From Text search to Patent data (targets)

[danidi]

It seems there are currently two issues complicating the retrieval of patent data for targets, when starting with a ConceptWiki Target.
One is the missing conversion of HGNC names (Data source H, http://www.genenames.org/data/hgnc_data.php?match=$id) to http://rdf.ebi.ac.uk/resource/surechembl/target/ (it is included here openphacts/IdentityMappingService#14).
The other issue is the missing transitives from CW to HGNC names (probably related to #319).

The current workaround to reach patent data for CW targets (ie a text search result) is the following:

• Use CW URI in MapURL call with http://identifiers.org/entrez.gene/ as targetUriPattern.
• With the result, do a MapURL call with http://www.genenames.org/data/hgnc_data.php?hgnc_id= as targetUriPattern.
• With the result, do a MapURL call with http://info.identifiers.org/hgnc.symbol/ as targetUriPattern.
• With the result, do a text replace of the pattern with http://rdf.ebi.ac.uk/resource/surechembl/target/.

[danidi]

Just to confirm, this will also affect queries made with uniprot uris:

with HGNC name: https://ops2.few.vu.nl/2.0/mapUri?app_id=f91c5b2b&app_key=18a5d823d0e4933ac5fe22a3d52974c1&Uri=http%3A%2F%2Fidentifiers.org%2Fhgnc.symbol%2FDAPK1 finds only the HGNC accession number.

with HGNC accession number: https://ops2.few.vu.nl/2.0/mapUri?app_id=f91c5b2b&app_key=18a5d823d0e4933ac5fe22a3d52974c1&Uri=http%3A%2F%2Fidentifiers.org%2Fhgnc%2F2674 finds unigene and ensembl, but not uniprot

[danidi]

In 2.2, the given HGNC accession number example (https://beta.openphacts.org/mapUri?app_id=f91c5b2b&app_key=18a5d823d0e4933ac5fe22a3d52974c1&Uri=http%3A%2F%2Fidentifiers.org%2Fhgnc%2F2674) does no longer link to unigene and ensembl.

[ianwdunlop]

Hi @danidi do you know any of the unigene & ensembl ids that HGNC:2674 used to map to?
Also, are those issues from 2016 above still open and relevant? Maybe the CW one is not so important although uniprot should be.

[Chris-Evelo]

Hi @ianwdunlop @egonw checked that: Ensembl has that given HGNC number (2674) for ENSG00000196730...

So it should be in our linkset. @JonathanMELIUS is checking whether it really is.

It is a Death related protein BTW, so maybe that is why it is acting weird ;-)

[ianwdunlop]

I did a bit of digging last night. Here are my first findings after a bit of 'grep'. No conclusions, just that there are some mentions of the ensembl & hgnc gene.

grep '2674' linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_hgnc.dependent.LS.ttl

ensembl:ENSG00000196730
        ensemblterms:DEPENDENT  <http://identifiers.org/hgnc/2674> .

grep -r 'ENSG00000196730' linksets/

/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_insdc.inferred_from_translation.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_ncbigene.dependent.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_refseq.inferred_from_transcript.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_uniprot.inferred_from_translation.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_pdb.inferred_from_translation.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_unigene.dependent.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_ena.embl.inferred_from_translation.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_wikigenes.dependent.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_hgnc.dependent.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_ccds.inferred_from_transcript.LS.ttl:ensembl:ENSG00000196730
/home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_refseq.inferred_from_translation.LS.ttl:ensembl:ENSG00000196730

grep 'ENSG00000196730' /home/ubuntu/d/linksets/ops-ensembl-homosapiens-linksets/Ensembl_Hs_uniprot.inferred_from_translation.LS.ttl

ensembl:ENSG00000196730
        ensemblterms:INFERRED_FROM_TRANSLATION
                <http://identifiers.org/uniprot/F8WCQ3> , <http://identifiers.org/uniprot/P53355> .

[danidi]

I would assume that the original issue is still present (the related github issues openphacts/IdentityMappingService#14 and #319 are at least still open). Of course CW is no longer used, but at least uniprot should work ideally.

I was wondering if it is an issue, that we have dependent and direct as predicates for the ensembl linksets (not exact or close match). As far as I am aware, #319 was not solved. It might be, that in earlier versions of the IMS the mappings were provided from CW linksets which were circumventing the ensembl linksets.