Property talk:P7963

Documentation

Cell Ontology ID
ID in the Cell Ontology

Applicable "stated in" value

Cell Ontology (Q55118285)

Data type

External identifier

Domain

According to this template: cell type (Q189118)

According to statements in the property:
anatomical structure (Q4936952) or cell type (Q189118)

When possible, data should only be stored as statements

Allowed values

CL_[0-9]{7}

Usage notes

Map instances of cell types to their Cell Ontology IDs. Include the CL_ Prefix.

Example

melanocyte (Q247101) → CL_0000148 (RDF)
red blood cell (Q37187) → CL_0000232 (RDF)
neuron (Q43054) → CL_0000540 (RDF)
macrophage (Q184204) → CL_0000235 (RDF)
bronchial epithelial cell (Q30029740) → CL_0002328 (RDF)

Source

https://github.com/obophenotype/cell-ontology
http://www.ontobee.org/ontology/CL

Formatter URL

http://purl.obolibrary.org/obo/$1

type constraint

Latest comment: 4 years ago16 comments4 people in discussion

Fizzled, should be fixed but won't happen soon. Create a new section to fix. Iwan.Aucamp (talk) 17:56, 13 April 2020 (UTC)Reply

@TiagoLubiana, ArthurPSmith, Gtsulab: I made the subject type constraint (Q21503250) here relation (P2309)instance of (Q21503252) AND class (P2308)cell type (Q189118) as per the proposal - but I think the subject type constraint (Q21503250) should allow either:

If only one constraint can be captured I think relation (P2309)subclass of (Q21514624) AND class (P2308)cell (Q7868) may be better.

I'm unsure how to capture such a constraint though and I am trying to figure it out, but in the mean time if you disagree with the proposed constraint please raise objections here. Iwan.Aucamp (talk) 10:02, 15 March 2020 (UTC)Reply

WikiProject Molecular_biology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

Notified participants of WikiProject Biology

Notified participants of WikiProject Anatomy

Discussion

Comment After some contemplation I have come to the conclusion that the proposed constraint cannot be captured, still would like input as to what constraint makes the most sense. As I said before I favor (relation (P2309)subclass of (Q21514624) AND class (P2308)cell (Q7868)). Iwan.Aucamp (talk) 15:50, 15 March 2020 (UTC)Reply

Hello @Iwan.Aucamp:. As far as I know, there are no instances of cells in the cell ontology, only instances of cell types. An instance of a cell would be something very specific, like the ovule that gave rise do Dolly, the cloned sheep. It is so specific that I believe there is not even an item for that. All cell types that are currently classified as instances of cells are actually misclassified. The concepts are often mixed, even in the academic literature.TiagoLubiana (talk) 16:26, 15 March 2020 (UTC)Reply

@TiagoLubiana: Currently the constraint is (relation (P2309)instance of (Q21503252) AND class (P2308)cell type (Q189118)) - which is how I understood your proposal. If we look at your examples though:

However, for macrophage (Q184204) there is only:

macrophage (Q184204)subclass of (P279)cell (Q7868)

There is no

macrophage (Q184204)instance of (P31)cell type (Q189118)

Therefore adding Cell Ontology ID (P7963) to macrophage (Q184204) will violate the property constraint (which it does, as I added it from examples).

The alternative property constraint (relation (P2309)subclass of (Q21514624) AND class (P2308)cell (Q7868)) will not be violated by having Cell Ontology ID (P7963) to macrophage (Q184204). Now I would guess that macrophage (Q184204)instance of (P31)cell type (Q189118) should exist but this still leaves the question which is the better subject type constraint (Q21503250) of these two:

I think the first, but that is what I want to confirm here. If you prefer something completely different please clarify. I don't want to make anything instance of (P31)cell (Q7868). Iwan.Aucamp (talk) 16:41, 15 March 2020 (UTC)Reply

@Eihel: following from this. As stated here already I think the right constraint is relation (P2309)subclass of (Q21514624) AND class (P2308)cell (Q7868). I think bronchial epithelial cell (Q30029740)instance of (P31)anatomical structure (Q4936952) is wrong and violates the guidelines in Help:Basic_membership_properties - it should be removed and replaced with bronchial epithelial cell (Q30029740)subclass of (P279)cell (Q7868). Iwan.Aucamp (talk) 10:14, 21 March 2020 (UTC)Reply
I advised Iwan on his TP without knowing the discussion here and I already made a change (still without knowing this discussion). Sorry. I would start by saying that this is also a subject specific to WD WikiProject property constraints. These participants may have a different, even more neutral perspective.
- For your proposal of Item bronchial epithelial cell (Q30029740), I do not agree. Yes it works, but the basic concepts of belonging are still not followed. I allow myself a small summary:
Translate the following terms by:
P31 - is an example of ...

