cell motility query parentage

[gocentral]

Hi

I am not convinced that GO:0048870 cell motility: Any process involved in the controlled self-propelled movement of a cell that results in translocation of the cell from one place to another. should be a child of GO:0006928 cellular component movement: The directed, self-propelled movement of a cellular component without the involvement of an external agent such as a transporter or a pore.

I can see that in order for a cell to move cellular components will need to move but it just seems odd.

Ruth

Reported by: RLovering

Original Ticket: geneontology/ontology-requests/11239

[gocentral]

Hi,

Because 'cell' is_a 'cellular component', 'cell motility' is_a 'cellular component movement'.

Original comment by: tberardini

[gocentral]

• assigned_to: Tanya Berardini

Original comment by: tberardini

[gocentral]

This is probably beyond my intellectual capacities...... but why is cell a cell component ? Cell component is the root; perhaps the name does not correctly represent all classes that it contains ?

Original comment by: pgaudet

[gocentral]

I don't know if we want to go into the whole history of 'cell component' as a root node here, various shot down proposals to change the top level. We can revisit the proposal to use CARO in Barcelona.

Even though our hands are tied at the top level, I don't think we should reify the weirdness in shadowed classes. How about

1. Adding an exact syn "cell or subcellular entity" to CC.
2. Tagging this somehow
3. For GO:0006928, s/cellular component movement/movement of cell or subcellular entity/
4. Adding a do_not_annotate_tag (AFAICT this is just a grouping. Presumably if you see movement you have more specific sense of what is moving)
5. Also adding to a do-not-show-in-TE antislim

Original comment by: cmungall

[gocentral]

Maybe we can change the def of CC too.

Original comment by: cmungall

[gocentral]

Hi Chris,

That sounds like a sensible solution.

Thanks,

Pascale

Original comment by: pgaudet

[gocentral]

I like Chris' suggestions. I'm going to leave this issue open a little longer before implementation.

Original comment by: tberardini

[gocentral]

I implemented these: -Adding an exact syn "cell or subcellular entity" to CC. -For GO:0006928, s/cellular component movement/movement of cell or subcellular entity/ -(For GO:0006928) Adding a do_not_annotate_tag (AFAICT this is just a grouping. Presumably if you see movement you have more specific sense of what is moving) -(For GO:0006928) Also adding to a do-not-show-in-TE antislim

I did not change the definition of CC as I couldn't come up with good wording. Open for suggestions on that one.

Original comment by: tberardini

[gocentral]

Still looking for help in redefining cc to include 'cell'.

Current: The part of a cell or its extracellular environment in which a gene product is located. A gene product may be located in one or more parts of a cell and its location may be as specific as a particular macromolecular complex, that is, a stable, persistent association of macromolecules that function together.

Original comment by: tberardini

[tberardini]

I can't think of a good way to augment the 'cellular component' definition to include cell. If someone else has a flash of inspiration, please add it here.

Have things progressed in other ontologies so that we could perhaps drop 'cell' from GO at this point in time? @cmungall ?

[cmungall]

On 17 May 2016, at 17:13, Tanya Berardini wrote:

I can't think of a good way to augment the 'cellular component' definition to include cell. If someone else has a flash of inspiration, please add it here.

Have things progressed in other ontologies so that we could perhaps drop 'cell' from GO at this point in time? @cmungall ?

Having a CL ID in GO will cause bioinformatics pipelines around the world to crash loudly. Sometimes forcing breakage is good to make sure people conform to the actual spec, but not really worth it here just for cell, when we can just obsolete the CL class instead

---

You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/geneontology/go-ontology/issues/11414#issuecomment-219891184

[dosumis]

Why does the CL ID have to be in GO? As GO:cell has no SubClasses in GO, we can just obsolete it and import a relevant term. This will be filtered out in everything but go-plus.owl. So, we just need a suitable general term - one that encompasses encapsulating structures. I could make sure the CARO def applies and make this the parent class to the one in CL...

[tberardini]

Thanks for input. I was hoping that reviving discussion might spur a solution that was previously not possible. I think I understand what David is getting at but I'm not sure how to implement that solution. There are more than a handful of experimental annotations directly to this term, not including annotations made by the BP-CC annotation inference pipeline. They would have to be dealt with by the annotating databases if we obsoleted the GO:cell.

[cmungall]

This is interesting, it looks like all those AgBase ones are proxies for tissue-specific gene expression smuggled in via annotation extensions. If they could coordinate with Bgee they may be able to ditch these.

Without smuggled anatomy:

It's looking more doable, and I'm coming round to @dosumis POV. We should also check that go-basic doesn't end up looking odd with GO:cell gone

[ValWood]

the original issue pointed out by @RLovering has gone away. Anything else to do here?