Hi, could you please stop merging wikidata entries about genes and proteins? The community consensus is that that these should be modeled as different wikidata items that are linked to each other by the encodes Property:P688 and encoded_by P702 properties. The merges you are performing are breaking a community-agreed upon data model and making it difficult to use the data - for example, in Wikipedia infoboxes. --I9606 (talk) 23:19, 9 November 2015 (UTC)[reply]
You are doing many merges of like this: https://www.wikidata.org/w/index.php?title=Q2745322&oldid=272631096 where you are taking a gene item (which is about a region of genome) and a protein item (which is about the molecular product of a gene) and merging them into one item. As explained in the discussion about this topic and the community consensus on the molecular biology project page, these merges are semantically incorrect and are causing a lot of problems for the people maintaining the gene and protein items. May I ask why you are doing this and what tool you are using to do it? Thanks --I9606 (talk) 19:38, 16 November 2015 (UTC)[reply]
I Don't understant the problem : english article STK11 and french article (LKB1) have the same subject : they are both about the gene AND the protein and I think it is quite logical to make interwiki. For that, I do it manually; I find the english article (STK11) corresponding to the french article and I click on "add links", just under "languages", fill the form and click OK.Nguyenld (talk) 19:49, 16 November 2015 (UTC)[reply]
The problem here is that there is not a 1-1 correspondence between items in wikidata and articles in the various Wikipedias. You are right, those articles are about both concepts. That does not mean that we should only have one item in wikidata for everything in that article. As an example, one gene can encode multiple proteins that are truly different entities (see 'alternative splicing'). The data model that the wikidata community agreed upon (through lengthy discussion linked to above) was to separate gene items from protein items and establish links between them via the 'encodes' and 'encoded by' properties. This makes it possible to manage the information much more effectively. For the Wikipedia article use case, all of the information on both item types (protien and gene) is accessible for use on the infoboxes in the article. See for example, the infobox that is currently displayed on https://en.wikipedia.org/wiki/Reelin . It is composed by queryng wikidata for properties of the gene and for properties of the proteins that the gene encodes. If you merge those to items you will destroy that infobox. --I9606 (talk) 19:56, 16 November 2015 (UTC)[reply]
I don't want to merge anything in wikidata. I just want to link english article to equivalent foreign language articles. How can I do that simply ? Nguyenld (talk) 20:25, 16 November 2015 (UTC)[reply]
The way to handle this is to adjust the interwiki links on the wikidata items. One wikidata item (e.g. on the gene) can link to many different language articles. (one per language wikipedia) . Let me know if you need help with that. --I9606 (talk) 22:16, 16 November 2015 (UTC)[reply]
To be clear, we want the wikidata item about the gene to be the center for the inter-language links. That picture there is showing all of the different languages that link to the wikidata item for the human Reelin gene (and through that item, to each other).--I9606 (talk) 23:09, 16 November 2015 (UTC)[reply]
