current F-P inferences gaf

[ValWood]

where are we getting this file from currently?

here?
http://build.berkeleybop.org/view/GAF/job/gaf-check-pombase/gene_association.pombase.inf.gaf

https://curation.pombase.org/pombe/stats/annotation

[ValWood]

Once we have established if we are getting the file from the correct place, can we check if we still see this drop in numbers? This was the original issue.

@mah11
That said, there is still a problem with inf.gaf - it's still a lot smaller than the one I get running owltools (Jenkins) GAF checks locally, a la geneontology/go-site#2226

[kimrutherford]

We're getting it from here:
http://build.berkeleybop.org/job/gaf-check-pombase/lastSuccessfulBuild/artifact/gene_association.pombase.inf.gaf

I don't know if it's the right one. I'm very confused by it all. That file is updated daily which I guess is a good sign.

[ValWood]

So I don't think we are getting the inferred gaf from the above location.
I did a spot check.

http://build.berkeleybop.org/view/GAF/job/gaf-check-pombase/gene_association.pombase.inf.gaf

we don't get these annotations:
PomBase SPCC663.12 cid12 GO:0000775 PMID:25274039 IMP C poly(A) polymerase Cid12 csp10|csp8 protein taxon:4896 20141110 GOC
PomBase SPCC663.12 cid12 GO:0034507 PMID:25274039 IMP C poly(A) polymerase Cid12 csp10|csp8 protein taxon:4896 20141110 GOC

but we do get this one:

PomBase SPCC737.09c hmt1 GO:0071994 PMID:7876244 IMP P vacuolar phytochelatin transmembrane transporter Hmt1 SPCC74.08c protein taxon:4896 20120531 GOC

odd?

[ValWood]

The one I pointed to is also updated.
!gaf-version: 2.0
!
! Date: 2018/01/22

I don't know what the difference is between these 2 files.

So we will work on the assumption that the one you are using is the correct one. Midori does this file still show reduced numbers?

I will do some spot checks on this one.

[ValWood]

OK both of these files are the same.

[ValWood]

I will list any problems I find with the file here. I don't thhink this file is built from current data.

It contains this annotation:

PomBase SPCC74.06 mak3 GO:0034599 PMID:20919928 IC GO:0032874 P

this must be generated from an MF IC annotation, but as far as I can see pombe mak3 does not have any IC annotation ( I checked in artemis to be sure that it is not filtered, I also checked amigo to be sure it isn't a submission timing issue)

[ValWood]

Does anyone understand this row:
this is a process ID, but in the aspect column it has F, for function
How does the loader handle that?

PomBase SPBC2F12.13 klp5 GO:1990758 PMID:21664573 IC GO:0005515 F kinesin-8 family plus-end microtubule motor Klp5 sot1 protein taxon:4896 20161018 GOC-OWL part_of(GO:1990758)

[mah11]

also see geneontology/go-site#524. it's a ghastly mess.

[kimrutherford]

I'm confused again. Is this the right file?:
http://build.berkeleybop.org/job/gaf-check-pombase/lastSuccessfulBuild/artifact/gene_association.pombase.inf.gaf

In geneontology/go-site#524 (comment) you mention a new location. I thought the URL we're using is the new location. Is there a new new location?

[kimrutherford]

Is there a new new location?

I've read the GO issue more carefully now. Is this the new new location?:
http://snapshot.geneontology.org/annotations/pombase-prediction.gaf

I compared that file with the one we're using. They are very similar except that column 9 ("Aspect") is different. That's alarming. Although it shouldn't cause problems for PomBase as we ignore that column.

Here's an example:

gene_association.pombase.inf.gaf:

PomBase  SPAC1006.07  tif1  GO:0006412  PMID:10462529  IMP  P  translation init...

pombase-prediction.gaf:

PomBase  SPAC1006.07  tif1  GO:0006412  PMID:10462529  IMP  C  translation init...

The only other different between the two files is that gene_association.pombase.inf.gaf has one extra annotation:

PomBase  SPBC18H10.16  can1  GO:1903826  PMID:18357653  IMP  X  plasma membrane arginine/lysine transmembrane

[ValWood]

Don't worry yet, if we don't include the file I don't think we will lose so much.
We can wait until resolved....

I'll check the file contents when I get the go ahead. There appears to still be a problem.

[mah11]

I've read the GO issue more carefully now. Is this the new new location?:
http://snapshot.geneontology.org/annotations/pombase-prediction.gaf

Yes, this is the new new location, and what GO is now billing as the correct one. The billing is still fairly subtle, and Tanya-at-TAIR has asked them to make an announcement when they've got the bugs fixed. Oh yeah, as your snippet shows, they haven't fixed all the bugs yet, so as Val says, no rush.

[ValWood]

@kimrutherford are we loading any of the process-from function prediction gafs?
I didn't think we were, but we are definitely getting them from somewhere...

because I see some in here:
http://curation.pombase.org/dumps/latest_build/logs/log.2018-02-25-20-00-50.excluded_go_terms_softcheck

SPAC21E11.06.1 GO:0065009 regulation of molecular function GOC
SPAC24H6.09.1 GO:0065009 regulation of molecular function GOC
SPAC343.14c.1 GO:0065009 regulation of molecular function GOC
SPAC4D7.09.1 GO:0065009 regulation of molecular function GOC
SPBC19G7.08c.1 GO:0065009 regulation of molecular function GOC

it's not a big deal, because they will be subject to our normal filtering I was just wondering where we get them from (probably from the GOA file?)

[ValWood]

I see it in goa so I think that's it. can you confirm.

[kimrutherford]

We load:
http://build.berkeleybop.org/job/gaf-check-pombase/lastSuccessfulBuild/artifact/gene_association.pombase.inf.gaf

I haven't changed the loader to use the new location yet:
http://snapshot.geneontology.org/annotations/pombase-prediction.gaf

But the two files are nearly identical.

The annotations you list in your comment are in gene_association.pombase.inf.gaf

[ValWood]

Yep the new file looks much better now, but we'll wait to switch. I'd still like to see better redundancy filtering (because it is difficult for us to discard experimental annotations, even when less specific than ours from the same source), but we get
rid of most ... after our various filtering we only retain 64 annotations, total....

[kimrutherford]

I've changed the load script to get the inferred GAF file from the new location: http://snapshot.geneontology.org/products/annotations/pombase-prediction.gaf

[ValWood]

Ok when the issues are solved with the "occurs in" we can reassess whether to import the component annotations too.

phew...

(I'll check tomorrow though anyway)

[kimrutherford]

Can this be closed now?

[ValWood]

I think so....