Jan 21st, 2016 Ontology Working Group Meeting

Planteome Ontology WG Zoom Meeting

• Date: Thursday Jan. 21st, 2016
• Time: 8:15am PST (GMT-8)
• Connection details: Join from PC, Mac, Linux, iOS or Android: Join from PC, Mac, Linux, iOS or Android: https://zoom.us/j/757453328

• Attendees: LC, AM, JE, PJ, EA, MAL, BS, GG.

File:Ontology WG Meeting 1-21-2016 video.mp4

File:Ontology WG Meeting 1-21-2016 audio.m4a

Report on PAG 2016

• strong emphasis on high throughput phenotyping, machine learning is hot right now
• data standardization, field scale measurements
• Utilization of ontologies for data annotation
• everyone wants specialized ontologies for THEIR data, not something predefined
• We need to demonstrate how ontologies are useful (as always)

Planteome Collaborators Workshop Monday Jan 11th

- http://planteome.org/outreach/workshops/PAG_2016

• Goals of the Meeting:
  • Introduce the Planteome project and launch Planteome 1.0 beta
  • Provide an update on the progress of the collaboration with the Crop Ontology to map crop-specific trait dictionaries to the Reference Trait Ontology
  • Provide a summary report and overview of the Planteome Curators Workshop, held Nov 2015 in Corvallis, OR
  • Engage Planteome collaborators and the plant genomics community by demonstrating use cases of data annotation for rice, cassava and lentils
  • Seek engagement of the plant genomics community through tools and applications to integrate into your workflow

• Who attended?

- 29 attendees (link to spreadsheet: Participants list CO-Planteome Meeting PAG 2016)

- Representatives from: CG centers (IITA, CIP, CIMMYT, IRRI) and Bioversity; INRA, IRD, LIRMM, USDA-ARS databases (SoyBase, MaizeGDB, GDR, PeanutBase, LegumeInfo), EBI, Transplant, ELIXIR Consortium, Rothamsted Research.

Comments and feedback

• People are concerned about the GAF2 format, but hopefully it will be expanding to more than 16 columns- modification of the GAF2 format
  • we will develop standards for conversion to GAF2

• Data download functionality on the browser needs to be refined
  • PJ: Suggest to have the column headings displayed- that would help people (JE open issue @ column headings)

From JE: Can't really do it as the "download" is actually a direct query to the solr and can't add stuff to it. However, there is a "Gaf chunk download" button that should actually have the gaf format, but again for technical reasons, we can't add a header line to. But, that button currently doesn't work, and a ticket has been created for it. Apparently, that button is not used much as it is broken on their end as well: https://github.com/geneontology/amigo/issues/293

• Modification and customization of the fields for download (open issue @)

• figure out what format people want to download data in, allow for download in this format
  • GeneBanks use the MCPD format (?)
  • Need a few use-cases for questions that researchers will be trying to answer

Update on Planteome browsers

Main Planteome website: http://planteome.org/

• drupal front pages and AmiGO2 browser serving ontologies and annotation data

Development site: http://dev.planteome.org/amigo

• Currently just ontology files but can also load annotations on there to test
  • loads nightly with all the ontologies
  • would be good if it was a different color or if there was a banner to alert users that it is the dev site. Done

Issues being sorted out:

• Many of our relations are not being recognized and displayed in AmiGO 2.0 (https://github.com/Planteome/planteome-ontology-browsers/issues/20)
  • only showing up- is_a, part_of and occurs_in; this is an issues for the Gene Ontology as well
  • Gene Ontology tracker: https://github.com/geneontology/amigo/issues/220

• Synonym scopes are not being displayed- all the broad, narrow, related and exact synonyms are being lumped together- (https://github.com/Planteome/planteome-ontology-browsers/issues/17)
  • Working with CM and Seth on resolving this. It may require us fixing it on our end

• Links to GitHub trackers are not functional-
  • Need to add these as dbxrefs in the AmiGO file- AM will do
  • CM has suggested going to using the url of the issue instead. LC will test, could write a script to implement this.

From MAL:

• The CO terms are showing up as subclasses, but they get confused with other things that are subclasses- We need special relationship to link the CO and TO classes
• Currently, the method_of, scale_of and variable_of relations are not being displayed- MAL will post a comment on the tracker: (https://github.com/Planteome/planteome-ontology-browsers/issues/20)
  • Perhaps they should be 'taxonomic equivalents' AND subclasses? ; perhaps you can have double-relationships??? (Suggestion from BS)

This requires further discussion

• Suggestions for improving the display:
  • Perhaps color coding in the graph
  • Changing CO_### to the correct name- e.g. CO_334:0000432 Plant Height ---> lentil plant height
  • How could this be done?

• Merged version of TO will contain all the mappings to the Crop specific ontologies, so they will not have to be loaded individually (From PJ)
• The breeders and data managers from the CG centers will maintain their Trait Dictionaries (TDs), which should include the mappings to the TO
• MAL's script will convert from the TD to the ontology file
• We should work on creating and testing a merged version with cassava, rice and lentil

• Visualization: need a better way to view the ontology trees
  • Choices now are OLS-vis, or the static .png display (note: OLS-Vis is currently only on the dev site)
  • OLS-Vis tool (http://ols.wordvis.com/)- pulls all the ontologies from the urls, so it is the Release version.
  • Could this be modified to display the planteome ones specifically on our browser?? Does not appear to be available for local install.

Trait vs descriptors- from EA- breeder terminology is important

• Descriptors (need to avoid this word), these are not traits
  • Usually measured on a small number of plants, and locations, with no stress
  • qualitative scale, but may be converted to ???
  • Often used for taxonomic reasons, measured at their maximum

• breeder evaluation of trait: done at multiple sites, multiple generations, include stress evaluations
  • are traits, looking for stability
  • indexed or quantitative, can be categorical scale
  • evaluation: happens at multiple times

AberOWL- Georgios V. Gkoutos

• An ontology repository and semantic search engine- link to paper on PubMed: http://www.ncbi.nlm.nih.gov/pubmed/25627673

• • semantic queries looking at relationships
• can search over all ontologies in obo
• reasoner is run, then query.
• can use SPQL

Next meetings:

• need to break up the calls into:
  • Ontology- Tuesday Feb 2nd
  • technical things (mappings and such)??
  • Browser/annotation ??
  • Planning

Upcoming meetings and workshops

• Phenotype RCN meeting Feb 2016
• GARNet/Egenis Workshop: Integrating Large Data into Plant Science: From Big Data to Discovery

April 21st-22nd 2016 (Link)

• Meeting in Montpellier, May 2016
• ICBO 2016 Corvallis