Preparing for ISMB Bio-Ontologies Task

[callahantiff]

Task

Complete the needed tasks in order to get PheKnowLator prepared for the 2021 ISMB Bio-Ontologies Challenge. The specific tasks are organized and described below.

SPARQL Endpoint

Not sure this is actually needed to support the challenge, but if it is, we need to answer the following question:

• [ ] What do we use?
• [ ] Where do we host it?
• [ ] Do we want to think about using something other than SPARQL (i.e. Neo4J) that can support other

Frequent Refresh of KG Builds

• [x] Confirm with organizing committee which pkt build type to use - OWL-NETS versions of the builds are not true property graphs and may not be compatible with the evaluation framework. Probably want an instance and subclass build and their
• [ ] Discuss potential changes that may be needed in order to create monthly KG updates to support challenge
  • What needs to be done in the build framework for ontology preprocessing and other preprocessing steps in order to automate monthly builds
  • Figure out about API set-up and requiring/providing keys for folks to download resources
  • Probably best to support a system that refreshes existing edge lists and not yet worry about sustaining/adding new edges sources yet
  • Set-up CI system that builds it once a month and deposits files (see if this works with GitHub Actions)

[callahantiff]

@bill-baumgartner - confirmed the following details with Rob H this morning:

• OK, SPARQL Endpoint primarily used for evaluation, OK to provide KG data in other formats
• Onboard with also providing a baseline embedding method

[callahantiff]

@bill-baumgartner - I am going to close this and open some new issues that will cover this and the work we discussed last night.