P279 - is a subset of
Think of set theory in mathematics. The sets can consist of other sets and simple elements. But on WD, once you get to the end of the sets (classes), you practically only find simple elements (Items). These concepts are used so that Items are properly linked. On WD, there are around 75 million instance of (P31)s and only 2 million subclass of (P279)s. It seems to me that bronchial epithelial cell (Q30029740) is a simple item. When in doubt use Special:WhatLinksHere/Q30029740. So instance of (P31) should be used for bronchial epithelial cell (Q30029740). Intuitively, we understand that subclass of (P279) will be used less and less. Using subclass of (P279) where it is not required is a persistent error on WD.
- In the Items, it is essential to choose instance of (P31) and subclass of (P279) well, but within the constraints, we can be more "flexible". Other constraints can be added to satisfy the use of the property; the aim being also not to clutter the pages of violations even more. In type constraint, there is no concept of "linked data", but of the use of a property. I am also counting on the future uses of this property and on the bad introduction of Items on WD. If I found an Item which does not satisfy the constraint, we can surely find other Items. Indeed, anatomical structure (Q4936952) is a compromise, but since I am not a biologist, I imagine that cell type (Q189118) is the element that will strictly encompass everything (with relation (P2309)subclass of (Q21514624)). Again, "strictly" means that errors may still appear. I am voting Support anatomical structure (Q4936952) as currently used. Cordially. —Eihel (talk) 11:40, 21 March 2020 (UTC)Reply
ps. If you want to change the Items to conform to the constraint, there is a big job. Normally, cell (Q7868) should be used little and Items should have in P31 blood cell (Q211709), somatic cell (Q725508), myoepithelial cell (Q1476681), etc. —Eihel (talk) 12:26, 21 March 2020 (UTC)Reply

@Eihel: You really should try and consult in property pages before making changes to them as per Help:Properties#Editing_properties and at least get some consensus. Further from Help:Basic_membership_properties "subclass of (P279) is used to state that all the instances of one class are instances of another". Translating that here bronchial epithelial cell (Q30029740)subclass of (P279)cell (Q7868) would imply that all bronchial epithelial cells are also cells - which is the case.

If on the other hand something is bronchial epithelial cell (Q30029740)instance of (P31)cell (Q7868) it means that bronchial epithelial cell (Q30029740) is an individual cell (Q7868) - it is no longer a class of cell, or a type of cell - which would be wrong because bronchial epithelial cell (Q30029740) is in fact a class of cell. Put differently, say I have a particular cell (Q7868) which is noteworthy, maybe it had left-handed DNA, and this particular cell (Q7868) was designated "Sample Cell 9994". This cell seems notable so I want to make a wikidata item for it, and I now want to say "Sample Cell 9994" is a cell (Q7868), which it is - so I put Sample Cell 9994instance of (P31)bronchial epithelial cell (Q30029740) - which is correct - but now, if bronchial epithelial cell (Q30029740)instance of (P31)cell (Q7868) - I have an instance of an instance ... which does not make any sense and will raise a warning anyway.

This is in line with Help:Basic_membership_properties. I'm fairly sure the amount of times a property is used is not relevant. If there are specific guidelines that is in line with what you suggest please reference them.

Property constraints are there to tell you when you are doing something wrong - in this case if the property constraint is (relation (P2309)subclass of (Q21514624) AND class (P2308)cell (Q7868)) it will indeed say there is a violation if you put Cell Ontology ID (P7963) on bronchial epithelial cell (Q30029740) - but this is by design - because bronchial epithelial cell (Q30029740) is wrong. Iwan.Aucamp (talk) 13:54, 21 March 2020 (UTC)Reply

Notified participants of WikiProject property constraints

WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘I am not a specialist in biology Iwan.Aucamp. Okay, if there are notable cells at bronchial epithelial cell (Q30029740), let's put bronchial epithelial cell (Q30029740)subclass of (P279)cell (Q7868) as you suggested. Attention: I never said that bronchial epithelial cell (Q30029740) was right: for me, even instance of (P31) is wrong in this Item. Regarding Help:Properties#Editing properties, my change to the property you created is not a major change and therefore does not require the opinion of the community. It is precisely your responsibility that it works. Precisely, that leads us to the next paragraph.

It is not cell (Q7868) that I introduced in the constraint, but anatomical structure (Q4936952). One more reason it works, because bronchial epithelial cell (Q30029740) nowhere contains cell or cell type. There are flaws in my choices on the constraint, but they are minimal and thwarted by the other constraints. I repeat again: here, the desired effect is not to "link data", but to prevent someone from taking this property in a bad Item. My choice is not in contradiction with what we write. You get stuck on cell type and cell, but I went further. I write my choice more clearly:

property constraint

subject type constraint

edit

class

anatomical structure

relation

instance or subclass of

0 references

add reference

add value

—Eihel (talk) 18:00, 21 March 2020 (UTC)Reply

@Eihel: First off - you are suggesting that we accommodate bronchial epithelial cell (Q30029740) with the property constraint - if you don't think that bronchial epithelial cell (Q30029740)instance of (P31)anatomical structure (Q4936952) is right then it is really hard to understand why you want to accommodate it and I would appreciate if you can explain the reason why we should change this property to accommodate something which me, you and and others agree is wrong. Iwan.Aucamp (talk) 21:15, 21 March 2020 (UTC)Reply

