20170410 Ontology Change Improvement Call

Date: April 10, 2017

Attendees: Brian Lowe, Marijane White, Tatiana Walther, Muhammad Javed, Graham Triggs, Juliane Schneider, Damaris Murry, Christian Hauschke, Dong Joon (DJ) Lee,

Goals: Last time we talked about impact of changes, issues Javed identified, asking sites to run SPARQL queries

Note: Marijane is going to do a terrible job of taking notes today.

Starting off with Javed's list of known issues. https://www.dropbox.com/s/976depqlqrnah65/ontologyErrors.docx?dl=0

Javed: While working on the new ontology documentation/listing tool, realized a number of errors in the ontology

• vivo:dateTime is typed as an Object Property, should be a Datatype Property

Marijane: source of this error is mysterious, not in ISF or in VIVO source code

Javed: Continued summary of issues:

• Membership diagram was wrong, had MembershipRole
• another mistyped property
• OCRe - Ontology of Clinical Research: this does not feel like a core namespace, purl is dead, URIs have changed, some properties no longer exist
• OBO Relation Ontology: has_agent ID has changed
• SKOS: rdfs:labels have changed. skos:prefLabel is not in VIVO
• VCard: vcard:Telphone is missing?

Marijane:

• Agree that Clinical Research should be pulled out into a module, perhaps a clinical module
• Github issues for the rest? Getting into process prematurely?
• Extract VIVO domain ontology from ISF? Before we start making changes we need to figure out where the changes are going to be made.

Javed: Nobody is using the ISF?

Marijane: It's unclear. ISF is the CTSA data standard, and the eagle-i ontology is extracted from the ISF. VIVO domain ontology could be extracted too.

Brian: yes, no current process, was done once manually in 2013. why not start with what eagle-i is doing?

Marijane: Agreed, we could start with what eagle-i is doing. Two approaches

• Shahim's tools, which are used in eagle-i https://github.com/ShahimEssaid/OwlCL
• OBO ROBOT tool, does similar things https://github.com/ontodev/robot

Shahim's tools may have documentation at eagle-i somewhere ROBOT has a good example set, fork the repo and run on your machine locally.

Javed:

• so if we extract vivo-core from VIVO-ISF, who decides?
• how do you differentiate between the ISF and the VIVO domain ontology? Will entities in VIVO domain ontology ALWAYS comes from ISF? Can new entities be defined in VIVO domain ontology that are NOT extracted from ISF? If No, We are good. If Yes, it is possible, then there should be differentiation between them.

Marijane:

• the community decides - VIVO sites and developers, eagle-i, openRIF. mostly a question of process and communication.
• differentiating is a good question. would like to change names.

Christian:

• thank you for explanation of relationship of the projects
• can this be documented?

Marijane gave a tour of the existing github repositories. Which one should we use? we have "community" and "vivo-isf-ontology", which is confusing. decision: triage the issues in https://github.com/openrif/community. move to https://github.com/openrif/vivo-isf-ontology issues as appropriate. Goal: remove the community repository, put issues in the vivo-isf-ontology repo.

Javed: still have a question about namespaces. Term "core" in the URI. The ISF ontology uses "core" in its namespace. That means, we cannot differentiate using terms "ISF" and "Core". They represent the same?

Brian:

• "core" ontology vs domain ontology: it used to be that "core" referred to the entities defined in the http://vivoweb.org/ontology/core# namespace. The subset of VIVO-ISF that the VIVO software shipped by default was referred to as the "VIVO" ontology, while the actual file served from the http://vivoweb.org/ontology/core URL was the "public" ontology.

Javed: agrees with marijane about name changes now =)

Marijane: ISF has a confusing structure

Brian: Explains why VIVO-ISF still refers to VIVO: Originally, the top level -- encompassing both VIVO and eagle-i -- was just supposed to be the "ISF." But then there was pushback because this name seemed too generic and confusing, and there was concern about not making the CTSAs, who had voted the "VIVO Ontology" as a "standard," feel like things had been fundamentally changed. So there was pressure to include the VIVO brand name at the top level.

Javed: Question for Graham - where are files saved in the VIVO source code?

Graham: Yes only ontology I work with is in the VIVO source code

Javed: maybe we should compare that too

Action items:

• triage the issues in the community repo, with the goal of eliminating it.
• obtain a copy of the eagle-i OWL file (see browseable ui at https://search.eagle-i.net/model/)
• Compare contents of eagle-i with the public vivo OWL file
• Comparing contents of ontology in the VIVO source code with the public vivo OWL file
• improve documentation on the ISF
• define process for keeping ISF up to date
• start figuring out how to extract the vivo domain ontology
• Compare extraction approaches: Shahim's OwlCL tool vs ROBOT
• figure out better names for all these things