logical defs of movement: cilium & microtubule-based

[malcolmfisher103]

Unexpected inference in gomodel:5f46c3b700004144

I made a small GO-CAM model in Noctua for Xenopus fam183b from a paper that covered both Xenopus and mouse, MGI were already curating the mouse stuff.

One of the annotations is for 'epithelial cilium movement involved in extracellular fluid movement [GO:0003351]'. The reasoner adds an inferred process of 'establishment of localization in cell [GO:0051649]'. This doesn't seem to be an obvious inference. I thought initially it might be because I said the process occurred in multiciliated cells but even when I removed that relationship the inference was still there.

Is this inference correct?

[pgaudet]

@vanaukenk @ukemi

[ukemi]

The inference is correct based on all of the rules. But it points to a problem in the ontology.

'epithelial cilium movement involved in extracellular fluid movement' is asserted to be a subclass of 'extracellular transport'. 'extracellular transport' is a subclass of 'transport'. 'transport' is a subclass of 'establishment of localization'. The 'epithelial cilium movement involved in extracellular fluid movement' in your model occurs_in some 'multiciliated epidermal cell'. 'multiciliated epidermal cell is a subclass of cell. Therefore 'epithelial cilium movement involved in extracellular fluid movement' is a type of 'transport' that occurs in a cell. Therefore it is a type of 'establishment of localization' that occurs_in a cell. 'establishment of localization in cell' is defined as 'establishment of localization' and ('occurs in' some cell). Following this logic, if the 'epithelial cilium movement involved in extracellular fluid movement' occurs in a cell, then it is 'establishment of localization in cell'.

I would question whether 'epithelial cilium movement involved in extracellular fluid movement' is a subclass of 'extracellular transport'. The 'involved in' phrasing of the term suggests that when it was created, the movement of the cilium was part_of 'extracellular transport'. This would be the easiest solution, but brings up some interesting issues about the spatial relations of transport. It would be fun to look at logically defining the term with nested expressions.

'epithelial cilium movement involved in extracellular fluid movement' = ('cilium movement' occurs_in some 'epithelial cell') and results_in_movement_of some (fluid? adjacent_to some 'epithelial cell'). Or a new cc term extracellular fluid?? I can't find fluid in an ontology. It would be nice to indicate the same instance of the epithelial cell.

[malcolmfisher103]

Thanks for that detailed answer David. One small thing is that the inference persisted even when I removed the 'occurs in' relationship to the cell type. So the singular term seems to be enough to draw the inference, perhaps because the cilium 'is a' cellular component?

[ukemi]

Ah yes. I suspect that once the occurs_in is removed, the inference is because 'cilium' is a subclass of 'plasma membrane bound cell projection'. 'plasma membrane bound cell projection' is part of a 'cell'. 'cilium movement' results in movement of some 'cilium'. @balhoff do you agree?

[deustp01]

I'm hanging up on the occurs_in part. If a cilium protruding outwards from a cell beats and thereby sets up a flow of fluid in the extracellular space (e.g., to clear mucus from lung epithelia, or to set up a directional flow sensed by other cells in a developing embryo to establish left - right asymmetry), then even though a cell part is doing the agitating and definitely depends on things inside the cell to provide energy and leverage, the flow is happening entirely on the outside. Is there any mileage in this?

[ukemi]

Yes. There is something funny about this and the way we represent it.

[deustp01]

Drag Karen into the inside - outside discussion. She really understands the weedy details!

---

From: David Hill notifications@github.com Sent: Tuesday, November 3, 2020 10:43 AM To: geneontology/go-ontology Cc: D'Eustachio, Peter; Comment Subject: Re: [geneontology/go-ontology] Unexpected inference in gomodel:5f46c3b700004144 (#20258)

[EXTERNAL]

Yes. There is something funny about this and the way we represent it.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_geneontology_go-2Dontology_issues_20258-23issuecomment-2D721208603&d=DwMCaQ&c=j5oPpO0eBH1iio48DtsedeElZfc04rx3ExJHeIIZuCs&r=2Sysx5329O--oOLYkHqNwFvpSt4Hk2pxc3MQOQUUTGQ&m=FFD67Q5b5hFUOZ6K9XWHQetk_9C_bG62VqQ0IV2FiD4&s=tK0VFfF3_SLRNzvAefBrmk7QjYmMBDt0H51Rbyb_Sjc&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_ADBFYVZLTAM7S4BEDE5X3ILSOAQLNANCNFSM4TIM3SKQ&d=DwMCaQ&c=j5oPpO0eBH1iio48DtsedeElZfc04rx3ExJHeIIZuCs&r=2Sysx5329O--oOLYkHqNwFvpSt4Hk2pxc3MQOQUUTGQ&m=FFD67Q5b5hFUOZ6K9XWHQetk_9C_bG62VqQ0IV2FiD4&s=rVXIwAWDsAxaArhMKcsRDLh4Pk9khnhtTuBYl4qUygo&e=.

