Annotation Conf. Call, February 11, 2014

From GO Wiki
Jump to: navigation, search

Agenda

Present:
SGD: Rama, Diane, Paul
MGI:David, Li
EBI- Paola, Rachael, Pablo
WB- Kimberley
Dicty- Petra, Pascale
Berkeley: Moni
UCL: Ruth, Paul and Ana


Review of apoptosis related genes (Pascale)

http://wiki.geneontology.org/index.php/PAINT_-_Apoptosis

QA checks in UniProt and Jenkins (Rachael)

There are checks that are not included in Jenkins but that UniProt-GOA does have checks for.

Jenkins checks for obsolete relations and checks GO terms in annotation extensions, but does it check for appropriate use of relations, e.g. if an entity type is out of range? Does it check for orphan ISS?

If Jenkins can't incorporate these checks, are groups interested in UniProt sending each group an error log for their annotations?

For example, the PomBase Jenkins check is clean; http://build.berkeleybop.org/job/gaf-check-pombase/lastSuccessfulBuild/artifact/gaf-validation-report.txt

but the UniProt-GOA check brings up several errors in 'domain/range';

e.g.
[[GO:0005829] [happens_during(GO:0034605)] [Invalid domain for happens_during: GO:0005829]]
[[GO:1990188] [not_dependent_on(PomBase:SPCC736.11),occurs_at(SO:0001843)] [Invalid range for occurs_at: SO:0001843]]
[[GO:1990188] [not_dependent_on(PomBase:SPCC736.11),occurs_at(SO:0001843)] [Invalid range for occurs_at: SO:0001843]]

ACTION:

We will add rules to GO rule xml file

Coming soon

  • Feb 24th: David OS will talk about formalizing the representation of symbiotic relationships and annotations to describe them in the ontology
    • GO annotations include many that involve an interaction between two organisms. These include:

annotation to a cell component of the host of the organisms expressing the gene being annotated. annotation to a biological process involving a virus and and its host.

In these cases we have a mechanism for recording multiple taxon IDs for interacting organisms, but not the relationship between the two organisms , e.g one might be host to the other or vector for the other. We also don't have a way to record which of the two interacting organisms a cell component is located in or a biological process is occurring in.

My proposal aims to solve this problem, by allowing annotators to link taxon IDs using a relationship such as host_of or vector_for. Additionally you will be able to use annotation extension to record which of the two organisms an annotated CC is located in or a biological process occurs in.

Along with this extension to annotation, I plan to rearrange parts of the CC hierarchy - particularly those relating to host cell. My presentation will include details of these changes.

  • TBA: Sandra Orchard will talk about the IntAct Complex portal

Discussion

Apoptosis

  • lot of annotations moved upstream. Assay read out and the process involved are different. Paola has added comments to terms where applicable to clarify.
1) A term with detailed annotation guidance and general annotation examples:
GO:0097527 necroptotic signaling pathway
(http://amigo2.berkeleybop.org/cgi-bin/amigo2/amigo/term/GO:0097527)

2) A term with specific annotation examples, including a PMID reference and protein identifiers:
GO:0097345 mitochondrial outer membrane permeabilization
(http://amigo2.berkeleybop.org/cgi-bin/amigo2/amigo/term/GO:0097345)

3) A term with annotation guidance on the dangers of mistaking an assay readout for a true process:
GO:0006309 apoptotic DNA fragmentation
(http://amigo2.berkeleybop.org/cgi-bin/amigo2/amigo/term/GO:0006309)
  • Apoptosis annotation guide is available on the wiki - http://wiki.geneontology.org/index.php/Apoptosis_Curation_Manual
  • Ruth has started a table cataloging some information. She will circulate it when ready.
  • We can now store annotation example in the ontology itself. When you are requesting a term, please try to include an example