Open pgaudet opened 1 month ago
Hi @raymond91125 Can you take this ticket?
These terms have a RHEA in the definition xref that is not in the general cross ref. The task is to check that it seems right, and add it, if appropriate as skos:exactMatch.
Thanks, Pascale
Sure. But the table has a few 'false positives' if I understand what you asked for correctly. @pgaudet
Below are action items:
I tried using a python script (oaklib) to exhaustively search for RHEAs in definition but not in general xrefs. The script is incomplete but it did whittle down to 22 candidates. I manually checked them and found 3 valid cases that need attention.
[x] GO:0043883 RHEA:46276 There appears to be a typo in the general xref RHEA:46277. Change it to RHEA:46276.
[ ] GO:0090555 RHEA:36440 Xref in the definition, RHEA:36440 is a left-to-right rxn (from in to out) whereas the textual definition states from exoplasmic to cytosolic leaflet of a membrane. Therefore it seems the assignment is incorrect. Perhaps we should change it to RHEA:36439, the directionless form. In the general xref, there is RHEA:66132, which appears to be of an identical rxn as that of RHEA:36439. Although the UniProKB annotations are probably non-overlapping (RHEA:66132 has 7,282 proteins; RHEA:36439 has 13). @pgaudet I am not sure about this one, could you take a look to see if I missed something? Thanks.
[x] GO:0140859 RHEA:35811 Definition xref is RHEA:35811 whereas general xref is RHEA:35407, which is the general rxn of RHEA:35811. Thus, it makes sense to use RHEA:35407 in both places.
Thanks for this @raymond91125 For GO:0090555 Did you see the comment in the parent term flippase activity
Nomenclature note. Flippases and floppases are ATP-dependent transbilayer lipid translocators. According to an extensively used, though not universal, nomenclature, they catalyze lipid transfer towards the inward monolayer (flippases) or towards the outward monolayer (floppases). Scramblases are ATP-independent, non-selective, inducing non-specific transbilayer movements across the membrane. The direction of the translocation should be taken into account for annotation (from the cytosolic to the exoplasmic leaftlet of a membrane).
So: since RHEA does not assign which side is in out out, these mappings cannot be exact WRT to the directionality of flippases and floppases. Therefore, I suggest we add RHEA:66132 as broad match to ATPase-coupled intramembrane lipid transporter activity and all its children.
Does that sound OK to you?
Thanks, Pascale
@raymond91125
You write
I tried using a python script (oaklib) to exhaustively search for RHEAs in definition but not in general xrefs. The script is incomplete but it did whittle down to 22 candidates. I manually checked them and found 3 valid cases that need attention.
What about the other 19?
The other 19 are false positives that my incomplete script is not able to eliminate. So I had to manually do it. The reasons for the incompleteness are silly and technical.
So: since RHEA does not assign which side is in out out, these mappings cannot be exact WRT to the directionality of flippases and floppases. Therefore, I suggest we add RHEA:66132 as broad match to ATPase-coupled intramembrane lipid transporter activity and all its children.
RHEA:66132 is specifically about one type of lipid phosphoethanolamine, not general enough, I think. According to PMID:20043909, flippases are P-type ATPases (I think therefore EC:7.6.2.1 xref GO:0140326) whereas floppases are ABC-type ATPases (possibly EC:7.6.2.-).
Yesterday, I did miss the important difference between RHEA:66132 and RHEA:36439, the side where ATP hydrolysis takes place. Therefore, I think RHEA:36440 was incorrectly applied to GO:0090555 phosphatidylethanolamine flippase activity. This is confirmed by UniProt https://www.uniprot.org/uniprotkb/P39524/entry, which annotates DRS2, the gene studied in GO:0090555 definition reference, with RHEA:66132.