Ontology meeting 2015-02-19

Attendees: DavidOS, Paola, Harold, Heiko, Tanya, Chris, DavidH

Minutes: Tanya

Regulation Chains

  David H went over Nikolai's results chaining regulates relations and introducing direct and indirect regulation. Presented the 'algebra' of
  the chains. Showed the inferred classification that the chains provided.
  David H will introduce his property chains into David OS's branch file, run assert links report and see
  what happens with terms in the development and apoptosis. The branch file pulls in live everything else except live GO.
  Chris concerned about aligning what Nikolai did with the current RO. Note that development part of GO doesn't have many ends_with details.  
  Find a specific developmental process that can have molecular functions attached to it and use that for modeling/test.
  Chris - test using Noctua? This brings in live RO.  Manually redo Nikolai's example in Noctua and see what happens with live RO in place. 
  Write this up (Nikolai's work) and submit to PeerJ or BioRxiv for giving him credit for his work.

- Related: Inference resulting from implementing regulation chain reasoning results_in relation - how does this fit? Currently no axiomatisation linking this to regulates. Shouldn't there be?

  Re: results_in . Need to go back and redo this.  Has come up in some SF items. History: bunch of related terms with a certain pattern
  This relationship allowed making logical definitions for them and specify the directionality.  Why not use 'regulates'?  
  Because of indirectness, upstream nature. Better to
  use 'indirectly regulates'?  Consider folding 'results_in' into this work.  Replace with what? Right now, logical definition: results_in 
  xxx regulation term, unfold to the generic regulates xxx bp term or indirectly regulates.

  EXAMPLE:  
  id: GO:2000531 
  name: regulation of fatty acid biosynthetic process by regulation of transcription from RNA polymerase II promoter 
  namespace: biological_process def: "Any process that modulates the frequency, rate or extent of the biosynthesis of fatty acids, 
  by modulating the frequency, rate or extent of transcription from an RNA polymerase II promoter." [GOC:vw] 
  is_a: GO:0042304 {is_inferred="true"} ! regulation of fatty acid biosynthetic process 
  is_a: GO:0072364 {is_inferred="true"} ! regulation of cellular ketone metabolic process by regulation 
  of transcription from RNA polymerase II promoter    
  intersection_of: GO:0006357 ! regulation of transcription from RNA polymerase II promoter 
  intersection_of: results_in GO:0042304 ! regulation of fatty acid biosynthetic process

  Desired:  regulates some (TF activity and regulates some FA biosynth) OR: directly_regulates some (TF activity and regulates some FA biosynth)
  Name: "TF activity and regulates some FA biosynth"  OR "TF activity that regulates some FA biosynth"
  Force genus to be direct and differentia to be indirect?
  Don't know how direct the regulation of tsc (in example above).  Knocked out a TF and seen fatty acid biosynthesis go down.
  Proposal: new class here TF activity that regulates some FA biosynth, existing term GO:2000531 gets defined this way 
  Problem: need to teach people rules in using these precise terms, no different from current situation, huge possibility for proliferation of
  terms
  None of these terms may be relevant if we can chain together annotations using Noctua in the lego model. 
  How many?  less than 200.  8 added in 2014. Most from 2012.
  Chris could do experiment with the expansion and show people the results.  Sandbox?

PUNT: Issues With Ion Channels

Right now the ion-specific channels are not defined as being selective for a type of ion. Therefore, non-selective channels are being annotated to the ion specified terms. In the long run this is misleading and is not intuitive to electrophysiologists.

Punt to next week, Huaiyu wants to be on call.

Disjoints in ChEBI

Background from last week: David OS to report back on today's meeting with Janna.

We will set up a Jenkins job to send to them about disjoints. We can make a branch file with high-level disjoints that we can edit ourselves and test. We can then feed this to them. Perhaps we can set up a Jenkins job to run Chebi and find violations that will become obvious to address. Janna says that current disjoint violations are being fixed. Who will follow up on any future ones? Not clear that there will be an ontology person after Janna is gone. Collaboration on roles? Need a contact person there to touch base with from time to time.

DavidOS to flesh out the rest of the update, please.

PUNT: How to define protein families in enzyme binding terms

See background here: http://wiki.geneontology.org/index.php/Ontology_meeting_2015-02-12#Enzyme_Binding