---

This email message, including any attachments, is for the sole use of the intended recipient(s) and may contain information that is proprietary, confidential, and exempt from disclosure under applicable law. Any unauthorized review, use, disclosure, or distribution is prohibited. If you have received this email in error please notify the sender by return email and delete the original message. Please note, the recipient should check this email and any attachments for the presence of viruses. The organization accepts no liability for any damage caused by any virus transmitted by this email.

[ukemi]

@malcolmfisher103 , are you sure you refreshed after you deleted the occurs_in 'multiciliated epithelial cell'? I just tried it with @balhoff and we didn't see the inference.

[ukemi]

@krchristie anything to add? An interesting discussion about cilia, where they are moving and location of the things that are a result of their movement. Perhaps bring this up on an ontology call?

[malcolmfisher103]

Hi David, you are right I hadn't refreshed it, I think I expected the reasoner to update it when I saved the changes.

[ukemi]

I'm kind of surprised it didn't when you saved. Weird.

[krchristie]

Ah yes. I suspect that once the occurs_in is removed, the inference is because 'cilium' is a subclass of 'plasma membrane bound cell projection'. 'plasma membrane bound cell projection' is part of a 'cell'. 'cilium movement' results in movement of some 'cilium'. @balhoff do you agree?

I think it is a tautology to say that cilium movement results in movement of itself, and probably not really what we want in GO.

Here are the logical definitions of cilium movement and microtubule-based movement

cilium movement (GO:0003341)

'movement of cell or subcellular component'
 and ('results in movement of' some cilium)
 and ('process has causal agent' some microtubule)

microtubule-based movement (GO:0007018)

'movement of cell or subcellular component'
 and ('process has causal agent' some microtubule)

I think that the logical definition of cilium movement should probably be changed so that it does not say that the cilium is the causal agent. I don't think it's correct to say that the microtubule is the causal agent of cilium movement anyway, as the MTs are present in non-motile cilia too. For cilia, the causal agent, or agents, would include the dynein arms. I'm wondering if it's really correct to say that the causal agent of microtubule-based movement is the MTs. It seems to me that the MTs are often the highway upon which motors move.

[krchristie]

@malcolmfisher103 - Regarding your model, it seems that you should be able to put the two GO terms in line with each other since we know that cilium assembly is upstream of cilium motility. Basically, whatever Fam183b.L is doing in the process of cilium assembly is upstream of epithelial cilium motility involved in extracellular fluid movement.

[malcolmfisher103]

@krchristie I had been considering making cilium assembly upstream of the motility term, thanks for the advice.

I think David and Jim have identified the 'occurs in' relationship with the 'multiciliated epidermal cell' as the culprit for where the inference came from, with some issues with my interaction with Noctua being why it persisted. So is this an incorrect relationship that I should just remove? The cilium is definitely part of the cell but should I say the motility 'occurs in' the cell?

[krchristie]

@krchristie I had been considering making cilium assembly upstream of the motility term, thanks for the advice.

I think David and Jim have identified the 'occurs in' relationship with the 'multiciliated epidermal cell' as the culprit for where the inference came from, with some issues with my interaction with Noctua being why it persisted. So is this an incorrect relationship that I should just remove? The cilium is definitely part of the cell but should I say the motility 'occurs in' the cell?

As a biologist, I think that the cilium movement is occurring within the cell since there are motors moving within the cilium that create the movement of the organelle as a whole and that movement is occurring within a part of the cell. However, due to its location on the external surface of the cell, the cilium is also moving through the extracellular region and the movement of extracellular fluid is definitely occurring outside of the cell. Many things about the cilium are tricky in the ontology. It isn't the first time we've had a hard time defining cilium terms in the ontology. I've added the editors-discussion tag so that we'll discuss this at an ontology editors' call.

[ValWood]

The problem is the

GO:0003351 epithelial cilium movement involved in extracellular fluid movement parent GO:0099111 microtubule-based transport

movement is not always transport. This parentage is wrong on a number of levels. The fluid is being transported by the epithelia. cillium movement is a microtuble based movement, but not a microtubule based transport I guess this term would need an additional parent to describe the fluid movement

[ValWood]

movement is not always transport.

was an understatement.

