Actions

ConferenceCall 2023 02 15: Difference between revisions

Ontolog Forum
Line 84: Line 84:


== Discussion ==
== Discussion ==
[11:57] RaviSharma: welcome everyone
[12:06] James Overton: Here's a link to the slides: https://docs.google.com/presentation/d/1oKxBzSCZGuFlIhw6jAXw00xgEGs7KgmaVR54NEdPprw/edit#slide=id.g20803e73efb_0_85
[12:09] Charlie Hoyt: not just prefixes but also CURIEs and IRIs
[12:11] RaviSharma: James  and Nico - shared design patterns - are these used from a design repository so when people develop ontologies, they can use the design templates?
[12:14] James Overton: Ravi: Many OBO projects use their own templates and patterns, but we are only beginning to use shared patterns. This is a key part of the COB effort.
[12:18] Charlie Hoyt: https://dashboard.obofoundry.org/
[12:18] Charlie Hoyt: actual dashboard: obofoundry
[12:18] Charlie Hoyt: https://dashboard.obofoundry.org/dashboard/index.html
[12:25] RaviSharma: I found a neat cross reference as to which other ontologies use a particular ontology that you are referencing.
[12:25] DOS: Very granular cell ontologies may define classes using reference data rather than textual definitions.  It can therefore be impossible to fulfill the unique text def requirement.  At best one can add a text definition that described the general class => non-uniqueness.  Because of this, the most detailed cell type ontology on the Foundry - FBbt - fails this test.  Would it be possible to modify the test to allow it to be fulfilled by links to reference data?
[12:26] RaviSharma: How do you determine content accuracy by engaging the experts like doctors or nurses before you determine its readiness for practicing users operational maturity!
[12:27] Charlie Hoyt: i am not familiar with the mondo custom OBO dashboard! This is so cool! For sure we need more strict requirements than what is currently in OBO Foundry
[12:30] James Overton: DOS:  Good point. The automated check for textual definitions is more strict than then principle, which says "most" terms should be defined. <https://obofoundry.org/principles/fp-006-textual-definitions.html>, we can change the automated check, but we need to make a decision about whether to err on the side of strictness or leniency.
[12:30] James Overton: The principles were not written with automated checks in mind, and some are harder to automate than others.
[12:33] DOS: Will promotion to the top tier require passing all tests, or will there be some attempt to order based on how many tests are failing?
[12:34] James Overton: My opinion is that better dashboard score should rank higher, but there's no decision yet.
[12:35] James Overton: OQUAT: https://cthoyt.github.io/oquat
[12:35] Nico Matentzoglu: Presentation link: https://docs.google.com/presentation/d/1oKxBzSCZGuFlIhw6jAXw00xgEGs7KgmaVR54NEdPprw/edit#slide=id.g1dcecf06728_0_7
[12:37] DOS: @James -I'd like to explore discuss how we might extend the principle to allow for a combination of free text def + formalised reference data links to be allowed.  Without this, I feel like FBbt is being penalised for being so detailed - where less detailed and so less useful ontology would pass.
[12:38] Nico Matentzoglu: CL is the cell ontology, CLO is the cell line ontology
[12:39] Alexander Diehl: Jonathan Bard, one of the originators of the Cell Ontology, in case anyone is curious (JB:JB)
[12:39] James Overton: DOS: I agree in general. I work on MHC Restriction Ontology (MRO) which is mostly complexes of proteins, where the logical definition says everything important. We could auto-generate textual definitions, but would that really MRO better? So I would like to see OBO refine the textual definition check.
