1 of 23

Mondo Disease Ontology:

Building a Community-driven Disease Resource

Sabrina Toro

University of North Carolina at Chapel Hill�Monarch Initiative | Translational and Integrative Sciences Lab

GA4GH rare disease working group

June 26th, 2024

2 of 23

Slides are available here:

tislab.org/mondo-GA4GHrare2024

3 of 23

Mondo Team

mondo.monarchinitiative.org

Nico Matentzoglu

Semantic Engineer

Joe Flack

Semantic Engineer

Harshad Hegde

Semantic Engineer

Katie Mullen

Ontology Curator

Sabrina Toro

Lead ontology curator

Nicole Vasilevsky

Ontology Curator

Ada HamoshMedical Expert, PI

Chris MungallCreator

Semantic Engineer, PI

Peter RobinsonMedical Expert, PI

Melissa Haendel

PI

Mondo�Community

Trish WhetzelOntology Curator, Semantic Engineer�Mondo Technical Lead

Monica Munoz Torres

Program Director

Sarah GehrkeProject Manager

4 of 23

Integration and comparison of disease data support diagnosis and treatment discovery

Disease Diagnosis

Treatment discovery

data comparison

GENE

VARIANT

TREATMENT

EXPOSURE

Source 1

Source 2

GENE

VARIANT

TREATMENT

EXPOSURE

PHENOTYPE

PHENOTYPE

DISEASE

DISEASE

data integration

5 of 23

Integration and comparison of disease data support diagnosis and treatment discovery

Disease Diagnosis

Treatment discovery

data comparison

GENE

VARIANT

TREATMENT

EXPOSURE

Source 1

Source 2

GENE

VARIANT

TREATMENT

EXPOSURE

PHENOTYPE

PHENOTYPE

DISEASE

DISEASE

data integration

6 of 23

Different communities annotate diseases at different levels of granularity and use different vocabularies and terminologies

7 of 23

Mondo was created by evidence-based merging of equivalent disease terms

8 of 23

Mondo covers a broad scope of diseases and reconciles disease classification and mappings between terminologies

  • Mondo integrates key medical and biomedical disease terminologies
  • Scope:
    • all disease types (genetic, rare, cancers,etc)
    • human and non-human animals
  • Mondo unifies disease classification
  • Mondo reconciles mappings between disease resources

disease

human disease

non-human disease

infectious disease

hereditary disease

cancer or benign tumor

nervous system disorder

respiratory system disorder

Mondo high-level classification

9 of 23

Example of Mondo term: Alzheimer disease (MONDO:0004975)

definition

hierarchy

synonyms

+

source

X-ref* / IDs of corresponding concepts in other sources*

unique permanent ID

  • exact synonyms
  • related synonyms �(used in the literature, but the usage is not strictly correct)

*mappings also available in a separate file

'disease has location' some 'central nervous system' (UBERON:0001017)

‘disease has feature' some 'Neurofibrillary tangles' (HP:0002185)

relations to other ontologies

10 of 23

Mondo is an open community driven resource for diseases

  • Mondo is a logic-based ontology that
    • integrates key medical and biomedical disease terminologies
    • provides a unified classification of diseases
  • Mondo provides precise, curated semantic mappings between terminologies
  • Mondo is a resource for the community by the community.

mondo.monarchinitiative.org

11 of 23

Mondo users and use cases

  • Disease annotations
    • Databases
    • ClinGen : disease - gene variants annotations
    • Pombase : disease annotations
    • OMIA (Online Mendelian in Animals) use Mondo to refer to their diseases
  • Data integration:
    • European Bioinformatics Institute (EBI)
      • Mondo is used as the primary ontology for disease concepts integrated into the Experimental Factor Ontology (EFO) for integration of data across EBI.
    • Gabriella Miller Kids First Data Resource Portal
      • use to structure diagnosis information
  • Data comparison and diagnostic/treatment prediction
    • Monarch Initiative (cross-species disease discovery)

12 of 23

Representation of rare diseases in Mondo

13 of 23

The “rare disease problem” : what is a rare disease?

