Missing xref abbreviations in go-dbxrefs.yaml

[cmungall]

From mike:

I mentioned yesterday at lunch that there are at least 3 abbreviations that are used in the WITH column of the goa_uniprot_all file. I checked that these abbreviations are not presented in the db-xrefs.yaml and thus not the GO.xrf_abbs file.

• irgsp - rice annotations, thousands of annotations
• araport11 - Arabidopsis annotations, tens of thousands of annotations
• insdc - International Nucleotide Sequence Database Collaboration, tens of annotations

Perhaps there are others abbreviations that are missing but these are the ones I found. The impact of this is the filtering script is that these are considered errors and thus the annotation is removed from the goa_uniprot_all file in the main directory.

[tonysawfordebi]

Quick update: these identifiers are all coming from Ensembl.

[tonysawfordebi]

I've been talking to Ensembl, and they're going to bring the database prefixes that they use into line with what's in db-xrefs.yaml; they'll ask us to add any that aren't in the file.

[jmcherry-zz]

In the goa_uniprot_all.gaf.gz files from 2017-10-23 I find 429844 errors in the WITH column, most of these are from araport11. Do you know the timeframe from Ensembl when they will get in line? As well as being in amigo, SGD would like to include some of their yeast IEAs.

[tonysawfordebi]

I don't, but I'll give them a nudge.