[12:40] RaviSharma: So how did you find resolution of prefix issues?
[12:40] James Overton: https://github.com/cthoyt/obo-community-health
[12:41] James Overton: ^^ I would like to see OBO refine both the textual definition principle and the automated check.
[12:42] Nicole Vasilevsky: For the OBO Community Health, is this up to date? I pushed to Mondo yesterday and it says the last push was in September
[12:42] RaviSharma: I am surprised that Envo is not being updated since August last year, does that mean that ENVO is stabilized or that development is slow?
[12:43] James Overton: Nicole: I see "2023-02-15" as the update date for MONDO.
[12:43] DOS: I love the Community Health dashboard.
[12:43] Nicole Vasilevsky: Oh, I see it was probably an old screenshot! Ignore me
[12:44] RaviSharma: Is quality score or effort a metadata?
[12:44] Gary BC: @Ravi  I think that ENVO was updated 2 days ago.
[12:45] DOS: Great to be able to point potential users to this so they can tell if an ontology they want to use is active.
[12:45] DOS: Also great to see some super-editors getting credit.
[12:44] Nico Matentzoglu: Just to make it clear: Charlie’s view on “trust” is his own :P not the OBO foundry view!
[12:45] Nico Matentzoglu: (On a personal level I agree with much of it)
[12:46] Nico Matentzoglu: Anita you can start sharing
[12:46] Allen Baron: How is the OBO community health score calculated?
[12:46] Asiyah: Some experts have significant technical barriers to use GitHub. The Geeks here should think of a way to help them!
Reply by Nico: I think sending people to special training is key. Organisations use tools like JIRA - GitHub is not too different then these. I would start by not calling it Geeky - I would think that using project management tools is pretty much a must in the modern era, in GitHub is not much different. I would recommend organising training in your organisations to raise awareness!
[12:47] Allen Baron: Nice presentation Charlie.
[12:48] Nico Matentzoglu: Instruction: https://oboacademy.github.io/obook/howto/deploy-custom-obo-dashboard/
Example custom dashboard: https://monarch-initiative.github.io/monarch-ontology-dashboard/dashboard/index.html
[12:50] Nico Matentzoglu: https://github.com/OBOFoundry/dashboard-template
[12:48] RaviSharma: Gary thanks for reply but I guess it was an old screenshot
[12:48] Gary BC: Trust is often discussed in terms of factors like Transparency of scope, goals, users, policies and capabilities as well as Responsibility who is responsible & what are steward standards?
[12:48] Charlie Hoyt: Apologies for the mix-up between CL and CLO. I'm a big fan of both resources!
[12:49] Alexander Diehl: You are forgiven!
[12:50] Nico Matentzoglu: https://github.com/OBOFoundry/dashboard-template
[12:51] Charlie Hoyt: > DOS: Also great to see some super-editors getting credit.
It's true that super-editors are the total backbone of OBO, but I also consider that these people are the ones whose time is most thin. It's not often you can get a Chris Mungall reply on GitHub
[12:52] Charlie Hoyt: @allen the community score code is implemented here: https://github.com/cthoyt/obo-community-health/blob/a2cf266cc6d0c432d719158cb658a79d9fb8a4de/build.py#L210-L290, it's not the documented in the nicest way like the scores on the OBO Dashboard. definitely a place for improvement
[12:53] Charlie Hoyt: @nico we should externalize the schema that Anita is talking about now since it's used in several places (OBO Foundry repo, obo dashboard, also to some extent OLS)
[13:02] Anita Caron: https://anitacaron.github.io/ont-summit2023/
[13:02] Charlie Hoyt: Thanks James!
[13:03] csbjohnson: Thank you!
[13:03] RaviSharma: thanks, it was great deeper dive in OBO efforts.


