gp2protein, gp2rna

[lpalbou]

We have gp2protein and gp2rna (I think for most revisions) which offer mapping to UniProt IDs. I personally think those would be useful to keep as those accession numbers can still be referred to in UniProt. Example:

FB:FBgn0000008  UniProtKB:Q9W283;UniProtKB:A0A0B4LG21
FB:FBgn0000014  UniProtKB:P29555
FB:FBgn0000015  UniProtKB:P09087
FB:FBgn0000017  UniProtKB:P00522
FB:FBgn0000018  UniProtKB:Q9VKM4;UniProtKB:Q8MR99;UniProtKB:Q95W10
FB:FBgn0000022  UniProtKB:P10083
FB:FBgn0000024  UniProtKB:P07140
FB:FBgn0000028  UniProtKB:P24350
FB:FBgn0000032  UniProtKB:Q8I0P9;UniProtKB:Q9VAD0
FB:FBgn0000036  UniProtKB:P09478
FB:FBgn0000037  UniProtKB:P16395
FB:FBgn0000038  UniProtKB:P04755
FB:FBgn0000039  UniProtKB:P17644

If we keep them, in which destination folder do we want them ? annotations/gp2protein (probably) ? or just gp2protein/ ? Reminder of the current folder hierarchy:

@pgaudet @cmungall

[pgaudet]

@vanaukenk What do you think, do we need those files ? Given that the specs were not great and not all groups provided them, I would be tempted to drop this. But GOA has them since 2013.

Any strong feeling one way or the other ?

Thanks, Pascale

[vanaukenk]

I'm inclined to say we don't need to archive these files.
I can't immediately think of a use case for them and it seems that, if really needed, they could be reconstructed from UniProt records. Perhaps we should just confirm this on tomorrow's managers call, though.

[lpalbou]

I think I would tend to keep these files, as they could be very useful for a bioinformatician trying to remap the entities of the time and compare enrichments at different time points. Sure we can discuss it tomorrow

[pgaudet]

Discussion with @thomaspd and @lpalbou

Suggest to keep them (since we have them), under a new folder in /metadata -> gp2protein

@kltm Is that OK for you ?

Thanks, Pascale

[pgaudet]

Or - should we call the folder 'id_mappings', since it seems like this is what they are ? and have it in the /annotation folder

[pgaudet]

@kltm and I propose /annotation/gp2protein

OK @lpalbou @thomaspd

Thanks, Pascale

[lpalbou]

This is done for both the CVS & SVN:

• SVN example: https://geneontology-test.s3.amazonaws.com/2013-05-01/annotations/gp2protein/index.html
• CVS example: https://geneontology-test.s3.amazonaws.com/2005-06-01/annotations/gp2protein/index.html

[pgaudet]

Are the SVN and CVS examples leading to similar pages ? I couldn't see a difference.

It looks great !

[lpalbou]

They are leading to different pages but that’s the whole purpose that a release created from CVS (2002-2011) or SVN (2011-2018) look the same 😃

[pgaudet]

Will everything be integrated into a single page ? I suppose we dont want to keep the distinction of whether the original data was on SVN or CVS ?

[lpalbou]

I am not following, you have two different pages because they point to two different releases. The SVN example was to show how the migration worked for the 2013-05-01 release (built from SVN) while the CVS example was to show how the migration worked for the 2005-06-01 release (built from CVS). Both links were provided to show that indeed, whatever the release a user will look at throughout that period and independently on how I rebuilt the archive, it will be shown in a consistent way..

[pgaudet]

OK thanks for the precision !