Orphan Diseasome

http://research.cchmc.org/od/01/index.html

An Orphan Diseasome permits investigators to explore the orphan disease (OD) or rare disease relationships based on shared genes and shared enriched features (e.g., Gene Ontology Biological Process, Cellular Component, Pathways, Mammalian Phenotype).

The red nodes represent the orphan diseases and the green ones the related genes. A disease is connected to a gene if and only if a mutation which is responsible of the disease has been identified on this gene.

UMLS = Unified Medical Language System ICD = International Classification of Diseases

Since 1863 by WHO Used by most countries to code medical activity, mortality data

MeSH = Medical Subject Headings controlled vocabulary thesaurus used for indexing articles for PubMed by

National Library of Medicine (USA) SnoMed CT = Systematized Nomenclature of Medicine--Clinical Terms

clinical terminology by the International Health Terminology Standards Development Organisation (IHTSDO) in Denmark

Used in the USA and a few other countries MedDRA = Medical Dictionary for Regulatory Activities

medical terminology to classify adverse event information associated with the use of medical products

by the International Federation of Pharmaceutical Manufacturers and Associations (IFPMA)

Different resources, different terminologies

(e)HR:SNOMED CT

Others?

Free text

Mutation/patient registries,databases:

HPOLDDB

PhenoDBElements of morphology

Others? Free text?

Tools for diagnosis:

HPOLDDB

Orphanet

Each terminology has a purpose–driven approach

Indexing health status of individual patients for health management (SnoMED) Detailed, focus on manifestations and complaintsAdapted to clinical habitsAnalytical approach

Indexing health status of individual patients for statistical purpose in public health (ICD)More agregated, interpreted phenotypic featuresAgregated conceptsUnambiguous to avoid blanks

Purpose–driven approach (2) Indexing health status of individual patients for clinical

research purpose (HPO / PhenoDB / Elements of morphology)Highly detailed to fit with the research questionsSpecific terminologies developed for disease-specific

patient registries Indexing health status of individual patients for retrieving

possible diagnoses (LDDB,POSSUM,Orphanet)Agregated concepts Requires a judgement of clinicians about phenomic

expressions that are relevantUnambiguous to avoid blanks

HOW TO MAKE ALL THESE TERMINOLOGIES INTER-OPERABLE ?

35

Convince the terminologies to converge in some way….

Sept 2012: start of mappings (Orphanet) EUGT2 – EUCERD workshop (Paris, September 2012)

ICHPT workshop (ASHG, Boston, October 2013)Selection of 2,300 core terms

PhenoDB

HPO

Orphanet

LDDBElements of Morphology

POSSUMSNOMED CT (IHTSDO)

ICD (WHO)

DECIPHER

Phenotype terminology project• Aims:

– Map commonly used clinical terminologies

(Orphanet, LDDB, HPO, Elements of morphology,

PhenoDB, UMLS, SNOMED-CT, MESH, MedDRA):

• automatic map, expert validation, detection and

correction of inconsistencies

– Find common terms in the terminologies

– Produce a core terminology

• Common denominator allowing to share/exchange

phenotypic data between databases

• Mapped to every single terminology

Mapping Terminologies

• Orphanet: 1357 terms (Orphanet database, version 2008)

• LDDB: 1348 dysmorphological terms (Installation CD)

• Elements of Morphology: 423 terms (retrieved manually from

publication AJMG, January 2009)

• HPO: 9895 terms (download bioportal, obo format, 30/08/12)

• PhenoDB: 2846 terms (given in obo format, 02/05/2012)

• UMLS: (version 2012AA) (integrating MeSH, MedDra, SNOMED

CT)

Tools• OnaGUI (INSERM U729):

ontology alignment tool• Work with file in owl format• I-Sub algorithm: detect

syntaxic similarity• Graphical interface to check

automatic mappings and manually add ones

• Metamap (National Library of Medicine): a tool to map biomedical text to the UMLS Metathesaurus

• Perl scripts: format conversion, launching Metamap, comparison of results…

Comparison of mappings and deduction

• Perl script to compare all the mappings and infer mappings of non-Orphanet terminologiesEg: Orphanet ID XX mapped to YY in HPO and ZZ in LDDB -> deduction: YY and ZZ should probably map

• Retrieve HPO mappings versus UMLS, MeSH

• First figures:

LDDB El. Morpho PhenoDB HPO UMLS…

Orphanet E: 1062 E: 416 E: 978 E: 2228 E: 6948

LDDB D: 275 D: 533 D: 1123 D:2678

El. Morpho D: 177 D: 716 D: 409

PhenoDB D: 1045 D:3268

HPO D: 6307+4800

UMLS…

Mapping of non-Orphanet terminologies

• Automatic and infered mappings were checked by experts– Using OnaGUI for all, except UMLS

Automatic I-Sub: 7.0 + deduction

– Metamap + deduction + HPO mappings• Figures:

El. Morpho PhenoDB HPO UMLS…

LDDB D: 257+23 added

D:528, 92%EA:674, 38%E

D: 1105, 87%EA: 2084, 23%E

D: 2654, 83%EA: 11731

El. Morpho D:174, 50%EA:189, 74%E

D:393, 93%EA: 436, 16%E

D:405, 84%EA:1248

PhenoDB D:1018, 91%EA: 4168, 6%E

D: 3222, 82%EA: 18776

HPO D: 7389A: 65535

UMLS…

First list of common terms Present in at least 2 terminologies Definition of rules for nomenclature Addition of terms present in each terminology as synonyms

Workshop on 21-22 October 2013 in BostonSuccess!