== Resources ==
== Resources ==

Revision as of 16:36, 8 March 2023

Session OBO Dashboards
Duration 1 hour
Date/Time 15 Feb 2023 17:00 GMT
9:00am PST/12:00pm EST
5:00pm GMT/6:00pm CST
Convener Gary Berg-Cross

Ontology Summit 2023 OBO Dashboards

Helping scientific researchers make better use of ontologies

Agenda

Conference Call Information

  • Date: Wednesday, 15 Feb 2023
  • Start Time: 9:00am PST / 12:00pm EST / 6:00pm CST / 5:00pm GMT / 1700 UTC
  • Expected Call Duration: 1 hour
  • Video Conference URL
    • Conference ID: 837 8041 8377
    • Passcode: 323309
  • Chat Room

The unabbreviated URLs are:

Attendees

Discussion

[11:57] RaviSharma: welcome everyone

[12:06] James Overton: Here's a link to the slides: https://docs.google.com/presentation/d/1oKxBzSCZGuFlIhw6jAXw00xgEGs7KgmaVR54NEdPprw/edit#slide=id.g20803e73efb_0_85

[12:09] Charlie Hoyt: not just prefixes but also CURIEs and IRIs

[12:11] RaviSharma: James and Nico - shared design patterns - are these used from a design repository so when people develop ontologies, they can use the design templates?

[12:14] James Overton: Ravi: Many OBO projects use their own templates and patterns, but we are only beginning to use shared patterns. This is a key part of the COB effort.

[12:18] Charlie Hoyt: https://dashboard.obofoundry.org/

[12:18] Charlie Hoyt: actual dashboard: obofoundry

[12:18] Charlie Hoyt: https://dashboard.obofoundry.org/dashboard/index.html

[12:25] RaviSharma: I found a neat cross reference as to which other ontologies use a particular ontology that you are referencing.

[12:25] DOS: Very granular cell ontologies may define classes using reference data rather than textual definitions. It can therefore be impossible to fulfill the unique text def requirement. At best one can add a text definition that described the general class => non-uniqueness. Because of this, the most detailed cell type ontology on the Foundry - FBbt - fails this test. Would it be possible to modify the test to allow it to be fulfilled by links to reference data?

[12:26] RaviSharma: How do you determine content accuracy by engaging the experts like doctors or nurses before you determine its readiness for practicing users operational maturity!

[12:27] Charlie Hoyt: i am not familiar with the mondo custom OBO dashboard! This is so cool! For sure we need more strict requirements than what is currently in OBO Foundry

[12:30] James Overton: DOS: Good point. The automated check for textual definitions is more strict than then principle, which says "most" terms should be defined. <https://obofoundry.org/principles/fp-006-textual-definitions.html>, we can change the automated check, but we need to make a decision about whether to err on the side of strictness or leniency.

[12:30] James Overton: The principles were not written with automated checks in mind, and some are harder to automate than others.

[12:33] DOS: Will promotion to the top tier require passing all tests, or will there be some attempt to order based on how many tests are failing?

[12:34] James Overton: My opinion is that better dashboard score should rank higher, but there's no decision yet.

[12:35] James Overton: OQUAT: https://cthoyt.github.io/oquat

[12:35] Nico Matentzoglu: Presentation link: https://docs.google.com/presentation/d/1oKxBzSCZGuFlIhw6jAXw00xgEGs7KgmaVR54NEdPprw/edit#slide=id.g1dcecf06728_0_7

[12:37] DOS: @James -I'd like to explore discuss how we might extend the principle to allow for a combination of free text def + formalised reference data links to be allowed. Without this, I feel like FBbt is being penalised for being so detailed - where less detailed and so less useful ontology would pass.

[12:38] Nico Matentzoglu: CL is the cell ontology, CLO is the cell line ontology

[12:39] Alexander Diehl: Jonathan Bard, one of the originators of the Cell Ontology, in case anyone is curious (JB:JB)

[12:39] James Overton: DOS: I agree in general. I work on MHC Restriction Ontology (MRO) which is mostly complexes of proteins, where the logical definition says everything important. We could auto-generate textual definitions, but would that really MRO better? So I would like to see OBO refine the textual definition check.

[12:40] RaviSharma: So how did you find resolution of prefix issues?

[12:40] James Overton: https://github.com/cthoyt/obo-community-health

[12:41] James Overton: ^^ I would like to see OBO refine both the textual definition principle and the automated check.