@Eihel: Next, the reason why bronchial epithelial cell (Q30029740)instance of (P31)anatomical structure (Q4936952) is wrong has absolutely nothing to do with whether there are notable cells. If there is never a notable cell ever on wikidata it still won't make bronchial epithelial cell (Q30029740)instance of (P31)anatomical structure (Q4936952) correct. CC @ChristianKl: Iwan.Aucamp (talk) 21:16, 21 March 2020 (UTC)Reply

@Eihel: Now, to get to your proposal. We start with the property constraint (P2302)subject type constraint (Q21503250)relation (P2309)instance or subclass of (Q30208840) aspect of it. For this to be justified (as opposed to property constraint (P2302)subject type constraint (Q21503250)relation (P2309)subclass of (Q21514624)) we would need one of the following:

An item which should have Cell Ontology ID (P7963) and should also have the statement instance of (P31)anatomical structure (Q4936952) - either directly or transitively. bronchial epithelial cell (Q30029740) is not such an item because the statement bronchial epithelial cell (Q30029740)instance of (P31)anatomical structure (Q4936952) is wrong, it is wrong because instance of would imply that bronchial epithelial cell (Q30029740) is not an a type of anatomical structure (Q4936952) but instead a particular anatomical structure (Q4936952) - but of course we know that bronchial epithelial cell (Q30029740) is in fact not a particular anatomical structure (Q4936952) - and we know this because there are many bronchial epithelial cell (Q30029740)s - not just one. So because the statement is wrong, we cannot say anatomical structure (Q4936952) should also have the statement instance of (P31)bronchial epithelial cell (Q30029740) - in fact we can say the opposite anatomical structure (Q4936952) should not have the statement instance of (P31)bronchial epithelial cell (Q30029740).
A statement of intent that Cell Ontology (Q55118285) will issue identifiers for an item which should also have the statement instance of (P31)bronchial epithelial cell (Q30029740)

If you present either of these things I will agree with property constraint (P2302)subject type constraint (Q21503250)relation (P2309)instance or subclass of (Q30208840) aspect of your proposed constraint. Without either of these the constraint is not justified. Iwan.Aucamp (talk) 21:53, 21 March 2020 (UTC)Reply

@Eihel: on to the second part of your proposal property constraint (P2302)subject type constraint (Q21503250)class (P2308)anatomical structure (Q4936952). For this aspect of the constraint to be justified we would need one of the following:

An item (existing or potential) which should have Cell Ontology ID (P7963) and should also have the direct (as opposed to transitive) statement subclass of (P279)anatomical structure (Q4936952) - and this would imply that no other item with statement subclass of (P279)anatomical structure (Q4936952) would be a more precise and correct alternative to a direct subclass of (P279)anatomical structure (Q4936952) statement. So say for example the statement haploid cell (Q515386)subclass of (P279)anatomical structure (Q4936952) existed (it does not as of the writing of this), this statement could be more replaced with the more precise and equally correct statement haploid cell (Q515386)subclass of (P279)cell (Q7868) (which does actually exists) - so the hypothetical (and actual) haploid cell (Q515386) would not be justification for the aspect in question.
A statement of intent that Cell Ontology (Q55118285) will issue identifiers for an item which should have Cell Ontology ID (P7963) and should also have the direct (as opposed to transitive) statement subclass of (P279)anatomical structure (Q4936952).

As with the other aspect, if you present either of these things I will agree with the property constraint (P2302)subject type constraint (Q21503250)class (P2308)anatomical structure (Q4936952) aspect of your proposal. Without either of those I don't see any justification for it. Iwan.Aucamp (talk) 23:47, 21 March 2020 (UTC)Reply

Hello Iwan.Aucamp

After 6 months, I expected more use of this property. So I cannot demonstrate that my strategy is working. You have the possibility to add several type constraints for several cases by following the violations on these constraints. Cordially. —Eihel (talk) 06:00, 5 October 2020 (UTC)Reply

Restriction for anatomical entities

Latest comment: 4 years ago2 comments2 people in discussion

Cell types are not necessarily anatomical questions, this is actually a very difficult conceptual question. I suggest we leave this restriction out for a while. TiagoLubiana (talk) 13:15, 24 August 2020 (UTC)Reply

You have to understand the usefulness of this constraint to know that leaving it as it was is more important than removing it. This constraint allows the addition of identifiers by having humanly logical predicates on the Items concerned. The introduction of an extended class still allows you to avoid that Items having no link with this property are discarded. The problem is

to have a extended class in this constraint or
to correct Items that may be suitable for this property or
to add classes in this constraint or
to add several different type constraints (P31, P279 or both).

Choice 1 is currently used. There is no restriction to be made since an Item currently has an identifier belonging to this class. —Eihel (talk) 04:30, 5 October 2020 (UTC)Reply

Add topic