Change annotations using 'has_direct_input' to 'has_input' extensions

[pgaudet]

Hello,

The GO-CAM specifications use only 'has_input'; not 'has_direct_input'.

We will get the annotations updated in Protein2GO; for groups not using this extension, please update your annotations.

Impacted groups:

AgBase ARUK-UCL BHF-UCL CAFA dictyBase FlyBase GO_Central HGNC MGI ParkinsonsUK-UCL PomBase SGD UniProt WB

AFAIK only MGI and pombase do not use Protein2GO in this list. https://docs.google.com/spreadsheets/d/1cOaBQrdW4fYfc3vVQlPWvAfFLsP_ItiPnqgKtvahUtU/edit#gid=645618912

Thanks, Pascale

[ukemi]

MGI done.

[pgaudet]

Hello,

In QuickGO there are still about 150 annotations using 'has direct input': https://docs.google.com/spreadsheets/d/1FpQY9u6kCgAKFcr79a1i_inq8vLXV5zZWGkZ7VCE5ow/edit#gid=0

Assigned by
ARUK-UCL	2 @RLovering
FlyBase	5 @hattrill
GO_Central	1 /fixed)
pombase via GOC-OWL	16 @ValWood Can you look at those ?
MGI	106 @ukemi
SGD	28 @srengel

@alexsign Can you update UCL, FlyBase and SGD annotations ?

Thanks, Pascale

[hattrill]

I quickly checked and fixed ours.

[alexsign]

@pgaudet ARUK-UCL and SGD is done. still in the database: MGI 106 GOOI 16

[ukemi]

This is redundant. @dustine32 will make this conversion as part of the import.

[mah11]

In light of @ukemi's comment, we won't worry about it for this ticket, but for future reference:

• What does "pombase via GOC-OWL" mean?

• The GAF (sorry, no GPAD yet) we submit does not contain any has_direct_input relations.

[alexsign]

@ukemi will they disappear from http://www.informatics.jax.org/downloads/reports/mgi.gpa ?

[ukemi]

Only after we finish with the round tripping, but it made sense to us that in cases like this, rather than go in and hand-edit all of our annotations, we would just fix it computationally when the pipeline is in place. @vanaukenk, please correct me if I am mistaken.

[mah11]

What does "pombase via GOC-OWL" mean?

it is data extracted from GO-CAM models converted to gpad.

In that case, the annotations are not assigned by PomBase.

[alexsign]

@mah11 Please ignore my previous comments I was looking at the different pipeline. Actual data is coming form inferences pipeline.

[mah11]

@alexsign - Ah, OK; tho' it still looks like something we can't address directly, since the annotations we submit don't use has_direct_input.

[alexsign]

@mah11 the file I'm getting is form here: http://build.berkeleybop.org/view/GAF/job/gaf-check-pombase/lastSuccessfulBuild/artifact/gene_association.pombase.inf.gaf

[mah11]

Thanks, @alexsign. I guess we'll have to let the relevant people improve that inference pipeline.

[pgaudet]

@mah11 Those come from the ontology. But it's so weird - the inferences state that the process is part of the same process:

UniProt ID	Gene name	GOID	label	Assigned by	extension
O13286	srw1	GO:1905786	positive regulation of anaphase-promoting complex-dependent catabolic process	GOC-OWL	happens_during(GO:0000080),has_direct_input(GO:0005680),part_of(GO:1905786)
Q9Y703	alp31	GO:0007023	post-chaperonin tubulin folding pathway	GOC-OWL	has_direct_input(PomBase:SPBC26H8.07c),part_of(GO:0007023)
O74476	sal3	GO:0006606	protein import into nucleus	GOC-OWL	has_direct_input(PomBase:SPAC24H6.05),part_of(GO:0006606)
P32587	pyp3	GO:0010971	positive regulation of G2/M transition of mitotic cell cycle	GOC-OWL	has_direct_input(PomBase:SPBC11B10.09),part_of(GO:0010971)
P14068	xpo1	GO:0006611	protein export from nucleus	GOC-OWL	happens_during(GO:0071280),has_direct_input(PomBase:SPAC31A2.11c),part_of(GO:0006611)
Q10164	rga2	GO:0035024	negative regulation of Rho protein signal transduction	GOC-OWL	has_direct_input(PomBase:SPAC16.01),part_of(GO:0035024)
O13924	hrk1	GO:0072356	chromosome passenger complex localization to kinetochore	GOC-OWL	has_direct_input(PomBase:SPBC1105.11c),part_of(GO:0072356)
P46595	ubc4	GO:0006511	ubiquitin-dependent protein catabolic process	GOC-OWL	has_direct_input(PomBase:SPBC2A9.04c),part_of(GO:0006511)
Q00619	mam2	GO:0000750	pheromone-dependent signal transduction involved in conjugation with cellular fusion	GOC-OWL	has_direct_input(PomBase:SPCC1795.06),part_of(GO:0000750)
O94537	ppk4	GO:0036498	IRE1-mediated unfolded protein response	GOC-OWL	has_direct_input(PomBase:SPAC22A12.15c),part_of(GO:0036498)
Q10156	lkh1	GO:2000134	negative regulation of G1/S transition of mitotic cell cycle	GOC-OWL	has_direct_input(PomBase:SPBC32F12.09),part_of(GO:2000134)
Q09763	gef1	GO:2000784	positive regulation of establishment of cell polarity regulating cell shape	GOC-OWL	has_direct_input(PomBase:SPAC110.03),part_of(GO:2000784)
O60175	SPBC21H7.06c	GO:0006606	protein import into nucleus	GOC-OWL	has_direct_input(PomBase:SPCC1739.13),part_of(GO:0006606)
P04551	cdc2	GO:0031031	positive regulation of septation initiation signaling	GOC-OWL	has_direct_input(PomBase:SPAC222.10c),part_of(GO:0031031)
Q9Y7J8	SPBC216.01c	GO:0045875	negative regulation of sister chromatid cohesion	GOC-OWL	has_direct_input(PomBase:SPBC26H8.05c),part_of(GO:0045875)
Q9HGN5	SPBC36B7.05c	GO:0071629	cytoplasm protein quality control by the ubiquitin-proteasome system	GOC-OWL	has_direct_input(PomBase:SPAC17G8.12),part_of(GO:0071629)

@cmungall @kltm this seems like a bug ???

[mah11]

From the first few, it looks like our manually curated extensions are being copied over unchanged, even when one of the extensions uses the same GO ID/term as the inferred annotation. E.g. we have these annotations (minimal columns shown):

Gene name	GOID	extension
srw1	GO:1990757	happens_during(GO:0000080),has_input(GO:0005680),part_of(GO:1905786)
alp31	GO:0044183	has_input(PomBase:SPBC26H8.07c),part_of(GO:0007023)
sal3	GO:0061608	has_input(PomBase:SPAC24H6.05),part_of(GO:0006606)

So I'm in the "it's a bug" camp ...