#564 follow on from new gene specific db_xrefs (Documentation and announcement)

[Valerie Wood]

We can now make gene specific db_xrefs from specific gene pages

implemented in Chado https://sourceforge.net/p/pombase/chado/436/
display will follow https://www.ebi.ac.uk/panda/jira/browse/PB-2215
migration of existing annotation will follow https://sourceforge.net/p/pombase/curation-tasks/563/

Syntax
/controlled_curation="term=external_link, description; db_xref=GEO:GSE57040"
where
description= description of the link which will appear on the gene page
Current allowed list is:
deletion mutant expression profiling
WT overexpression expression profiling
genome-wide localization

the accession will usually be of the form GEO:GSE57040 or ArrayExpress:E-MTAB-2543

To do,
1. Add syntax to curator documentation and list of allowed 'types'
Q where should the list of allowed descriptions be maintained?

1. check that these are the correct prefixes in the DBxrefs file (GEO and ArrayExpress is that correct?)

2. Let uses know that they can submit gene specific link outs (maybe a submission form for this?, initially just document the requirements in the FAQ (gene, description (from selection or suggest new, DB:ACC) and ask to submit to help desk)

[Valerie Wood]

whilst compiling the list of Arrayexpress and GEO links for the existing terms to migrate them
https://sourceforge.net/p/pombase/curation-tasks/563/

I realised that the current syntax we will lose the PMID link, and that there is nowhere to put this in the external references section anyway.

You can get to the paper from the array express or GOE link but it is a shame not to have this still on the gene page.

Should we change the proposed display of these links slightly. I think people would be more likely to see them as links in the "Gene expression section", than buried in external references, maybe?

What do you think?
(we would need to extend the syntax to include a PMID, would need to discuss with Kim if this could be done with | separated in the db-xref)

Or should we just continue as we are....

[Antonia Lock]

where would you put the genome wide localization?

[Valerie Wood]

Good point. Should we just continue as planned and lose the Pubmed links?
Any other ideas?

[Midori Harris]

if we want to avoid massive changes to a thing that seems likely to work as is, we could put a PMID in as part of the descriptive text, e.g.

controlled_curation="term=external_link, WT overexpression expression profiling (PMID:00000); db_xref=ArrayExpress:E-MTAB-2543"

It might be hard or impossible to make the PMID link, but at least it would be visible on the gene page, and people could copy & paste if they care enough.

[Valerie Wood]

I guess we could. Or we could ask MArk to make an extra column on the xrefs table for PMID....

I'll ask Kim first if it would be OK to do
/controlled_curation="term=external_link, description; db_xref=GEO:GSE57040|PMID:1111111"
if this is tricky it may force us into the

controlled_curation="term=external_link, WT overexpression expression profiling (PMID:00000); db_xref=ArrayExpress:E-MTAB-2543"

option