Movement and transport are different things, but often conflated in the ontology. "movement is physical motion between points in space while transport is an act of transporting"

microtubules move things (i.e as cillia), or by pushing (the nucleus), or pulling (chromosomes apart). I don't think this is transport?

they are also part_of the process of transport, when they provide the tracks for vesicles and cargo to be moved (for instance components of the spindle pole body, are transported along the spindle microtubules towards the SPB). Lost chromsomes are pulled toward the spindle pole body during mitosis. I think this happens on 'depolymerizing mnicrotubules', so it might look a bit like transport, but it probably isn't.

I don't know if this helps to disambiguate?

At present we have GO:0099111 microtubule-based transport as part of transport, but not all microtubule based movement is transport.

I think you might fix the problem here by moving

GO:0003351 epithelial cilium movement involved in extracellular fluid movement

to be a child of GO:0007018 microtubule-based movement

instead of GO:0099111 microtubule-based transport

[pgaudet]

Maybe the genera should be 'GO:0046907 intracellular transport' rather than 'movement of cell or subcellular component'

(or another term - just 'movement of cell or subcellular component' is a strange grouping class.

[krchristie]

11/9/20 - Discussion on ontology editors call

• The logical definition of microtubule-based movement (GO:0007018) = 'movement of cell or subcellular component' and ('process has causal agent' some microtubule) is not sufficient. It also requires an additional causal agent of some type of microtubule motor.
• In the logical definition of cilium movement (GO:0003341), it seems very odd that the genera is a term that is so broad: 'movement of cell or subcellular component'. @krchristie will look into what it should be.

[krchristie]

Maybe the genera should be 'GO:0046907 intracellular transport' rather than 'movement of cell or subcellular component'

I don't think that 'GO:0046907 intracellular transport' will work since the microtubule motors that move the cilium don't transport anything; they just change the relative positions of the microtubules within the cilium resulting in cilium bending, aka movement.

(or another term - just 'movement of cell or subcellular component' is a strange grouping class.

Agreed!

[ukemi]

Is that always true? I thought there were cases where there was transport of receptors along microtubules in the cilium so that they could be deposited in the cilium and the cilium would act as an antenna for the ligands.

[krchristie]

Is that always true? I thought there were cases where there was transport of receptors along microtubules in the cilium so that they could be deposited in the cilium and the cilium would act as an antenna for the ligands.

Transport of receptors along MTs in the cilium occurs by IFT (intraflagellar transport) and is caused by movement of the IFT motors along ciliary MTs, so this is transport. This occurs in non-motile primary cilia for sure, and probably also in motile cilia, but I haven't seen much research characterizing signaling in motile cilia. IFT motors do not generate movement of cilia, though IFT is required upstream of ciliary movement because IFT is required to transport things into the cilium as part of assembly, so when IFT is broken, ciliary movement is generally also broken.

Ciliary movement is caused by the motors of the dynein arms, which are present in motile cilia, but not in non-motile ones. Movement of dynein arms does not transport anything, but shifts the relative connection points between two ciliary MTs such that bending of the cilium occurs.

I edited my previous comment to accurately specify the microtubule motors THAT MOVE the cilium rather than the the microtubule motors IN the cilium

[krchristie]

microtubules move things (i.e as cillia), or by pushing (the nucleus), or pulling (chromosomes apart). I don't think this is transport?

MTs are often the scaffold along which a MT-motor may move along a MT to generate movement; this is how motile cilia move and a common mechanism of transport. However, MTs alone are often not sufficient for movement, e.g. non-motile cilia which have MTs, but no dynein arms with motors to shift the relative positioning of the MTs.

My recollection is that there are other types of microtubule mediated movement where the movement is generated by shortening or lengthening an anchored MT that has some cargo attached at other end. @ValWood - do you know if this is what is occurring in the examples you mentioned above for "pushing (the nucleus), or pulling (chromosomes apart)"?

[krchristie]

I see that we already have this term: microtubule polymerization based movement (GO:0099098). But this term chromosome movement towards spindle pole (GO:0051305) is due to shortening rather than polymerization so it isn't a child of that term.

I think we're going to need some new grouping terms under microtubule movement. Here's some initial thoughts.

-- microtubule movement (GO:0007018) --- term to group movement based on MT length change (GO:NEW) ---- microtubule depolymerization based movement (GO:NEW) ---- microtubule polymerization based movement (GO:0099098) --- term to group movement based on movement of motor along MT (GO:NEW)

not sure whether existing term microtubule sliding is related to motors or not...

We may also need new child terms under microtubule-based transport to reflect the subgroupings I'm thinking about above.

