Closed ValWood closed 11 months ago
The spt4 gene (and therefore the Q9P7K8 protein) is currently getting the term GO:0032784 (regulation of DNA-templated transcription, elongation) because it was suggested in a 2001 paper (https://europepmc.org/article/MED/11182892) that this protein regulated transcription. If this information is now obsolete we will change our ARBA rule accordingly.
There will be a lot of problems with "regulation of" ....terms generally because biologists use the phrase "regulation" often to mean "affects". In GO we use it in the strictest sense to mean "control"
From the GO perspective this gene is part_of the transcription elongation process (specifically "co-transriptional chromatin reassembly"), yes it 'affects' this process, but it isn't controlling it.
So the information in the paper is still correct, but we are looking through a slightly more refined lens with more detailed mechanistic information.
Actually I do not see the paper state that these are involved in regulation of elongation. The state that they are "Spt4, Spt5, and Spt6 have characteristics of general transcription-elongation factors"
Thank you for the explanation. In the paper I could see that it stated: "These new studies, taken together with earlier work, portray Spt4, Spt5, and Spt6 as general transcription-elongation factors, controlling transcription both positively and negatively in important regulatory and developmental roles".
My explanation is not quite as detailed as it needs to be.
So when a gene product operates within a pathway (such as Spt4, Spt5, and Spt6 as general transcription-elongation factors) we make the connections between them if they operatte in a negative or positive direction. We don't use "regulation of process" for regulation of the activity of one gene product by another within a pathway. These connections are made by connecting activities, not by a "regulation of biological process" term.
So for example if within the "transcription elongation" pathway A has_ a_positive_effect_on B we would capture
A has_ a_positive_effect_on B as part of "transcription elongation"
"regulation of transcription elongation" would only be used for "upstream regulation" i.e. to begin the process or alter its rate. Otherwise, every single gene product would be regulating the process it is involved in , in some way. We now know that spt6 is involved in reastablishing chromatin that was disrupted by RNA polymerase to prevent cryptic trancription. It's role is not directly to 'regulate' elongation although it has a positive effect on this process (or a negative effect when mutated), and is considered as part of the process of transcription elongation.
@pgaudet @thomaspd have I explained this corrrectly from the GO perspective?
Yes you have - how authors use 'regulates' or 'regulation' is often quite different from the GO definition, which is a lot narrower - a regulator should act directly on a protein in a way that is rate-limiting on the pathway.
Thanks, Pascale
I see. Thank you so much for the detailed explanation. I will discuss this internally with my team and get back when we have actionable points.
Hello, This ARBA rule stopped existing, probably due to updates on Swiss-Prot entries. But the GO term GO:0032786 (positive regulation of DNA-templated transcription, elongation) is still being propagated through InterPro family IPR009287 (https://www.ebi.ac.uk/interpro/entry/InterPro/IPR009287/), so I will send an email to InterPro for them to stop associating this GO term to that family.
Removing ARBA, tagging InterPro
Here the. same applies to GO:0006355 | regulation of DNA-templated transcription | IEA with ARBA00028138 Spt4 https://www.pombase.org/gene/SPBC21C3.16c
I cannot see this GO term GO:0006355 in this protein (https://rest.uniprot.org/uniprotkb/Q9P7K8.txt). Am I missing something? Also, it would be easier to determine which InterPro signature this term should be removed. This is because ARBA (https://www.uniprot.org/arba/ARBA00028138) is touching multiple signatures.
I could see it here https://www.pombase.org/gene/SPBC21C3.16c yesterday, but I filtered it so it has gone today. There was a new GO release a couple of days ago so maybe this was already cleaned up. There has been quite a bit of effort to remove the "regulation" terms when things are "part_of"
Oddly there is a PAINT annotation in AmiGO https://amigo.geneontology.org/amigo/gene_product/PomBase:SPBC21C3.16c from PANTHER:PTN000306261 I will check that that is removed.
I see it in QuickGO https://www.ebi.ac.uk/QuickGO/annotations?geneProductId=Q9P7K8&aspect=biological_process I don't know why it isn't in the Swiss-Prot entry?
Sorry Val, my mistake - Q9P7K8 in fact has the GO term GO:0006355. Yes, UniProt does not have the same data as QuickGO, as we filter out the less relevant GO terms and only keep the most important ones. But the complete list of GO terms are always present in QuickGO. For this protein, the GO term was also added by the InterPro signature IPR009287 (which we already reported to InterPro).
@sarach06 @blazaropinto Can you please have a look at IPR009287?
Hi Pascale, Thank you very much. I had a look at this family last week, when Pedro sent and email. Ther GO term has been removed, and this will be visible in InterPro 97.0. Best wishes, Sara
Thanks @sarach06 !
Looks like everything is addressed and annotations will be fixed when the updates trickle through the various pipelines.
https://www.pombase.org/gene/SPBC21C3.16c GO:0032784 | regulation of DNA-templated transcription, elongation | IEA with ARBA00027284
this is now annotated to
co-transcriptional chromatin reassembly
co-transcriptional chromatin reassembly not to regulation of elongation