Fix GAF loading for RNAcentral annotations

[kimrutherford]

These errors should be fixable. There are about 1000 RNAcentral annotations that are failing because of this:

line 140430: gene feature not found, none of the identifiers (URS000030D4E9_4896 URS000030D4E9_4896) from this annotation match a systematic ID in Chado - skipping
line 140431: gene feature not found, none of the identifiers (URS000030D4E9_4896 URS000030D4E9_4896) from this annotation match a systematic ID in Chado - skipping
line 140435: gene feature not found, none of the identifiers (URS0000314D2B_4896 URS0000314D2B_4896) from this annotation match a systematic ID in Chado - skipping

https://curation.pombase.org/dumps/builds/pombase-build-2025-01-23/logs/log.2025-01-22-23-42-44.gaf-load-output

Probably we just need to remove the _4896 from the IDs like URS000030D4E9_4896 and look up in Chado using the base ID URS000030D4E9.

[kimrutherford]

This is a little trickier than I thought. The URS IDs aren't unique so some annotation rows from the GOA GAF file will map to multiple genes. Before I go ahead with implementing it, is it OK to do that?

[ValWood]

I think its the other way around. One pombe genes will have multiple URS IDs .

If it's the other way around it sounds wrong and we should look into that!

[kimrutherford]

I think its the other way around. One pombe genes will have multiple URS IDs

There are two genes in Chado with 2 URS IDs but those have 2 transcripts - each transcript has its own URS.

There are quite few URS IDs that are attached to more than one gene.

Some examples:

https://rnacentral.org/rna/URS0000314D2B/4896
URS0000314D2B │ SPRRNA.10 ║
URS0000314D2B │ SPRRNA.17 ║
URS0000314D2B │ SPRRNA.20 ║
URS0000314D2B │ SPRRNA.28 ║
URS0000314D2B │ SPRRNA.34 ║
URS0000314D2B │ SPRRNA.39 ║

URS0000415965 │ SPATRNAALA.01 ║
URS0000415965 │ SPATRNAALA.04 ║
URS0000415965 │ SPATRNAALA.05 ║
URS0000415965 │ SPBTRNAALA.07 ║
URS0000415965 │ SPBTRNAALA.08 ║
URS0000415965 │ SPBTRNAALA.09 ║
URS0000415965 │ SPBTRNAALA.10 ║
URS0000415965 │ SPBTRNAALA.11 ║
URS0000415965 │ SPCTRNAALA.12 ║

URS00002BA4D5 │ SPATRNATRP.01 ║
URS00002BA4D5 │ SPBTRNATRP.02 ║
URS00002BA4D5 │ SPBTRNATRP.03 ║

[ValWood]

OK, that's a bit weird. We should report that too.

[kimrutherford]

Fix GAF loading for RNAcentral annotations

That's fixed for the next load. Once it's done I'll compare with the previous load to see how many extra annotations we have. I suspect it won't be many.

[kimrutherford]

Once it's done I'll compare with the previous load to see how many extra annotations we have. I suspect it won't be many.

168 :-)

On Monday I'll double check to make sure we're getting all the possible annotations.

[kimrutherford]

168 :-)
On Monday I'll double check to make sure we're getting all the possible annotations.

I think it's OK. There only 630 pombe RNAcentral annotations (I miscounted earlier because I included the japonicus annotations). There a 160 or so URS IDs that don't have corresponding pombe genes which causes 334 of the 630 annotations to fail to load.

Some of the remaining annotations are filtered because of: pombe-embl/supporting_files/GO_terms_excluded_from_pombase.txt

And then some annotations are filtered because you have a more specific annotation already.

[kimrutherford]

URS0000415965 │ SPATRNAALA.01 ║
URS0000415965 │ SPATRNAALA.04 ║
URS0000415965 │ SPATRNAALA.05 ║
...

OK, that's a bit weird. We should report that too.

It looks like they group identical genes into one entry:
https://rnacentral.org/rna/URS0000415965/4896