Deep phenotyping for everyone

Melissa Haendel, PhDJuly 9th, 2016

Phenoday!!

@monarchinit@ontowonkahaendel@ohsu.edu

The genome is sequenced, but…

…we still don’t know very much about what it does

3,435 OMIM

Mendelian Diseases with no known genetic basis

?66,396 ClinVar

Variants with no known pathogenicity

Why we need all the organisms

Model data can provide up to 80% phenotypic coverage of the human coding genome

The prevailing clinical diagnosis pipelines leverage only a tiny fraction of the available

data

PATIENT EXOME/ GENOME

PATIENT PHENOTYPES

PATIENT ENVIRONMENT

PUBLIC GENOMIC DATA

PUBLIC HUMAN & MODEL PHENOTYPE,

DISEASE DATA

PUBLIC HUMAN & MODEL ENVIRONMENT,

DISEASE DATA

POSSIBLE DISEASES

DIAGNOSIS & TREATMENT

Under-utilized data

monarchinitiative.org

PROBLEM Diagnosis / treatment / prognosis on gestalt

(Experience, intuition, and pattern recognition) Things are not always what they first seem Errors are common, and up to 35% of errors

cause harm It takes patients @ six years from noticing

symptoms to being diagnosed with trips to eight physicians

25% of patients having to wait between 5 and 30 years

HYPOTHESISDiagnosis, treatment and prognosis may be informed and complemented by democratized deep phenotyping that is easier to compute, collect, and exchange

Ulcerated paws

Palmoplantar

hyperkeratosis

Thick hand skin

Image credits:

"HandsEBS" by James Heilman, MD - Own work. Licensed under CC BY-SA 3.0 via Commons – https://commons.wikimedia.org/wiki/File:HandsEBS.JPG#/media/File:HandsEBS.JPG

http://www.guinealynx.info/pododermatitis.html

Challenge: Each database uses their own vocabulary/ontology

MPHP

MGIHPOA

Challenge: Each database uses their own phenotype vocabulary/ontology

ZFA

MPDPO

WPO

HP

OMIA

VT

FYPO APOSNOMED

………

WB

PB

FB

OMIA

MGI

RGD

ZFIN

SGD

HPOA

EHR

IMPCOMIM

…QTLd

b

Can we help machines understand phenotype terms?

“Palmoplantar hyperkeratosis”

Human phenotype I have absolutely no

idea what that means

???

The Human Phenotype Ontology for deep phenotyping

Decomposition of complex concepts allows interoperability

Mungall, C. J., Gkoutos, G., Smith, C., Haendel, M., Lewis, S., & Ashburner, M. (2010). Integrating phenotype ontologies across multiple species. Genome Biology, 11(1), R2. doi:10.1186/gb-2010-11-1-r2

“Palmoplantar

hyperkeratosis”

increased

Stratum corneum

layer of skin

=Human phenotype

PATO

Uberon

Species neutral ontologies, homologous concepts

Autopod

keratinization

GO

Harmonizing diseases, phenotypes, anatomy, and genotypes

We learn different phenotypes from different organisms

Diagnosing an undiagnosed disease

Putting all that data to use to diagnose a rare platelet syndrome

http://bit.ly/stim1paper

Phen

otyp

ic profi

le

Gene

s

Heterozygous, missense mutationSTIM-1

MGI mouse

N/A

Heterozygous, missense mutation

STIM-1

N/A

Ranked STIM-1 variant maximally pathogenic based on cross-species G2P data,

in the absence of traditional data sourceshttp://bit.ly/exomiser

Stim1Sax/Sax

What about patients? Can they phenotype themselves?

The GenomeConnect surveyEg. Machado-Joseph disease might present:

With: Ptosis, andAbnormalities of:

eye movement,

globe size, vision, optic nerve

Without: Myopia, Hypermetropia, andAbnormalities of:

the retina, the iris, and the lens

Annotation sufficiency determination per disease

Patient self-reported HPO profile HPO reference profile

Comparison

Ensure that the survey is maximally diagnostic

Disease 1

Disease 2

Disease 3

Disease 4

Disease 7500

HPO reference profile

HPO reference profile

HPO reference profile

HPO reference profile

Patient self-reported HPO profile

Patient self-reported HPO profile

Patient self-reported HPO profile

Patient self-reported HPO profile

Simulated GenomeConnect survey HPO Profiles

Monarch Initiative reference HPO Profiles

Ensure that the survey is maximally diagnostic

Patient

Expert

Phenotypic Profile overlap

Compare phenotypic profiles

For every known disease, fill the survey and ask:Does the profile match the disease best based on the survey mapping?

GC HPO profile

HPO reference profile

GC HPO profile

GC HPO profile

GC HPO profileHPO reference

profile

HPO reference profile

HPO reference profile

Comparisons

Assess patient-derived profile generation

HPO profile

HPO profile

HPO profile

HPO profile

HPO profile

Disease 1

Disease 2

Disease 3

Disease 4

Disease 7500 HPO reference profileGC HPO profile

Assess patient-derived profile generation

Patient

Expe

rtPh

enot

ypic

Pro

file

over

lap

Compare phenotypic

profiles

For every diagnosed patient:Can the patients utilize the survey and retrieve the correct disease?

