Are multi-organism process and cellular process disjoint?

From GO Wiki
Jump to: navigation, search

Proposal 1 - Multi-organism process and cellular process should no longer be disjoint

From examination of the taxon inconsistency file it would seem that these two processes are not disjoint in principle, though they currently are recorded as being disjoint in GO.

  • According to current thinking, the cellular process term covers processes occurring at the cellular level in a single organism.
  • Multi-organism process covers processes that involve more than one organism.
  • These two processes are considered disjoint (mutually exclusive), and so cannot have is_a children in common.

So the question is, where do we put processes ocurring at the cellular level, and including multiple organisms? Examples are 'receptor mediated endocytosis of virus by host' and 'viral transcription'. The process 'receptor mediated endocytosis of virus by host' clearly happens at the cellular level but requires the participation of viral gene products, and so is a cellular multi-organism process. We cannot currently make the term 'cellular multi-organism process' because it would be an is_a child of both multi-organism process and cellular process. This is not allowed, as the two parent terms are disjoint.

This stucture is not permitted as the two parent terms are disjoint:

[i]cellular process
---[i]cellular multi-organism process
[i]multi-organism process
---[i]cellular multi-organism process

Somehow we have to decide whether to place 'receptor mediated endocytosis of virus by host' under cellular process or multi-organism process, but it doesn't make sense to choose, as it is clearly happening at the cellular level, and involves two organisms.

Currently we have terms that should be descendents of cellular multi-organism process placed as descendents of cellular process, and they have virus annotations. As we start to make more virus annotations, it is inevitable that there will be more terms like this. Such terms are immediately flagged up if a taxon rule is made as follows:

[Term]
id: GO:0009987
name: cellular process
relationship: only_in_taxon NCBITaxon:131567 ! cellular organisms

(Cellular organisms excludes viruses.)

It would be helpful to resolve this problem at some point, but this can only be done if the multi-organism process and cellular process are no longer stipulated to be disjoint from one another.

Proposal 1

I would like to propose that the disjoint relationship between cellular process and multi-organism process be removed, and a child term cellular multi-organism process be created.

Proposal 2 - Direct or indirect viral annotations to cellular process should be allowed

Following on from the proposal above, if a viral annotation was made to a term such as 'receptor mediated endocytosis of virus by host' and then slimmed up to cellular process, then the annotation would be wrong according to current thinking. (Current GO community understanding is that cellular processes can only be carried out by gene products of cellular organisms, and viruses are not cellular organisms.)

Proposal 2

I would like to propose that the term cellular process should not be restricted to use only for annotation of gene products of cellular organisms, either in the taxon checking file or in the definition or established understanding of the term. With or without the acceptance of proposal 1, cellular process will still be an ancestor of terms such as 'receptor mediated endocytosis of virus by host' and so it will be as well that annotations should be able to be slimmed up to cellular process without being incorrect.

Comments

Midori: "I think it would help Val quite a bit with the yeast mating (i.e. conjugation with cellular fusion)."

Jane: I think we do need to keep multi-organism processes disjoint with single-organism processes - at the moment cellular processes are always single organism so there isn't a problem. Otherwise I think we'll end up with oddities in slims etc. In the example you give, genes involved in multi-organism processes would end up slimmed to a cellular process parent - that might be very confusing.

However, to keep them disjoint would lead to this sort of structure:

single-organism process
---[i] single-organism cellular process
---[i] single-organism multicellular process
multi-organism process
---[i] multi-organism cellular process
---[i] multi-organism multicellular process
multicellular process
---[i] multi-organism multicellular process
---[i] single-organism multicellular process
cellular process
---[i] single-organism cellular process
---[i] multi-organism cellular process

which I'm not happy with either. I think this warrants more thought. I do see the problem though.

Thread carries on on the go list:

Re: [Go] disjoint proposal
28/01/2010

Concludes that in fact 'multi-organism process' was really intended to be 'multi-organism process occuring at the cellular, tissue, organ or organism level' whereas 'cellular process' was intended to be 'single organism process occurring at the cellular level'.

This means that we need to come back to these terms and properly define them and then move the multi-organism processes out of under 'cellular process'.

Later, others disagree and it is proposed to take this to the consortium meeting.