  • Different criteria of “rare” depending on the geographical location
    • “prevalence”, i.e. the number of affected individuals per population at a given point in time
      • 1:1500 individuals in the US population
      • 1:2000 individuals in the EU population
      • 1:2500 individuals in Japanese population
    • Additional criteria specific to some countries�e.g “life-threatening or chronically debilitating diseases” is a criterion for “rare disease” in the EU
  • One disease can be rare in one geographical location, but not in another
    • e.g. “Kawasaki Disease” is considered rare in Europe, but common in Japan
  • Different definitions of “rare”
    • What diseases should be counted as “rare”?
    • e.g. should subtypes of disease considered “rare”?

14 of 23

Mondo aims to be a global resource, relevant and available for any group, regardless of their location or definition of rare disease.

  • Mondo harmonizes and integrates all rare disease concepts, according to rare disease authoritative sources.
  • Mondo includes metadata allowing users to choose which rare disease criteria is appropriate for their use case.
  • Rare diseases are integrated in the ontology alongside other non-rare diseases, offering context and facilitating computational analysis.
  • Open-source alignment system is designed to be robust and scalable, ensuring continuous synchronization with external resources within the Mondo framework.

15 of 23

Mondo rare subsets according to rare disease authorities.

NORD rare list

NORD

GARD rare list

GARD

Orphanet rare list�(=“disorder”)

Orphanet

MONDO

subset: orphanet-rare

subset: nord-rare

subset: gard-rare

16 of 23

Mondo rare subsets based on manual curation and ontology inference.

MONDO

MONDO

  • Only subset manually curated
  • Must have PMID + ORCID as source

NORD rare list

NORD

GARD rare list

GARD

Orphanet rare list�(=“disorder”)

Orphanet

MONDO

subset: orphanet-rare

subset: nord-rare

subset: gard-rare

Rare diseases from publication/curation/…

Ontology inference

Ontological children of a rare disease are themselves rare diseases

subset: mondo-rare

subset: inferred-rare

17 of 23

Mondo “rare” subset is the union of all rare subsets.

subset: rare

MONDO

subset: orphanet-rare

subset: nord-rare

subset: gard-rare

subset: mondo-rare**

subset: inferred-rare

NORD rare list

NORD

GARD rare list

GARD

Orphanet rare list�(=“disorder”)

Orphanet

MONDO

subset: orphanet-rare

subset: nord-rare

subset: gard-rare

Rare diseases from publication/curation/…

Ontology inference

18 of 23

Examples diseases are rare according to all rare authorities, or only according to one

19 of 23

Only 35% of rare disease terms in Mondo are “rare” according to all rare disease authorities

20 of 23

Mondo includes cross-references and mappings to disease resources, including rare disease resources.

Authoritative rare disease source

21 of 23

Rare disease subsets available in Mondo

mondo-rare subset

mondo

  • orphanet-rare (rare diseases according to Orphanet)
  • nord-rare (rare diseases according to NORD)
  • gard-rare (rare diseases according to GARD)
  • mondo-rare (manually curated by Mondo team)
  • inferred-rare (ontologically inferred)�+ High level classification

Rare diseases can be filtered by a chosen rare disease authority

22 of 23

Upcoming work

  • Linkout to external curated sources (e.g. NORD, ClinGen, etc)
  • Including new “rare disease authoritative” sources based on user requests
  • “Disease groups”
    • E.g “neurological disorder”, “muscle disorder”,...
    • To support display and filtering based on disease group

23 of 23

Summary

  • Mondo is a logic-based ontology that integrates key medical and biomedical disease terminologies, including rare disease, and provides a unified classification of diseases, and precise, curated semantic mappings between terminologies
  • Mondo rare disease subset is based on rare disease authoritative sources
  • Mondo rare disease subset includes:
    • orphanet-rare (rare diseases according to Orphanet)
    • nord-rare (rare diseases according to NORD)
    • gard-rare (rare diseases according to GARD)
    • mondo-rare (manually curated by Mondo team)
    • inferred-rare (ontologically inferred)�+ high level classification
  • Mondo terms includes cross-references to disease sources/terminologies for both rare and non-rare diseases.
    • Support integration and prediction tools
  • Workflow to include more rare disease subsets is scalable: we could add “any” authoritative rare disease sources.
  • Rare diseases are in the context of non-rare diseases