Annotation Conf. Call 2019-09-24: Difference between revisions

From GO Wiki
Jump to navigation Jump to search
mNo edit summary
 
(29 intermediate revisions by 2 users not shown)
Line 12: Line 12:


== GO-CAMs and Annotation Extensions ==
== GO-CAMs and Annotation Extensions ==
* After USC hackathon in June, we formed a small working group to fully articulate the [https://github.com/geneontology/go-shapes/blob/master/shapes/go-cam-shapes.shex GO-CAM model specifications]
* Update on relations allowed for GO-CAM as specified in ShEx (Shape Expressions)
* This will allow for consistent GO-CAM curation using Noctua workbenches, import projects (e.g. Reactome and MOD imports)
* Approach:
* Will also allow us to move forward with harmonizing AE relations used in GO-CAM and conventional annotation
** Examine relations used in GO
** On today's call, we will review the annotation extension relations that need further discussion, starting with happens_during
*** Annotations extensions
** [https://docs.google.com/spreadsheets/d/1zB8UKtG3JRGMkHgXmMpby-K5RFmYxGGRMEXULCCtcPo/edit#gid=1539887761 Google summary spreadsheet of MF extensions]
*** Ontology
** Some relations are in the Relations Ontology (RO), others have just been in gorel (relations only used in the Gene Ontology)
** Trying to move towards using only RO relations, using a smaller set of relations to make curation easier (but still maintain accurate representation of the biology)
*Relations Usage
**Definitely in
**In, but usage restricted to specific ontology terms and their children
**Out for now (need further discussion)
**Definitely out


=== happens_during ===
* [https://docs.google.com/spreadsheets/d/1qoxnXM3q59jcZLIxnX5SY0H5syfK7GJc38s0VGxx2jM/edit#gid=0 Google spreadsheet for happens during ]
* Current recommendations are that 'happens during' is used to define a biological phase or life stage during which a molecular function occurs
* However, happens_during is not defined in the [https://www.ebi.ac.uk/ols/ontologies/ro/properties?iri=http%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FRO_0002092 relations ontology]
* There is a comment, however:
** X happens_during Y iff: (start(Y) before_or_simultaneous_with start(X)) AND (end(X) before_or_simultaneous_with end(Y))
** This implies that the start and end of the molecular function and phase are known
** For things like cell cycle phases or development stages this seems fairly straightforward
* But what about annotations that referred to a function that 'happens during' a biological process, such as 'response to oxidative stress'?
* What do we mean when we make these statements?
** WB had seven MF annotations with 'happens during' extensions that needed review.
*** One used a WBls term, so that was okay.
*** The remaining six used 'response to' terms and in reviewing them, I changed the extension relation to 'part of' the 'response to' term.  The MFs were all transcription-related functions that mediated the transcriptional response to unfolded protein,  or hypoxia, or heat.
** (Pascale): If we want to allow happens_during for 'response to' terms, then we should treat them as 'biological phases', and disallow direct annotations. Are there cases where using 'part_of' versus 'happens_during' makes sense ?
* Updated documentation on 'happens during' usage:  http://wiki.geneontology.org/index.php/Annotation_Extension_Relation:happens_during


== Ontology Updates and Annotation Reviews ==
* [https://docs.google.com/spreadsheets/d/1okodOSG7mWgS8ZE8T95XjPr0-9G1vEIkAy2BkipEfRw '''Biological Process''']
=== Cellular Component ===
**Definitely in:
* 'cell part' and child terms
***part_of
** creation of new term ' cellular anatomical entity'
***has_input
=== Protein multimerization terms ===
***has_output
* How are the reviews going?
***occurs_in
* Next terms up: protein homodimerization, protein heterodimerization (non-IPI evidence codes)
***transports_or_maintains_localization_of
***has_target_end_location (Reactome)
***has_target_start_location (Reactome)
***regulates
***negatively_regulates
***positively_regulates
**In but usage restricted to extending from 'activity regulating processes', e.g. transcription, translation:
***causally_upstream_of
***causally_upstream_of_negative_effect
***causally_upstream_of_positive_effect
**Out for now (need further discussion)
***results in.....
****curators would request appropriate terms for now
***others
**Definitely out
***chained relations, e.g. regulates_o_occurs_in
****annotations using these relations will be [http://noctua.geneontology.org/editor/graph/gomodel:5c4605cc00000396 modeled in GO-CAM] but the relations won't be used in the GO-CAM models themselves
***others


=== Transcription terms ===
* [https://docs.google.com/spreadsheets/d/1DuON03hEfEDDC7_6CCnnR3JcBA4ltQgGFP0otAeMZKs '''Molecular Function''']
* Any issues or questions here?
**Definitely in:
***enabled_by
***part_of
***occurs_in
***has_output
*** has_input
***regulates
***negatively_regulates
***positively_regulates
***directly_regulates
***directly_negatively_regulates
***directly_positively_regulates
***causally_upstream_of_or_within
***causally_upstream_of_or_within_negative_effect
***causally_upstream_of_or_within_positive_effect
***causally_upstream_of
***causally_upstream_of_negative_effect
***causally_upstream_of_positive_effect
***happens_during
**In but usage restricted to  specific function terms
***  directly_provides_input_for
**Out for now (need further discussion)
***occurs_at
***compound function relations
**Definitely out
***[https://docs.google.com/spreadsheets/d/1DuON03hEfEDDC7_6CCnnR3JcBA4ltQgGFP0otAeMZKs See spreadsheet]


