Ontologies and where to find them

Maja Magel
Charlie Pauvert

2024-07-19

Learning objective

  • find an ontology and assess its validity

Find ontologies

Find ontologies

From the general to the specific:

The OBO Foundry

is a community of Open Biology and Biomedical Ontologies (OBO) that share principles and practices

https://obofoundry.org

OBO ontologies

  • are used by many and maintained by few.

  • have non-overlapping terms in relation with others ontologies

  • are responsive to changes by the scientific community

Selecting ontology

A usable and reliable ontology

  • has sound terms definitions that you agree with
  • is actively developed and updated to reflect the scientific practices
  • has a community space to exchange with the community (tickets, issues)
  • has an adequate license and appropriate funding (for maintenance)

Exercise: Ontology look-up

Task using http://obofoundry.org

  • Look-up 1-2 ontologies that match your field or interests:
    • by typing keywords into the search bar
    • or using the Ontology Domains menu
  • Report them on the pad and compare with your neighbor
  • Assess if the ontology matches the quality criteria

(optional) List the dependencies of the ontology and the one of your neighbor

The paradox of standards - What to do when nothing fits

The paradox of standards

A classification [or any standard] fit to support data journeys needs to be dynamic enough to support the ever-changing understanding of nature acquired by the biologists who use it and at the same time stable enough to enable data of various sorts and significance to be quickly surveyed and retrieved. Leonelli (2016, 116)

An edge case

An edge case

What to do?

  1. Don’t give up on ontologies just now
  2. Get a different perspective by discussing your case with colleagues
  3. Contribute back to the community with your edge case

Contributing

Your edge case can help others!

  • Do you want to amend/extend the definition of a term?
  • No term fits your case?

Either way, you need to contact the maintainers of the ontology

Contributing new terms

  • Find the community space of the ontology (e.g., GitHub, Slack)
  • Look-up if someone has the same edge case already
  • Read the Contributing guidelines
  • Identify any request process

Exercise: NTR

NTR: New Term Request

Task

  • Look-up recent examples of ENVO contributions

  • Choose one and skim through the discussion

  • List the information required for a type of contribution

  • (optional) Find the template or recommendations for contribution

  • (optional) What about for UBERON?

Limitations

Limitations

  • Ontologies need to be maintained (funding)

  • Ontologies need to be updated to reflect current knowledge. This needs engaged biologists!

  • Limited accessibility because scientists needs adequate infrastructure and to speak English

  • Sometimes ontologies are not needed (Malone et al. 2016, see Rule 10)

A silver lining

A silver lining

Still, worth trying out ontologies, especially if they are used by metadata standards in your field!

Going further with ontologies

  • Knowledge graphs
  • Reasoning with ontologies
  • Large Languages Models
  • Extending publications

References

Workshop Biometadata-03

Goldmann, Ulrich, Emma Vos, and Philippe Rocca-Serra. 2023. “Requesting New Terms.” https://w3id.org/faircookbook/FCB021.
Leonelli, Sabina. 2016. Data-Centric Biology: A Philosophical Study. Chicago ; London: The University of Chicago Press.
Malone, James, Robert Stevens, Simon Jupp, Tom Hancocks, Helen Parkinson, and Cath Brooksbank. 2016. “Ten Simple Rules for Selecting a Bio-Ontology.” PLOS Computational Biology 12 (2): e1004743. https://doi.org/10.1371/journal.pcbi.1004743.
Rocca-Serra, Philippe, Susanna-Assunta Sansone, Danielle Welter, and Alasdair J. G. Gray. 2023. “Selecting Terminologies and Ontologies.” https://w3id.org/faircookbook/FCB020.