Change justification for Uniprot to Ensembl linkset

[stain]

The linkset Uniprot-to-Ensembl has the justification SIO_000985 (protein coding gene)

Linkset: http://data.openphacts.org/dev/ims/linksets/uniprot/uniprot_ensembl.ttl.gz

ENSG IDs (as given as example) are not included in the linkset. Was this due to the one to many issue? 

All Ensembl entities included here seem to be transcripts. Is protein coding gene really the right justification here? Maybe SO:0000233 (mature transcript) would be better.

Note that there are also the opposite linksets from Ensembl to Uniprot from @JonathanMELIUS's linksets, which are detailed such as HomoSapiens, translation. I have not checked if there are overlaps here.

Here's the SPARQL query for the Uniprot-to-Ensembl linkset.

[stain]

Need to evaluate if this linkset is even needed considering the opposite direction Ensembl-Uniprot linksets from Jonathan which have a probably stronger foundation than Uniprot's loose "see also" notation.

[Chris-Evelo]

I have suggested to remove it a few times in the past. It keeps popping up as a cause for inconsistent behaviour. Given the way we build other gene product related gene sets it doesn't make sense to have this. We only added it as a quick hack early on.

[nicklynch]

@Chris-Evelo So to summarise, we would remove this linkset Uniprot-to-Ensembl for human and non human? Is there value in keeping the non-human coverage? Do we risk missing data if we exclude the non-human?

[danidi]

My initial issue here was that this linkset contains only ensembl transcript IDs (starting with ENST), but no ensembl gene IDs (at least not for the human ones, I didn't check the other organisms).