In site search, taxon column should not be blank for phenotypes and diseases (models?)

[jmcmurry]

Related to https://github.com/monarch-initiative/monarch-app/issues/983 Currently taxon is only fetched for genes. Therefore it looks weird to see the column blank for disease / phenotype searches as taxon is relevant. See for example http://beta.monarchinitiative.org/search/kallmann

As a strawman, consider the following:

• [ ] Display the relevant taxon(s) for phenotypes based on the source of the term. HP-->human, MP-->mammals, FB-->fruit fly etc.
• [ ] Display the relevant taxon for disease. Most things described in terms of 'disease' and not 'phenotype' are very likely to be either exclusively or primarily human. A) Can we assume that this is true or are there non-human disease vocabularies/ontologies are out there we plan to incorporate? @cmungall please advise.
• [ ] Group the results: Genes, Phenotypes, Models(when do we actually have a model result from site search? see also https://github.com/monarch-initiative/monarch-app/issues/989)

In either case, I would also like to see an additional column for either the source or the CURIE to help distinguish and contextualize results.

[cmungall]

• We can veto the first one: ID space of ontology is not a reliable way to determine this.
• We will have animal diseases in at some point.
• Grouping may be the best option: gene/phenotype/disease.

Another option is simply to rename the column as something neutral, like 'context'

Looking ahead, we want any node in the graph to be findable here. Currently only the ontology-graph is searched, as there are issues doing full text indexing on the complete graph ( @jnguyenx - is there a SG ticket open?). Different categories of node will have different relevant neighbor nodes.

[jmcmurry]

ok, sounds good; let's perhaps:

• [ ] group: genes separately from phenotypes, separately from diseases
  • [ ] delete the category column altogether to avoid redundancy
• [ ] leave taxon as a column header in genes group
• [ ] continue the dimensions of column in genes and diseases, but change the header to either 'source' if people prefer listing just the source prefix or to 'id' if people prefer listing the CURIE. Either CURIE or source prefix gives people a hint regarding species context without running the risk of a false assumption. Like this:

Also Kudos to Github, it was a stab in the dark to paste this text from Excel into an issue, but it was immediately and automatically converted to png. Aaa-mazing.

[jnguyenx]

@cmungall here are the tickets in SciGraph for text search https://github.com/SciGraph/SciGraph/labels/vocabulary%20overhaul

[jmcmurry]

ps. Ultimately the taxon column issue will rear its head any time we have a multi-category table. Eg. http://monarchinitiative.org/labs/golr/NCBIGene:6469