Q9P6P9 pyruvate dehydrogenase (acetyl-transferring) kinase activity has a maapping to
GO:0006006 glucose metabolic process
but pyruvate can come from other places so this mapping causes a TPV (is too far upstream)
It's even worse. This reaction activates pyruvate dehydrogenase, which in turn catalyzes the conversion of pyruvate to acetyl CoA. Acetyl CoA has many fates, but in mammals it is explicitly UNable to be converted in net amounts to glucose.
The GO annotation to glucose metabolic process is coming from the UniProt keyword 'glucose metabolism' that was assigned to the pombe protein by similarity to the cerevisiae one (see http://www.uniprot.org/uniprot/Q9P6P9\). It's not clear how the UniProt curators got the link to glucose metabollism, but the phrase I keep finding when searching for this complex is "The resulting inhibition of Pyruvate Dehydrogenase prevents muscle and other tissues from catabolizing glucose and gluconeogenesis precursors. "
Regardless, the GO term pyruvate dehydrogenase (acetyl-transferring) kinase activity is not related to GO:glucose metabolic process in the ontology, so there is no TPV in GO.
Let me know if I'm misunderstanding.
I'm talking nonsense its not causing a TPV. Its just a mapping to a process which is too far upstream.
OK, that's alright then.
It's not a problem with the UniProtKW2GO mapping - KW:glucose metabolism is mapped to GO:glucose metabolic process, which is correct. It's just that UniProt has associated the KW glucose metabolism with this entry.
Just catching up on some SF, is the result of this issue that you would like the glucose metabolism keyword removed from Q9P6P9?
yep, its downstream.
I noticed in uniprot this also has
two-component signal transduction system (phosphorelay) GOC
signal transduction by phosphorylation GOC
do you know where these are generated from ? we don't have them in PomBase yet, but these aren't correct either
OK, I'll request the removal of the glucose metabolism keyword.
The GOC annotations to two-component signal transduction system (phosphorelay) and
signal transduction by phosphorylation were being created from inter-ontology links from the term GO:0000155 two-component sensor activity, which had been applied via InterPro2GO mappings from IPR005467. These GO mappings have since been removed from the InterPro entry and we are no longer getting these annotations in our database. You are seeing them in UniProt because of the different update cycles, they will eventually disappear.
Hope that helps,
Ah that accounts for it. thanks
here, we are talking about a central switch in glucose metabolism, one that decides whether cells run on (anaerobic) glycolisis, or burn their glucose via mitochondrial respiration (the Warburg effect). There's a wealth of old and new literature on the subject.
Inhibition of pyruvate dehydrogenase by pyruvate dehydrogenase kinases (yeast, human, etc) regulates glucose metabolism (see PubMed 21076574, 12676647 , 22195959, 22360721, 16606348, etc). This is really a fundamental metabolic switch, and we should show this information.
Seeing the key role of the pyruvate dehydrogenase kinases, I'd rather keep the UniProt KW.
From your description, it sounds like _regulation_ of glucose metabolism (corresponding to GO:0010906) would be a better annotation for pyruvate dehydrogenase. You would of course know better that I what KW changes this would entail.
yep, I was going to suggest the same, it should be annotated to regulation.
Problem is that UniProt keywords are not (and don't aim to be) as granular as GO terms are. We have no possibility to distinguish between _regulation of_ and simply glucose metabolism just via the KW mapping...
Then, unfortunately, you are forced to choose between having the KW-based annotations offer incomplete information or potentially misleading information.
For myself (and for PomBase), I would err on the side of incompleteness, since we have plenty of manual curation routes that will fill in what is missing from UniProt KW mappings. But it will be very useful for us to know if we will have to do some filtering at our end to achieve that.
"Glucose metabolism" is a child of "Regulation of...), so the information given via the UniProt KW is less granular, but it is correct. And we can use GO tterms for the full details.
> "Glucose metabolism" is a child of "Regulation of...
No, it comes closer to being the other way round, although the "parent/child" nomenclature doesn't work nearly as well for regulation (and several other relationship types) as it does for is_a or part_of.
At present there are no GO terms that explicitly mean "a process or regulation thereof".
This particular instance isn't an issue for us I have filtered this mapping, but it might cause problems for others. However, this practice might start to cause us some problems as we refine the annotations to distinguish between regulation / and directly involved in.
I'm going to ask Rachel and David how we should proceed.
Mapping is still present will resend to UniProt
Q9P6P9 still has the keyword 'glucose metabolism'. Forwarded to UniProt.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.