Reviewed 2736 terms appearing 2 or more times in the 6 terminologies in 17 hours

2302 terms chosen, including preferred term Definitions are from Elements of Morphology if

available, and HPO/Stedman’s Medical Dictionary, if not List of terms, mapping to HPO, PhenoDB, Elements of

Morphology will be available at http://ichpt.org by January 2015.

All tools will map to this terminology to allow interoperability among resources

Workshop on Terminologies for RD – Paris, 12 September 2012

Many terminologies in use to describe phenomes - No interoperability

Joint EuroGenTest and EUCERD workshop Organized by Ségolène Aymé Agreement to define a core set of terms

common to all terminologies and a methodology

Core set identified by cross referencing HPO PhenoDB Orphanet UMLS: MeSH, MedDRA, SnoMed CT LDDB Elements of morphology

44

Adoption of a core set of >2,300 terms common to all terminologies

Workshop of validation, Boston21-22 October 2013 Workshop supported by HVP and

EuroGenTest Organized by Ada Hamosh Expert review of the initial proposal Selection of 2,370 terms Decision to propose them for adoption by all

terminologies Establishment of the International

Consortium for Human Phenotype Terminologies – ICHPT

Publication on the IRDiRC website with definitions from HPO Elements of morphology

FROM A TERMINOLOGY TO AN ONTOLOGY

COMPUTERS ARE NOT SMART

45

Why ontologies are needed ?

Ontologies are representations of the knowledge in a way which is directly understandable by computers

Ontologies allow reasoning Ontologies define the objects AND the relationship

between the objectsDuchenne muscular dystrophy (disease) Is a neuromuscular

disease (group of diseases) Schistosomias (disease) Is a cause of anemia (manifestation)

46

Standardization of Phenotype OntologiesWorkshop Sympathy, 19 Apr 2013, Dublin Organized by IRDiRC, supported by the University of Dublin, Forge and EuroGenTestConclusion: Adopt HPO & ORDO & cross-reference with OMIM

Standardisation of Phenotype OntologiesRare Diseases

bioportal.bioontology.org/ontologies/ORDO

Phenotypic Features

bioportal.bioontology.org/ontologies/HP

Based on Orphanet multi-hierarchical classification of RD

Genes– diseases relationships

Cross-references: - For RD nomenclature : OMIM, SNOMED CT,

ICD10, MeSH, MedDRA, UMLS- For genes : OMIM, HGNC, UniProtKB, IUPHAR,

ensembl, Reactome

ICHPT (International Consortium for Human Phenotype

Terminologies)2,307 terms- core terminology

Mapped to: HPO Elements of MorphologyOrphanet LDDBSNOMED CT Pheno-DB (OMIM)MeSH UMLS

Available soon for download at ichpt.org

Please adopt/disseminate HPO and ORDO

to speed up R&Dto the benefit of the patients

49

THEY CAN HELP REPURPOSE DRUGSCOMPUTERS ARE VERY SMART

50

Rational: Make optimal use of molecules already known

Drug Repositioning or Repurposing is a strategy used to generate new or additional value for a drug, by targeting diseases other than those for which it was originally intended Address unmet medical needs Reduce time to market due to provided information on Unbiased clinical safety and efficacy data Add value to exiting porfolio Increase drug pipeline Decrease R&D failure risks Decrease development costs Creates new revenue potential

51

Graph Theory Enables Drug RepurposingGramatica et Al.: PLOS one, Vol 1 e84912, 2014

23 Million articles from PubMed Possible to link the gathered information on drugs,

physiological pathways and resulting biological activities with the pathophysiological signs & symptoms of diseases

Possible to rank the matches in order to identify the most promising leads

52

Graph Theory Enables Drug RepurposingGramatica et Al.: PLOS one, Vol 1 e84912, 2014

53

Graph Theory Enables Drug RepurposingGramatica et Al.: PLOS one, Vol 1 e84912, 2014

54

Conclusion

Open access to dataWe now leave in an open worldIt is an opportunity in researchEvidence that open-access to data is beneficial, especially

for the data producer !Orphadata is accessed by 3000 researchers/ month

Agreed standards to make data interoperable Responsibility of Institutions and of individual

researchers

55

Thank you for your invitation

56