Addresses #2786 NTR CD57-positive enterocyte #2847
Conversation
Fixes #2786 NTR CD57-positive enterocyte
|
#gogoeditdiff |
Here's a diff of how these changes impact the classified ontology (on -base file):Ontology comparisonLeft
Right
Ontology importsOntology annotationsCCL3-positive alveolar macrophage
|
Here's a diff of your edit file (unreasoned)Ontology comparisonLeft
Right
Ontology importsOntology annotationsCD57-positive enterocyte
|
|
CellMark comments have been imported when refreshing the imports, causing that some terms have two comments and failing the QC (I might be wrong on the assessment, please correct me if I am wrong). We might need a different AP to avoid having this problem. |
Your assessment looks correct to me. Looking at the CellMarks comments:
It seems to me that such an information should rather belong to a dedicated annotation. Using If we can’t get the CellMarks folks to change that we might have to strip |
The idea was to add markers and confidence scores in human readable form to go along with a formal specification - including marker IRIs and confidence scores as floats. The one comment restriction is an annoying OBO legacy. My ideal solution would be to add SPARQL update to the pipeline to concatenate comments. @gouttegd - what's your opinion on this option? (I have a feeling you might object.) |
I am certainly not very fond of it. That being said, it’s not like I had anything much better to propose… (A truly ideal solution for me would be, either: (1) use dedicated annotation properties for confidence scores and marker IRIs, and leave it to the ontology browser(s) to display those informations in a human-friendly form, or (2) get rid of the OBO format and its limitations entirely. Neither of which are really under our control.) But adding a step in the pipeline to merge comments could be quite tricky as it involves comments spread over several files… How about merging comments, but only for OBO artefacts? |
Works for me. (Although breaking OBO<->OWL round-tripping makes me a bit queasy) |
Pretty sure that horse has already left the barn… CL contains axioms that are not representable in OBO and that are rendered in the |
Use a new command in the Uberon ROBOT plugin to produce "customized" OBO artefacts in which: * all "untranslatable" OWL axioms (owl-axioms tag) are stripped (they were already stripped before, this just changes the way we do it); * all GCI axioms are stripped (#2856); * when a class has several rdfs:comment annotations (which is not allowed in OBO), they are merged into a single annotation (#2847).
Use a new command in the Uberon ROBOT plugin to produce "customized" OBO artefacts in which: * all "untranslatable" OWL axioms (owl-axioms tag) are stripped (they were already stripped before, this just changes the way we do it); * all GCI axioms are stripped (#2856); * when a class has several rdfs:comment annotations (which is not allowed in OBO), they are merged into a single annotation (#2847).
Fixes #2786 NTR CD57-positive enterocyte