Clarify and implement rules on inclusion of RNA gene products / RNA central

[cmungall]

There are various tickets in this repo that are very old and related to RNAcentral. Let's close these.

We should clarify the situation. I think it is as follows

• every authoritative group is reponsible for designating the IDs they bring in in their gpis, as usual
• some groups have MOD IDs for their RNA genes, they include in their GPI, that is fine, these should be working fine
• for human, the authoritative group is uniprot, and they use RNACentral - this is fine, and RNA genes seem to be loaded fine

Autocomplete works fine with the human RNAs as expected (though it's confusing as they don't have actual symbols)

If this is correct then this ticket can be closed with the addition of documentation and closing of the old tickets

[thomaspd]

Before we close, could this other issue be related to the problem of RNAs not passing the ShEx, and be traceable to NEO in some way?
geneontology/minerva#481

[balhoff]

@thomaspd yes it looks like there is a problem in NEO: #100.

[cmungall]

One other minor issue, I don't know much much it bugs curators:

RNAC GPIs lack symbols. This means that they show up with IDs on the display:

If we look at the source, e.g

https://rnacentral.org/rna/URS00005EB5B7/9606

We see that RNAC provide names "Homo sapiens let-7a-1 stem-loop (hsa-let-7a-1)" but not symbols

This is reflected in the GPI:

➜ neo git:(master) ✗ gzip -dc mirror/goa_human_rna.gpi.gz | grep URS00005EB5B7_9606 | cut -f3

➜ neo git:(master) ✗ gzip -dc mirror/goa_human_rna.gpi.gz | grep URS00005EB5B7_9606 | cut -f4
Homo sapiens let-7a-1 stem-loop (hsa-let-7a-1)

@alexsign -- is there a way to get something approaching a symbol into the GPI file (the thing in parens looks like a symbol)

If not, I propose that we add something to the neo perl scripts that uses the name for the rdfs:label if symbol is not present. Then the name will show up in the Noctua models, e.g. the ones here: geneontology/noctua#760

[pgaudet]

I dont knwow if this is true for all RNAC entities, but the label would be what's in the parenthesis - for e.g. hsa-let-7a-1

@RLovering is this correct?