GC HPO profile

Clinical evaluation HPO profile

GC HPO profile

GC HPO profile

GC HPO profile Clinical evaluation HPO profile

Clinical evaluation HPO profile

Clinical evaluation HPO profile

GC HPO profile Clinical evaluation HPO profile

Comparison

Determine the contribution and sufficiency of patient self-phenotyping

Patient 1

Patient 2

Patient 3

Patient 4

Patient 7500

Determine the contribution and sufficiency of patient self-phenotyping

UDN patient generated GenomeConnect survey

HPO profile

UDN patient Clinical evaluation HPO profile

Patient

Expert

Phenotypic Profile overlap

Compare phenotypic profiles

Human Phenotype Ontology, now with 6,200 plain language

synonyms for patients, families, and non-

experts

www.human-phenotype-ontology.org@HP_ontology

Almost half of the 14k synonyms are plain language

All synonyms Plain language synonyms

Introducing PhenoPackets

It’s exactly what you think it is:a packet of phenotype data to be used

anywhere, written by anyone

Genes Environment Phenotypes+ =

Biology central dogma

Standards for encoding and exchanging data

must be up to these challenges

@ontowonka

The relationships too must be captured

It is not just the bits…

G-P or D (disease)causescontributes tois risk factor forprotects againstcorrelates withis marker formodulatesinvolved inincreases susceptibility to

G-G (kind of)regulatesnegatively regulates (inhibits)positively regulates (activates)directly regulatesinteracts withco-localizes withco-expressed with P/D - P/Dpart ofresults inco-occurs withcorrelates withhallmark of (P->D)

E-Pcontributes to (E->P)influences (E->P)exacerbates (E->P)manifest in (P->E) G-E (kind of)expressed inexpressed duringcontainsinactivated by

Genes Environment Phenotypes+ =

Computable encodings are essential

Base pairsVariant notation (eg. HGVS)

Human Phenotype Ontology

Mammalian Phenotype Ontology

Medical procedure codingEnvironment Ontology

@ontowonka

Genes Environment Phenotypes

VCF PXFGFF

Standard exchange formats exist for genes …

but for phenotypes? Environment?

NEW

BED

@ontowonka

If it is alive, it can be PhenoPackaged

Some biodiversity images adapted from http://i.vimeocdn.com/video/417366050_1280x720.jpg

Model Organisms

Biodiversity Crops Domestic Animals

Disease vectorsEpidemiologic

al Monitoring

Drug discovery &

Development

Rare Disease

Diagnosis

Personalized

Medicine

Environmental Monitoring

Patients & Cohorts

Genetic Engineering

Mechanistic

Discovery

Phenopackets for organisms

This is “Maru”,a 4-year-old,

male cat of the Scottish Fold

breed

abnormal sheltering behavior

[MP:0014039] (onset at birth)Bi

ogra

phy

Phen

oty

pes &

qu

alifi

ers

youtube.com/user/mugumoguWeighs 6kg

Mea

sure

me

nts So

ur ce

title: "age of onset example"persons:- id: "#1" label: "Donald Trump" sex: "M"

phenotype_profile:- entity: "person#1" phenotype: types: - id: "HP:0200055" label: "Small hands" onset: description: "during development" types: - id: "HP:0003577" label: "Congenital onset" evidence: - types: - id: "ECO:0000033" label: ”Traceable Author Statement" source: - id: "PMID:1"

Image credits: upi.com

Phenopackets for humans

Canonical JSON format

Phenopackets for Patients

Image credits: ngly1.org

• Dry eyes• Developmental delay• Elevated liver function

phenotype_profile:- entity: ”patient16" phenotype: types: - id: "HP:0000522" label: ”Alacrima" onset: description: ”at birth" types: - id: "HP:0003577" label: "Congenital onset" evidence: - types: - id: "ECO:0000033" label: ”Traceable Author Statement" source: - id: ” https://twitter.com/examplepatient/status/123456789"

• Patient registries

• Social media

Phenopackets for journals

Each article can be associated with a

phenopacket

Robinson, P. N., Mungall, C. J., & Haendel, M. (2015). Capturing phenotypes for precision medicine. Molecular Case Studies, 1(1), a000372. doi:10.1101/mcs.a000372

Each phenopacket can be shared via DOI in any repository outside paywall (eg. Figshare, Zenodo, etc)

PhenoPacket formats

CSV JSON RDF OWL

Export phenopacket to

The PhenoPackets ecosystem

www.phenopackets.orghttps://github.com/phenopackets/

So, do you expect us to put these together ourselves?

Emerging tool: WebPhenote (based on Phenote)

create.monarchinitiative.org

WebPhenoteForm-based Graph-based

Noctua / LEGO inside

Thank you!Deep Phenotype and have a magical

day

Community engagement surveybit.ly/monarchcommunity

AcknowledgementsLawrence Berkeley

Chris MungallSuzanna LewisJeremy NguyenSeth Carbon

CharitéPeter RobinsonSebastian Kohler

RTIJim Balhoff

CyverseRamona Walls

U of PittsburghHarry Hochheiser

OHSUMatt BrushKent ShefchekJulie McMurryTom ConlinNicole Vasilevsky

Queen Mary College London

Damian SmedleyJules Jacobson

GarvanTudor Groza

Alfred Wegener Pier Buttigieg

GSoCSatwik Bhattamishra

FUNDING: NIH Office of Director: 1R24OD011883; NIH-UDP: HHSN268201300036C, HHSN268201400093P, Phenotype Ontology Research Coordination Network (NSF-DEB-0956049)With special thanks to Julie McMurry for excellent graphic

design