Wikidata:Property proposal/cell line used

cell line used edit

Originally proposed at Wikidata:Property proposal/Natural science

Not done

Description	cell line used in a given scientific article, experiment or experimental procedure
Represents	cell line (Q21014462)
Data type	Item
Domain	scholarly article (Q13442814)
Example	LC3, a mammalian homologue of yeast Apg8p, is localized in autophagosome membranes after processing (Q24597817) → HeLa (Q847482)
Source	mostly academic papers
Planned use	we plan to start experimenting with this property using WikiFactMine

Motivation

I've assembled a group of volunteers interested in extracting this kind of data from academic papers using semiautomatic (but curated) procedure - i.e. WikiFactMine. We're in touch with WikiFactMine team (tagging @Charles Matthews:) Vojtěch Dostál (talk) 17:59, 17 October 2017 (UTC)[reply]

Discussion

Support David (talk) 06:51, 18 October 2017 (UTC)[reply]
Support There are now obvious advantages in adding value to items about scientific papers, by means of properties that relate to the methods used, not just the bibliographical information. This proposal is a very good example for that direction. It is not only a use case for the WikiFactMine technology, but shows where to go with abstracting key technical points. As well as all that, it is well known that the use of cell lines causes some problems for science: see w:Immortalised cell line#Limitations. Therefore documenting this information on Wikidata will be useful. Charles Matthews (talk) 08:10, 18 October 2017 (UTC)[reply]
Support --Magnus Manske (talk) 13:31, 18 October 2017 (UTC)[reply]
Comment I certainly like the idea of annotating bibliographic items with more information, such that they can be linked to things like the cell lines used in the research described in a paper. Currently, the main way to do that is via main subject (P921), for which methodological details are often not a good fit, unless it's a methodological paper. On the other hand, we do not have a text equivalent for depicts (P180) as used on paintings and other visuals. What we do have is uses (P2283), which has begun to be used (example) on items about papers in much the same way as you are proposing here for cell lines. So instead of going for a specific "cell line used" property, it would seem more appropriate to go for this existing generic "uses" property. Having said that, I am uncertain as to whether attaching something like "uses" to the bibliographic item is the way to move forward. I may be splitting hairs too much here, but strictly speaking, it is not the paper that "uses" a given cell line, it was the researchers or the research project who used them, and the paper just reports about that usage. --Daniel Mietchen (talk) 16:52, 18 October 2017 (UTC)[reply]

You make a fair point. But if a paper "uses" test tubes, X-rays, chromatography, etc., would we be happy? Charles Matthews (talk) 19:43, 18 October 2017 (UTC)[reply]

@Daniel Mietchen, Charles Matthews: Related discussion: Wikidata:Property proposal/uses software Mahir256 (talk) 03:18, 19 October 2017 (UTC)[reply]

Expanding the thought, uses (P2283) would probably be fine, if the goal is just keywords. What I'd like to see is machine-readable abstracts of experimental method, that follow a standard pattern. While very few papers will actually have an infobox on Wikipedia, as a thought-experiment, one can think what the fields "should be", for major subject areas. Charles Matthews (talk) 10:06, 19 October 2017 (UTC)[reply]

uses (P2283) sounds good too; I'll happily leave this decision to our community --Vojtěch Dostál (talk) 07:42, 20 October 2017 (UTC)[reply]

Support Has great potential. Imagine if a known cell line was contaminated or found to be quite different from initially reported. If this property were included on entities that used this cell line, you could easily estimate the impact of such an event. Also, knowing what cell lines were used in the research can provide context on limitations and confidence on the generality of the work. Gtsulab (talk) 15:23, 19 October 2017 (UTC)[reply]
Strong oppose use generic uses (P2283). The type of used object can be inferred from instance of (P31) on the objects item. Imagine, in future you want also describe for neuroscience papers which type of neurons were used, or which animals and which brain regions were used. And for physics paper you want describe which accelerator and detector were used or for astronomy papers which telescope. Do you really plan to create a new property for each kind of objects which can be used in an experiment? --Pasleim (talk) 17:13, 23 October 2017 (UTC)[reply]

Well, Vojtěch wants to add information to items about papers, that is important in his field of molecular biology. This proposal arises out of a discussion we had at Wikimania in August. Using SPARQL, it would indeed be possible to read as far as instance of (P31) of the object of uses (P2283), as you say. On the other hand, I'm interested in "methodological abstracts", as explained above. Once the subject of a paper is established, a scientist is going to ask about just these things. There is clearly a discussion to be had about what is worth presenting as structured data. There should be little doubt in this case, however: a publisher friend told me recently that cell lines are checked in this field, before a paper is even considered for publication. Charles Matthews (talk) 08:20, 24 October 2017 (UTC)[reply]

Oppose in part per Pasleim, in part because the distinction between the paper using something and what is being described in the paper (the experiment or study) should not be restricted to cell lines. Mahir256 (talk) 00:29, 24 October 2017 (UTC)[reply]
Support ChristianKl (talk) 17:46, 29 October 2017 (UTC)[reply]

Notified participants of WikiProject Medicine ChristianKl (talk) 11:21, 30 October 2017 (UTC)[reply]

Not sure I think it is established that this is a characteristic which data researchers examine. I am unclear about when Wikidata distinguishes "uses" versus "uses a specific thing". "Cell line use" is probably a different sense of the word "use" than "using a tool". If subject matter experts want this property then it seems okay to me. Blue Rasberry (talk) 17:29, 1 November 2017 (UTC)[reply]

A huge problem with "uses" is that it can mean many things in different circumstances which can make it unclear. Additionally, I don't think that a paper that talks about an experiment which which a given cell line is featured is "using" the cell line in a way that's similar to the way binocular vision (Q609543) "uses" pair of human eyes (Q41890371). For the practical purposes of the project of Wikifactmine it also makes a lot of sense to do store this information on the items of the scientific paper and offer the information via the primary sources tool. ChristianKl (talk) 18:32, 1 November 2017 (UTC)[reply]

We could rename this proposal to "used in the related research project". If we restrict it to cell lines, we block the way to create methodological abstracts for other fields than cell biology. --Pasleim (talk) 18:44, 1 November 2017 (UTC)[reply]

@Pasleim: I created a separate proposal for it https://www.wikidata.org/w/index.php?title=Wikidata:Property_proposal/used_in_the_related_research_project ChristianKl (talk) 11:15, 2 November 2017 (UTC)[reply]

@Pasleim, Mahir256, Daniel Mietchen, Magnus Manske: @Charles Matthews, Vojtěch Dostál, Bluerasberry, ديفيد عادل وهبة خليل 2: We now have describes a project that uses (P4510), do we still need this one? ChristianKl (✉) 14:30, 12 November 2017 (UTC)[reply]
- As far as I'm concerned, no: discussion with Vojtěch produced another solution. Charles Matthews (talk) 14:35, 12 November 2017 (UTC)[reply]
- Same here. Thanks for finding a good solution --Vojtěch Dostál (talk) 15:07, 12 November 2017 (UTC)[reply]

@Pasleim, Mahir256, Daniel Mietchen, Magnus Manske: @Charles Matthews, Vojtěch Dostál, Bluerasberry, ديفيد عادل وهبة خليل 2: Not done, given that we now have describes a project that uses (P4510). ChristianKl (✉) 15:42, 13 November 2017 (UTC)[reply]

Seems satisfactory to me. Blue Rasberry (talk) 15:44, 13 November 2017 (UTC)[reply]