[12:42] Nicole Vasilevsky: For the OBO Community Health, is this up to date? I pushed to Mondo yesterday and it says the last push was in September

[12:42] RaviSharma: I am surprised that Envo is not being updated since August last year, does that mean that ENVO is stabilized or that development is slow?

[12:43] James Overton: Nicole: I see "2023-02-15" as the update date for MONDO.

[12:43] DOS: I love the Community Health dashboard.

[12:43] Nicole Vasilevsky: Oh, I see it was probably an old screenshot! Ignore me

[12:44] RaviSharma: Is quality score or effort a metadata?

[12:44] Gary BC: @Ravi I think that ENVO was updated 2 days ago.

[12:45] DOS: Great to be able to point potential users to this so they can tell if an ontology they want to use is active.

[12:45] DOS: Also great to see some super-editors getting credit.

[12:44] Nico Matentzoglu: Just to make it clear: Charlie’s view on “trust” is his own :P not the OBO foundry view!

[12:45] Nico Matentzoglu: (On a personal level I agree with much of it)

[12:46] Nico Matentzoglu: Anita you can start sharing

[12:46] Allen Baron: How is the OBO community health score calculated?

[12:46] Asiyah: Some experts have significant technical barriers to use GitHub. The Geeks here should think of a way to help them!

Reply by Nico: I think sending people to special training is key. Organisations use tools like JIRA - GitHub is not too different then these. I would start by not calling it Geeky - I would think that using project management tools is pretty much a must in the modern era, in GitHub is not much different. I would recommend organising training in your organisations to raise awareness!

[12:47] Allen Baron: Nice presentation Charlie.

[12:48] Nico Matentzoglu: Instruction: https://oboacademy.github.io/obook/howto/deploy-custom-obo-dashboard/

Example custom dashboard: https://monarch-initiative.github.io/monarch-ontology-dashboard/dashboard/index.html

[12:50] Nico Matentzoglu: https://github.com/OBOFoundry/dashboard-template

[12:48] RaviSharma: Gary thanks for reply but I guess it was an old screenshot

[12:48] Gary BC: Trust is often discussed in terms of factors like Transparency of scope, goals, users, policies and capabilities as well as Responsibility who is responsible & what are steward standards?

[12:48] Charlie Hoyt: Apologies for the mix-up between CL and CLO. I'm a big fan of both resources!

[12:49] Alexander Diehl: You are forgiven!

[12:50] Nico Matentzoglu: https://github.com/OBOFoundry/dashboard-template

[12:51] Charlie Hoyt: > DOS: Also great to see some super-editors getting credit.

It's true that super-editors are the total backbone of OBO, but I also consider that these people are the ones whose time is most thin. It's not often you can get a Chris Mungall reply on GitHub

[12:52] Charlie Hoyt: @allen the community score code is implemented here: https://github.com/cthoyt/obo-community-health/blob/a2cf266cc6d0c432d719158cb658a79d9fb8a4de/build.py#L210-L290, it's not the documented in the nicest way like the scores on the OBO Dashboard. definitely a place for improvement

[12:53] Charlie Hoyt: @nico we should externalize the schema that Anita is talking about now since it's used in several places (OBO Foundry repo, obo dashboard, also to some extent OLS)

[13:02] Anita Caron: https://anitacaron.github.io/ont-summit2023/

[13:02] Charlie Hoyt: Thanks James!

[13:03] csbjohnson: Thank you!

[13:03] RaviSharma: thanks, it was great deeper dive in OBO efforts.

Resources

Previous Meetings

 Session
ConferenceCall 2023 02 08ODK
ConferenceCall 2023 02 01ROBOT
ConferenceCall 2023 01 25COB
... further results

Next Meetings

 Session
ConferenceCall 2023 02 22Ubergraph
ConferenceCall 2023 03 01Panel
ConferenceCall 2023 03 08Modules and Patterns
... further results
Retrieved from "https://ontologforum.com/index.php?title=ConferenceCall_2023_02_15&oldid=4385"