Sept 29th, 2015 Ontology Working Group Meeting
(Redirected from Ontology Working Group Meeting, Sept 29th, 2015)
Who: participants: AM, LC, MAL, EA, CM, PJ, JE
- Link to recording- Video: File:Ontology WG 9-29-15 video.mp4
- Link to recording- Audio: File:Ontology WG 9-29-15 audio.m4a
1. Training workshop in Corvallis
- Location: Corvallis, 2nd week of Nov
- notes are on the wiki Planteome_Curator_workshop_Fall_2015 - dates, accommodations
- PJ said whomever is on the Planteome- has travel funds, and should go ahead and use them. We will book the others (CIP, etc...)
- 2 curators from CIGAR coming
- EA: Could check with ICARDA and ICRISAT curators, if there is money and they are interested to come
- Suggested: Karthika Rajendra - ICARDA, Morocco might come to the training. Worked on the lentil trait dictionary and was interested in getting trained. (She was apparently at the April workshop (not sure which one this is. Not at the Crop Ontology 2014 meeting in Montpellier).
- RS Hamilton- IRRI, very busy as he is also working at Rome office, but will be at the PAG meeting
- CM will be there all week, so we can have the Protege training at any time in the week. Seth will come for a couple of days at the end of the week.
2. TermGenie Demo
- TermGenie [1] is a tool for creating new ontology terms for either GO (Gene Ontology) or OBA (Ontology of Biological Attributes), as well as others. It uses a pattern-based approach to rapidly generate new terms and place them appropriately within the ontology structure. All terms are reviewed by a senior editor before the final commit to the ontology.
Link to TermGenie GitHub
- Two instances of TermGenie exist: for OBA and GO:
- CM
- log-in (anyone can sign up)- Firefox
- Select template (currently entity-quality)
- Enter the entity (pick it from the GO), the quality (from PATO), enter optional lit_ref, and DBXref
It checks for duplicates, or existing terms Gives the label (add "trait") to the end. Also give a definition it infers the relationships, need to check them Don't use the template it give you (shown in red under the name) select the little check-box, and click submit - Now it waits for a "gate-keeper" to accept the term from the queue - Gate-keeper reviews it (can edit), and performs the commit, once accepted it will update the ontology on GitHub
Works in GO
- You can jump to different ontologies, select the upper level term you want it under, and hit "use template" (only works if there is an upper level term that your term conforms to.)
Free form term creation
allows you to define all aspects of the new term using a free-template.
enter all the metadata, and submit.
There is a double-check that terms are not being created as duplicates. Doesn't have access to OWL, so it can't check duplicates of biological significance.
Doesn't exist for the TO, EO, and PSO. (we will have to develop this for them)
can run it on any server that we need it to be on.
create a yaml-file on git, that has metadata on all the users, and their permissions.
Don't know yet if we need separate instances for each different ontologies, or if we can use one 'master' for the whole Planteome
Caveats
inferences work proportionally to how well the logical definitions are in the ontology.
TermGenie really only works if the hierarchy follows the PATO
if you want the reasoner to work, you have to trust the primitive ontologies
can compromise, and manually enter terms that don't fit with our interpretation
TermGenie requires equivalence axioms to be in the same file.
3. Discussion About Plant Trait Ontology:
"Morphology" needs 'discussion'
need more templates to accommodate more complex traits/term requests.
Do we switch to 100% Protege? Or can we still use both.
We are pushing for a full-switch to Protege.
Training will focus on Protege only for the workshop.
Keep source files in OBO for instances where we need obo-edit.
Need to use the most recent Protege (5.0 beta snapshot) to save the source as .OBO
Link from Chris: PATO:plant_trait_xp