Fix db-xrefs.yaml for IMG

[cmungall]

E.g. http://amigo.geneontology.org/amigo/term/GO:0043770

Goes through a round of redirects and ends up on a login page. What should we do here @monicacecilia?

[monicacecilia]

ack! not good. I logged in but could not find the gene ID this entry is referencing. - I did find 75 other D. renio genes with the word "methyltransferase" in their name. - but I know that this still doesn't help us.

Who/what group added this reference?

What I can recall from the talks at the IMG workshop is that although all information is free-of-charge to the public, a login is required for viz and analysis. I can ask someone in Nikos' group whether/how they allow external referencing to their resources.

[cmungall]

Ah yes, more context was in order. It's actually an xref indicating the source of the GO term:

id: GO:0043770
name: demethylmenaquinone methyltransferase activity
namespace: molecular_function
def: "Catalysis of the reaction: 2-demethylmenaquinone + S-adenosyl-L-methionine = menaquinone + S-adenosyl-L-homocysteine." [IMG:01848]
subset: gosubset_prok
xref: EC:2.1.1.-
is_a: GO:0008168 ! methyltransferase activity

[monicacecilia]

For the record, email to members of IMG team: [...] Today I have a [hopefully] quick question for the IMG team.

We came across a Gene Ontology term on our database with a cross-reference indicating that the source of the annotation is IMG. The term can be seen at http://amigo.geneontology.org/amigo/term/GO:0043770

The source directs users to an IMG page, which after a round of redirects lands them in your login page. http://img.jgi.doe.gov/cgi-bin/pub/main.cgi?section=GeneDetail&page=geneDetail&gene_oid=01848

Our questions: Is it possible for the public to view IMG genes without login? Otherwise, do you have an alternative suggestion for how to reference IMG as a source on GO? We would like to ensure we are able to direct our users to the appropriate place and that GO maintains the most up-to-date ways to access your data. Any help would be greatly appreciated.

By the way, I used my credentials to enter IMG and found out that in the case of this gene, the ID cannot be found. I did find 75 other D. renio methyltransferases, including several of which have GO:0043770 activity annotated to them in GO, but the "demethylmenaquinone methyltransferase activity" function is not annotated to any gene products in the D. renio genome in IMG. Updates to the annotations since the term was originally assigned a number of years back, or a poorly captured gene ID "way back when" could be likely culprits, among other possibilities. I would very much like to dig into the forensics of that with someone from your team later, if there is interest. -- but for now, the question of whether (and how) access to the public is possible is most important. [...]

[cmungall]

did you hear back?

[pgaudet]

We have 143 xrefs to IMG - for info

[pgaudet]

IMG as an 'Integrated Microbial Genomes and Microbiomes' - as far as I can tell if has information about genes and genomes, ie not relevant as definition cross references.

From some archeological work it looks like Jane added these mappings in 2007 (http://wiki.geneontology.org/index.php/Enzyme_Alignment). I think they can be removed and the entry be deleted from the dbxref file.

[pgaudet]

GO:0043811 phosphate:acyl-[acyl carrier protein] acyltransferase activity

• [x] remove IMG:01871
• [x] add +xref: EC:2.3.1.274 +xref: RHEA:42292

---

GO:0043851 methanol-specific methylcobalamin:coenzyme M methyltransferase activity - 0 annotations IMG:01315, MetaCyc:RXN-8096

-> corresponds to EC:2.1.1.246/ RHEA:45208

• [x] merge into GO:1990088 [methyl-Co(III) methanol-specific corrinoid protein]:coenzyme M methyltransferase activity (0 annotations)
• [x] change label to methylcobamide:CoM methyltransferase activity to match UniProt Q48927