Proposal: Community Strategy for Archiving Orphaned/Defunct Chemical Ontologies (REX, FIX, Gainsville Core)

[StroemPhi]

Context

Several valuable chemical ontologies developed in the early 2000s and 2010s are currently in a precarious state, risking the loss of encoded domain knowledge. Specifically, Physico-chemical process (REX) and Physico-chemical methods and properties (FIX) by EBI, and the Gainsville Core Ontology (GC), developed by Chemical Semantics, Inc. for publishing computational chemistry results, are no longer served by their original institutions and lack active curation, effectively rendering them "orphaned."

While rescue efforts exist, they present distinct challenges for integration. Chris Mungall has notably rescued REX and FIX by creating ODK-based repositories in his personal GitHub space (https://github.com/cmungall/fix & https://github.com/cmungall/rex) (see also: OBOFoundry/OBOFoundry.github.io#2881); however, these versions currently lack an explicit license statement, which is actually an exclusion criterion for indexing in the NFDI4Chem Terminology Service (TS). In contrast, a mirror for GC has been established under the NFDI4Chem organization (https://github.com/NFDI4Chem/GC-Ontology-Mirror/). Yet, such fragmented hosting, spanning personal rescue repos, organization mirrors, or Internet Archive snapshots, poses several problems regarding their indexing in look-up services. This lack of formal indexing obstructs the vision of an Ontologies4Chem Canon, which treats the collection of ontologies and their documentation as a scholarly discourse to be studied, referenced, and understood. As a library-based service, the TIB/NFDI4Chem TS aims to provide permanent, discoverable access to these resources so researchers can properly look up, study, and cite legacy concepts. A community strategy is needed to consolidate these resources into a unified archive, verify or rectify licensing, and implement clear "defunct" warnings to prevent accidental reuse while fulfilling the archival mission of preserving scientific memory.

Expected outcome of the workshop

The primary goal is to reach a community consensus on a general strategy for handling outdated, defunct, or orphaned ontologies within the chemistry domain. Specifically, we aim to:

• Validate the proposal of creating a centralized "Archive" repository under the NFDI4Chem organization to host static versions of such ontologies, distinct from active ODK-based repos/projects.
• Agree on metadata standards for these archives, specifically the implementation of prominent warning banners in indexing services (like the NFDI4Chem TS) to clearly signal their "orphaned" status to users.
• Discuss the licensing remediation process (e.g., retroactive CC0/CC-BY declaration with original authors) to enable legal indexing and mapping.
• Gather feedback on the proposed workflow: Archiving -> Warning-Indexed Indexing -> Community-driven SSSOM mappings to active ontologies.

Prerequisites issues/tasks

• License Clarification: Contacting EBI (for REX/FIX) to secure permission for retroactive licensing (CC0/CC-BY) to resolve current indexing barriers.
• Technical Prototype: Preparing a mock-up or description of the proposed "warning banner" and metadata schema for the NFDI4Chem Terminology Service to demonstrate how "defunct" ontologies would be displayed.
• Stakeholder Input: Gathering initial thoughts from OBO Foundry representatives regarding the PURL redirection issues and the feasibility of a centralized archive approach.

Implementation issues/tasks

This session will be structured as a presentation followed by a guided community discussion:

• Presentation (10-15 mins):
  • Overview of the current status of REX, FIX, and Gainsville (history, current hosting issues, licensing gaps).
  • Presentation of the "Centralized Archive" proposal (NFDI4Chem org repo vs. personal repos).
  • Demonstration of the proposed "Warning Banner" concept in the TS.
• Interactive Discussion (30-40 mins):
  • Facilitated debate on the proposed strategy: Is a centralized archive the right approach? What are the risks?
  • Discussion on governance: Who decides when an ontology is "orphaned"? Who maintains the archive repo?
  • Consensus building on the metadata requirements (warnings, licensing) for indexing such ontologies.
  • Next steps: Defining a working group or action plan to implement the agreed strategy post-workshop.

