cellular processes missing cellular process parent

[ValWood]

I spotted

GO:0071840 cellular component organization or biogenesis
GO:0055085 transmembrane transport]have both had "cellular process" parent deleted.
GO:0016192 vesicle-mediated transport
GO:0006913 nucleocytoplasmic transport
GO:0007034 vacuolar transport

lost "cellular process parent"
Any reason for that?

GO:0097428 protein maturation by iron-sulfur cluster transfer
(but this never had it)

val

[ValWood]

Note that the definition of cellular process is:

Any process that is carried out at the cellular level, but not necessarily restricted to a single cell. For example, cell communication occurs among more than one cell, but occurs at the cellular level.

[ukemi]

Is it really worth distinguishing things 'at the cellular level' in GO?

[ValWood]

It might not be but we always have. Its probably more useful for multicellular eukaryotes?
In yeast everything is at the cellular level...

[ValWood]

"cellular process" It is in the pir and plant slims, so might be worth checking there (I thought it was in the generic slim, but it isn't....)

[ukemi]

With LEGO, we can capture the location of where functions and processes are occurring in a much better way. It doesn't have to be fuzzy 'at the level of'.

[ValWood]

What is 'it'?
"cellular process"

It doesn't have to be fuzzy 'at the level of'.
true....

Is there an easy way at present to distinguish all cell level things from the tissue level/behaviour/homeostasis of number of cells/ system process etc?
(it will be a long time before you can do that with LEGO derived annotation if there is not another way...)

[dosumis]

The transport one might have been my edits while cleaning up to improve inference. The ideal way to fix it would be to have a term for 'cellular transport' to use as the genus. Unfortunately we don't have this term. We could assert 'cellular process' as a parent, but the logical definition:

transport
and (mediated_by some vesicle)

may result in classifications that contradict that in future. For example, transport in exosomes between cells is vesicle mediated, but is this cellular (?)

Is there an easy way at present to distinguish all cell level things from the tissue level/behaviour/homeostasis of number of cells/ system process etc?

How about: A process occurring in or involving exactly one cell?

But note: cellular component organization includes organization of extracellular structures, which doesn't fit this def.

[ValWood]

Cell level processes shouldn't exclude "organization of extracellular structures".

For example, transport in exosomes between cells is vesicle mediated, but is this cellular (?)

We (pombase curators), are thinking that all things mediated by a single cell are cellular . However, we are now tripping up over "conjugation" in fungi, which is mediated by more than one cell, but we still think of as cellular, so this is not "sufficient". Unless of course we accept that conjugation is not cellular....?

[ValWood]

Some parts of conjugation are mediated by a single cell (shmooing), and some parts (cytogamy) are not cellular...so perhaps that would be fine. The parts which are cellular would have cellular process parents.

[dosumis]

"Cell level processes shouldn't exclude "organization of extracellular structures."

Then I think we're stuck with the existing def (and lots of tickets arguing over it...)

If we agree that everything that occurs in a cell is cellular, we could at least automate some of this by adding the axiom:

biological_process that occurs_in some cell SubClassOf: 'cellular process'

@cmungall Do we have a place where GCIs like this can live?

[cmungall]

In this case, it wouldn't go to the header, it would look like this:

id: GO:0008150
name: biological_process
is_a: GO:0009987 {gci_relation="occurs_in", gci_filler="GO:0005623"} ! cellular process

http://owlcollab.github.io/oboformat/doc/obo-syntax.html#5.2.2

This will look confusing in OE, as OE doesn't take into account qualifiers in making the hierarchy

We'd want to filter this axiom from the public obo releases.

Alternatively, we could PATO-shunt this to a cellular quality. The GCI would then live in pato_ext. Ad we'd have an equiv axiom on cellular process.

But overall I think we need to rethink the whole cellular-X thing

[cmungall]

So as not to divert this ticket I started a document with some thoughts about problems with cellular-X terms: https://docs.google.com/document/d/1QpfUY_LgeIryMj6EEAE05FXLE_894GalkerBU8dMuVU/edit

this is in the GO drive, in the ontology folder. I like keeping things in the tracker but google docs can be useful for things like this

[tberardini]

Is there any action that I could/should take on the particular terms that Val has brought to our attention?

Two suggestions from @dosumis:

1. The ideal way to fix it would be to have a term for 'cellular transport' to use as the genus. Unfortunately we don't have this term. We could assert 'cellular process' as a parent, but the logical definition [of 'vesicle transport']:

transport
and (mediated_by some vesicle)

may result in classifications that contradict that in future. For example, transport in exosomes between cells is vesicle mediated, but is this cellular (?)

2. If we agree that everything that occurs in a cell is cellular, we could at least automate some of this by adding the axiom:

biological_process that occurs_in some cell SubClassOf: 'cellular process'

(see Chris' response above to why that might be problematic)

[tberardini]

Linking to the bigger issue #12849

[ValWood]

Even though this might be a "bigger issue" it looks very odd to a biologist that none of these terms are is_a "cellular process"

Could these links be instantiated and this ticket closed pending the "bigger issue" #12849 ticket?

Note, this ticket was opened in 2016 so the "longer-term solution" does not seem to be imminent.

[ValWood]

Don't the logical definitions sort situations like this? It's quite 'high-level" stuff.

[pgaudet]

Changed subclasses:

• GO:0071840 cellular component organization or biogenesis
  from biological_process to 'cellular process'

• GO:0055085 transmembrane transport]have both had "cellular process" parent deleted.
  Added subclass 'cellular process'

• GO:0016192 vesicle-mediated transport
  Added subclass 'cellular process'

• GO:0006913 nucleocytoplasmic transport
  Added subclass 'cellular process' to (great-grand) parent 'cellular localization'

• GO:0007034 vacuolar transport
  Added subclass 'cellular process'

[pgaudet]

Also added 'cellular process' subclass to:

• GO:0007155 cell adhesion

• GO:0098743 cell aggregation *

• GO:0001906 cell killing *

• GO:0008283 cell population proliferation *

• GO:0051674 localization of cell

• GO:0061842 microtubule organizing center localization

• *were direct children of biological_process