[ValWood]

Yes, for chromsome pulling apart microtubules depolymerize, and the DAM/DASH complex acts as a rachet, so this is the microtubule kinetochore attachment. https://www.mdpi.com/1422-0067/20/24/6182/htm This is 'microtubule-dependent movement', but it isn't transport (at least, I do not think anybody would describe chromosome segregation as a type of transport, but the chromosomes move apart).

Then there is this situation: https://www.pombase.org/reference/PMID:31483748 microtubules-dependent processes also provide the force to oscillate the nucleus at interphase (I don't think this is transport intuitively either).

There is also the microtubule-based movement with kinesins providing the force moving the SPBs apart during spindle formation https://www.pombase.org/reference/PMID:29167352

These processes are all microtubule-based movement, but not transport. I looked at these terms last night and they all seemed to make sense except for the one parentage I pointed out above.

I think this proposal is making things more complicated. I wouldn't add any more grouping terms.

I still think you can resolve the current problem by removing

GO:0003351 epithelial cilium movement involved in extracellular fluid movement parent GO:0099111 microtubule-based transport

which is clearly wrong. Cillium movement isn't microtubule-based transport.

[ValWood]

I updated the above comments, I think the depolymerization terms are OK but I'm sure there are now aspects of these processes, especially around chromosomes segregation, where we know a lot of the activities and the processes could be simplified to allow more modular annotation.

But I don't think it is simple to classify as polymerization or depolymerization based movement. A single process can combine a number of mechanisms and forces.

[ValWood]

I requested some of these polymerization terms over the years when we knew less about the processes.

[krchristie]

I still think you can resolve the current problem by removing

GO:0003351 epithelial cilium movement involved in extracellular fluid movement parent GO:0099111 microtubule-based transport

which is clearly wrong. Cillium movement isn't microtubule-based transport.

It really isn't quite that simple.

This parentage is coming from reasoning based on the existing equivalence axiom of several terms including this term:

cilium movement (GO:0003341)

'movement of cell or subcellular component'
 and ('results in movement of' some cilium)
 and ('process has causal agent' some microtubule)

I've already commented about the fact that this equivalence axiom is not sufficient. Fixing this may require adding grouping terms under microtubule-based movement in order to make appropriate logical axioms for other terms.

Then this term, the one being used for annotation,

epithelial cilium movement involved in extracellular fluid movement (GO:0003351) does not have an equivalence axiom, but does have these ASSERTED SubClass Of relationships:

• cilium movement
• extracellular transport [def = The transport of substances that occurs outside cells.]

and this INFERRED SubClass Of relationship:

• microtubule-based transport

The ASSERTED relationships seem true, but the INFERRED one is incorrect. To remove the inferred relationship, it is likely that we will have to fix the logic in microtubule-based movement, which I hope will make it possible for me to define this term with an equivalence axiom, rather than just with some directly asserted relationships.

[ValWood]

OK got it.

I still think we should ask first whether "microtubule-based movement" is really a useful grouping term. All of the children should have other parentage describing specific processes, and "microtubule-based movement" groups so many unrelated concepts (other than that they are microtubule-based) it might not really be so useful (microtubule-based process would suffice).

Also, the persistance of the precomposed term- I thought GO was removing all of the precomposed process terms with "involved in" so the speciific term epithelial cilium movement involved in extracellular fluid movement (GO:0003351) presumably would go at some point?

[pgaudet]

I dont know very much about those processes, but it seems to me 'microtubule-based movement (GO:0007018)' is a good grouping term, and can include microtubule polymerization and depolymerization-based movements, but I am wary of creating too many grouping therms under that class - so I would NOT create those. In fact I suggest we go the other way sand obsolete 'microtubule polymerization based movement (GO:0099098) (has no direct annotations).

I suggest we dont create those:

--- term to group movement based on MT length change (GO:NEW)

The definition of 'microtubule-based movement (GO:0007018) already covers that: "A microtubule-based process that results in the movement of organelles, other microtubules, or other cellular components. Examples include motor-driven movement along microtubules and movement driven by polymerization or depolymerization of microtubules."

---- microtubule depolymerization based movement (GO:NEW) --- term to group movement based on movement of motor along MT (GO:NEW)

Thanks, Pascale

[ValWood]

I agree. 'microtubule-based movement (GO:0007018) is sufficient. The rest are mechanistic, so they now seem odd in BP.

[krchristie]

The reason I'm suggesting these terms is that I think that they may be needed in order to make equivalence axioms that are both necessary and sufficient. For some of these types of MT-based movement, a MT motor that moves along the MT is required. I don't think that is true for the types of movement that are based on MT length change, i.e. polymerization or depolymerization. I don't want them as annotation terms, but as logical definitions that can be used in order to define other terms.

[pgaudet]

Let's see if we can do without, first ?