[cthoyt]

a few thoughts on this:

Validate the proposal of creating a centralized "Archive" repository under the NFDI4Chem organization to host static versions of such ontologies, distinct from active ODK-based repos/projects.

I don't think NFDI4Chem has sufficient personnel with experience, time, and motivation to be an effective umbrella to centralize archival and maintenance of orphaned ontologies. I would suggest considering a different kind of governance structure like https://github.com/jazzband/ where as many barriers are removed from maintenance as possible. https://github.com/bio-ontology-research-group/unit-ontology also did something similar - anyone can make PRs and they will just get accepted.

Relevant: OBOFoundry/OBOFoundry.github.io#2881

Discussion on governance: Who decides when an ontology is "orphaned"? Who maintains the archive repo?

I am so excited about this! I would love to establish an actionable, objective measure of what it means to be orphaned.

Related discussion: OBOFoundry/OBOFoundry.github.io#1920 and see also other issues on the OBO Foundry tracker: https://github.com/OBOFoundry/OBOFoundry.github.io/issues?q=is%3Aissue%20state%3Aclosed%20orphaned

Technical Prototype: Preparing a mock-up or description of the proposed "warning banner" and metadata schema for the NFDI4Chem Terminology Service to demonstrate how "defunct" ontologies would be displayed.

FIX and REX are both annotated as orphaned on OBO Foundry and show a banner describing what OBO Foundry considers as orphaned, e.g., see https://obofoundry.org/ontology/fix

---

All that being said, I don't think this will be a very productive discussion. I wouldn't suggest setting aside so much time for it.

[K-r-ll]

Hi there, yes @StroemPhi I have created FIX more or less the same times as REX, that is, some 25 years ago... Can't say I am the maintainer coz I didn't touch it for more than 20 years now

[StroemPhi]

@K-r-ll thanks for your quick reply! It's not about maintainership but creatorship here. So we'll make sure to put you down as dcterms:creator in their header. Please let us know who else might need to be credited as creator or contributor.

[cthoyt]

updates for both of chris's repos:

• Add CC-BY 4.0 license and structured creator annotation cmungall/fix#3
• Add CC-BY 4.0 license and structured creator annotation cmungall/rex#2

[K-r-ll]

In case anyone is interested, which I doubt, the only publication explicitly mentioning FIX is this conference paper:

Degtyarenko, K. (2003) Chemical vocabularies and ontologies for bioinformatics. In Collier, H., Ed. Proceedings of the 2003 International Chemical Information Conference, Nîmes, France, 19–22 October 2003. Infonortics, Tetbury, pp. 144–162.
https://www.academia.edu/2428642/Chemical_vocabularies_and_ontologies_for_bioinformatics

I did not mention REX there but there is something named "Ontology of biochemical reactions" that must be a top-level REX, without IDs. Cf. Figure 1 (p. 164) in another paper:

Degtyarenko, K. (2004) Controlled vocabularies and ontologies in enzymology. In Hicks, M.G. & Kettner, C., Eds. 1st International Beilstein Workshop on Experimental Standard Conditions of Enzyme Characterization. Beilstein-Institut, pp. 143–174.
https://www.academia.edu/2428648/Controlled_vocabularies_and_ontologies_in_enzymology

[StroemPhi]

Thanks alot @K-r-ll, we are indeed intersted in such links ;)

[StroemPhi]

Given that the issue around the licence of FIX and REX is resolved, I'd now like to focus in this discussion, and the workshop (talk) proposal it represents, on the documentation and archiving part. So for me the preferred solution would be to have one central repo for any chemical ontology that is not maintained anymore, in which we can keep their source code and also collect as much as possible context information. This would allow to have one single point of truth for the chemistry community, to which we and ours various systems can point to. Given that the Gainsville Core is not an OBO ontology, pushing this documentation task onto the OBO comunity seems a bit unreasonable to me. I don't really care where such a unified archive repo should live, as long as we have it and can all openly contribute to the documentation.