= Minutes =
 
*Present: Colin, David, Dmitry, Dustin, Edith, Giulia, Helen, Harold, Karen, Kimberly, Laurent-Philippe, Michelle, Midori, Pascale, Paul T., Sabrina, Seth, Stacia, Tanya, Petra
* [https://docs.google.com/spreadsheets/d/17LcW5v4mX18jJjAXfXethJnnI0cTg--PZUGXtV7NWsY '''Cellular Component''']
**Definitely in:
***part_of
**In but usage restricted to specific component terms
***adjacent_to
****extracellular region
***overlaps
****cell junction-type terms
**Out for now (need further discussion)
***produced_by
**Definitely out:
***[https://docs.google.com/spreadsheets/d/17LcW5v4mX18jJjAXfXethJnnI0cTg--PZUGXtV7NWsY See spreadsheet]




== GO-CAMs and Annotation Extensions ==
*'''GO-CAM GPAD outputs'''
=== happens_during ===
**Still work to be done here
* We discussed review and use of 'happens during' as an annotation extension relation for MF terms.
***Should we limit the AEs generated and if so, how?
* Use of 'happens during' with a biological phase of life cycle stage is allowed.
***Should we try to get chained annotation extension relations back out of GO-CAM models?
* Use of 'happens during' with a 'response to x' BP term is still under review.
***How do we want to propagate 'contributes to'?
** If the MF is an integral part of the 'response to' term, then the curator should update the annotation extension relation to 'part of'.
** Other 'response to' extensions will need to be examined in more detail to better understand what information we were trying to capture and how it would be modeled in a GO-CAM
** PomBase has a number of these annotations and will get back to us on their thoughts.
* We will also address Barbara's questions about use of 'happens during' on a future call.
=== occurs_at ===
* We will not use SO terms with the 'occurs at' relation to capture the type of sequence or motif at which an MF occurs, e.g. 'occurs at' n_terminal_region (SO:0100014)


= Minutes =
*Present: David H, Dustin, Edith, Harold, Justin, Kimberly, Laurent-Philippe, Li, Michele, Midori, Pascale, Patrick M, Petra, Rob, Sabrina, Seth, Stacia, Stan, Suzi A, Tanya


[[Category:Annotation Working Group]]
[[Category:Annotation Working Group]]

Latest revision as of 11:27, 24 September 2019

Agenda

October GOC Meeting - Berkeley

Pipeline

  • There have been some glitches with the GOC pipeline.
  • New ontology terms may not be available for use yet.
  • Software team is aware of the issues and working on restoring the pipeline to its regular schedule.
  • Any updates here?

GO-CAMs and Annotation Extensions

  • Update on relations allowed for GO-CAM as specified in ShEx (Shape Expressions)
  • Approach:
    • Examine relations used in GO
      • Annotations extensions
      • Ontology
    • Some relations are in the Relations Ontology (RO), others have just been in gorel (relations only used in the Gene Ontology)
    • Trying to move towards using only RO relations, using a smaller set of relations to make curation easier (but still maintain accurate representation of the biology)
  • Relations Usage
    • Definitely in
    • In, but usage restricted to specific ontology terms and their children
    • Out for now (need further discussion)
    • Definitely out


  • Biological Process
    • Definitely in:
      • part_of
      • has_input
      • has_output
      • occurs_in
      • transports_or_maintains_localization_of
      • has_target_end_location (Reactome)
      • has_target_start_location (Reactome)
      • regulates
      • negatively_regulates
      • positively_regulates
    • In but usage restricted to extending from 'activity regulating processes', e.g. transcription, translation:
      • causally_upstream_of
      • causally_upstream_of_negative_effect
      • causally_upstream_of_positive_effect
    • Out for now (need further discussion)
      • results in.....
        • curators would request appropriate terms for now
      • others
    • Definitely out
      • chained relations, e.g. regulates_o_occurs_in
        • annotations using these relations will be modeled in GO-CAM but the relations won't be used in the GO-CAM models themselves
      • others


  • Molecular Function
    • Definitely in:
      • enabled_by
      • part_of
      • occurs_in
      • has_output
      • has_input
      • regulates
      • negatively_regulates
      • positively_regulates
      • directly_regulates
      • directly_negatively_regulates
      • directly_positively_regulates
      • causally_upstream_of_or_within
      • causally_upstream_of_or_within_negative_effect
      • causally_upstream_of_or_within_positive_effect
      • causally_upstream_of
      • causally_upstream_of_negative_effect
      • causally_upstream_of_positive_effect
      • happens_during
    • In but usage restricted to specific function terms
      • directly_provides_input_for
    • Out for now (need further discussion)
      • occurs_at
      • compound function relations
    • Definitely out


  • Cellular Component
    • Definitely in:
      • part_of
    • In but usage restricted to specific component terms
      • adjacent_to
        • extracellular region
      • overlaps
        • cell junction-type terms
    • Out for now (need further discussion)
      • produced_by
    • Definitely out:


  • GO-CAM GPAD outputs
    • Still work to be done here
      • Should we limit the AEs generated and if so, how?
      • Should we try to get chained annotation extension relations back out of GO-CAM models?
      • How do we want to propagate 'contributes to'?

Minutes

  • Present: David H, Dustin, Edith, Harold, Justin, Kimberly, Laurent-Philippe, Li, Michele, Midori, Pascale, Patrick M, Petra, Rob, Sabrina, Seth, Stacia, Stan, Suzi A, Tanya