Ontology meeting 2021-06-28

From GO Wiki
Jump to navigation Jump to search
  • Group members: Pascale, Karen, Harold, Raymond, Peter, Jim, Tanya, Kimberly, David, Chris, Paul
  • Present: Pascale, Karen, Harold, Raymond, Jim, Tanya, Kimberly, David, Chris
  • Regrets: Paul


Agenda

Announcements

  • Rhea-GO alignment jamboree- week of July 26th
  • No call next week - 4th of July observance day
  • No call the week after - No GO calls week

Taxon constraints

Can now add references supporting the TC, see https://github.com/geneontology/go-ontology/blob/master/src/taxon_constraints/never_in_taxon.tsv

Discussion points

Xenobiotic to multispecificity transporter

Karen

-> Will create a ABC-type multispecificity transporter and a 'channel type' (exact terms labels TBC)

Preparing RHEA jamboree

NAD(P) issues

  • Some enzymes use either NAD or NADP, at least in vitro. RHEA creates 2 reactions for these, but this is not the 'interesting' part of the reaction, https://github.com/geneontology/go-ontology/issues/19648
  • EC also sometimes creates a grouping class, NAD(P)+ - but this means that the enzyme can use BOTH, so it cannot be a parent class to the NAD/NADP terms in GO as we currently do (see Rule 18: "For oxidoreductases using NAD+ or NADP+, the coenzyme should always be named as the acceptor except for the special case of Section 1.6 (enzymes whose normal physiological function is regarded as reoxidation of the reduced coenzyme). Where the enzyme can use either coenzyme, this should be indicated by writing NAD(P)+.)"
  • GO often (but not always) follows that pattern. The 'NAD(P)+' term is used to mean either, "the enzyme can use both", or "we don't know which H donor is used", which leads to annotation inconsistencies.


Kristian 2021-06-25 (skype with Pascale - asking how many enzymes this may affect). His answer: 

 Many. There are currently 170 EC numbers for enzymes that accept both NAD+ and NADP+. In addition we can have EC numbers for enzymes that are specific for NAD+ and NADP+, respectively. This can be illustrated by Alcohol dehydrogenase:
* EC 1.1.1.1 Alcohol dehydrogenase (NAD+)
* EC 1.1.1.2 Alcohol dehydrogenase (NADP+)
* EC 1.1.1.71 Alcohol dehydrogenase (NAD(P)+)

 It is always a problem to find all the substrates of an enzyme, it requires that the scientists look for them. Co-enzymes (like NAD and NADP) are often only described in the assay mixtures, and for enzymes like P450 not at all since they are characterised using microsomes, that are purified membrane fractions.
  • Proposal:
    • Always use NAD(P), regardless, and define it as 'OR'
    • Map to both RHEA (NAD & NADP) as NarrowMatch
    • ECs would probably also be NarrowMatch
    • Generate definitions for these that include "NAD(P)"
    • Add a curation comment to say that this means the enzyme may use either NAD, NADP, or both. Annotation to this term doesn't imply one or the other.


=> agree

GO/Rhea scopes

See https://github.com/geneontology/go-ontology/issues/20114

Anne writes "@pgaudet, sorry but we will not create the parent reaction with Ser/Thr residues. This not the scope of Rhea. Similarly we don't create reactions with NAD(H) nor NADP(H). As mentioned earlier by @alanbridge, UniProt entries linked to EC 2.7.11.1 are annotated with RHEA:17989 ATP + L-seryl-[protein] = ADP + H+ + O-phospho-L-seryl-[protein] RHEA:46608 ATP + L-threonyl-[protein] = ADP + H+ + O-phospho-L-threonyl-[protein]

Note that these 2 reactions are linked to several EC numbers (see https://www.rhea-db.org/rhea/?query=ec%3A2.7.11.1&columns=%2Crhea-id%2Cequation%2Cenzyme-class%2Cuniprot%2Cgo).

This is typically a case where UniProt/Rhea have different annotation guidelines compared to GO. In these cases the 1:1 mapping will not be possible."

=> agree to obsolete terms as these are outside the scope of GO

  • Paul suggests not making RHEA cross-references for anything involving proteins - proposes to leave it out of scope for now. Could revisit later. Maybe this also applies to generic DNA and RNA.


Query lobed nucleus

https://github.com/geneontology/go-ontology/issues/20764

Was obsoleted (with no annotations) The reason for the obsoletion is that this represents the nucleus of specific cell types (basophils, eosinophils and neutrophils). The lobed nuclear appear to be due to differently packaged chromatin, see https://sciencing.com/lobes-nucleus-20628.html, they do not represent different types of nucleus.


David OS notes that this causes problems to define some cell types; does that make it a valid GO term? Note that we had the same problem with 'cytokine secretion' - Alex Diehl argued to keep these to define some leukocyte types.

=> Chris created a ticket to [create a script to test if an ontology class is used by another ontology, to be used prior to obsoleting https://github.com/geneontology/go-ontology/issues/21723]

Other Business

https://github.com/geneontology/go-ontology/projects/1

Retrieved from "https://wiki.geneontology.org/index.php?title=Ontology_meeting_2021-06-28&oldid=80567"