Closed ukemi closed 1 year ago
Here is the list of CHEBI classes in Reactome models that do not appear to be in NEO: nonresolving_chebi.txt
This list was made with a SPARQL query to check for any classes having a CHEBI:
prefix but no subClassOf
connection to chemical entity CHEBI:24431.
Some of these are subClassOf CHEBI:36342 "subatomic particle", which is a sibling of 'chemical entity'. In fact, one of them is 'subatomic particle' itself. Here is the list: CHEBI:10545 CHEBI:30212 CHEBI:33233 CHEBI:33252 CHEBI:36338 CHEBI:36340 CHEBI:36341 CHEBI:36342 CHEBI:36347
The remainder are classed under 'chemical entity' (at least in CHEBI).
Interesting. Thanks @nataled. I suspect the remainder just haven't been imported yet.
Here's an example where we have annotated an electron as a small-molecule participant in a reaction: BH4 is oxidised to the BH3 radical during the eNOS catalytic cycle. In this case, before figuring out with Jim, Chris, et al. how to make a ChEBI entry for electrons, we should also think about the chemistry and ask exactly what biochemistry we a trying to capture here, whether naked electrons are really a part of it (is this a curator hack for an unknown or anyway unspecified electron acceptor?) and if so how to annotate it in an orderly way.
We also have things like photons activating retinol - same questions.
Looking at the models, these don't seem to be the issue. I think I have determined what is causing most of the Shex problems. There is no shape for small molecule regulators.
@dustine32 and @deustp01. How do we want to handle ts ticket? If it was meant to just create the report, then I think it should be closed. If it was meant to cover the resolution of the issues, then it should remain open and we should create a specific list of to do items. What do you guys think?
Thinking that tickets should set out well-delimited problems / chores, it might be best for sanity to close this ticket - diagnosis is done - and create a new ticket that refers to this one for the clean-up?
Can do. I will make new tasks as tickets and then close this one. We can double check on Monday.
We would like to have a report that lists ChEBI identifiers that fail to be recognized during the Release 83 import. https://github.com/geneontology/pathways2GO/issues/221 If possible, the report should also include the model identifier and whether or not the model fails the ShEX.