uPheno 2: Alpha review - Githubissues

matentzn commented 4 years ago

Floating terms
- FIXED anterior_to posterior_to (ZP) and dorsal_to, in left_side_of, in_right_side_of,
- FIXED http://purl.obolibrary.org/obo/CARO_0010004 (missing import)
- FIXED http://purl.obolibrary.org/obo/CHEBI_316698 (bug in FYPO, fixed)
- FIXED http://purl.obolibrary.org/obo/ENVO_01001784 (seems to come from ENVO? No term exists with that id..)
- FIXED http://purl.obolibrary.org/obo/GO_00000236 (bug in FYPO, fixed)
- FIXED http://purl.obolibrary.org/obo/ZP_0103989 floating? - definitely should not
- TODO http://purl.obolibrary.org/obo/UPHENO_0031759 (-> floating because it is only defined, and the bearer is not in the ontology!)
- TODO http://purl.obolibrary.org/obo/NCIT_C3824 (the missing class from the above case)
- FIXED NCBI taxon terms:
- TODO All Species specific phenotypes are rendered under owl:Thing, but that is a bug I have now reported to Protege.
  - Maybe I could simply remove those restrictions in a post processing script
- PLANA terms (reported to sofia)
  - FIXED http://purl.obolibrary.org/obo/PLANA_0007501
  - FIXED http://purl.obolibrary.org/obo/PLANA_0000077
Obsolete terms with references:
Weird labels
- TODO http://purl.obolibrary.org/obo/UPHENO_0075938
  - “owl: Thing in label"
    - Pattern domains should be changed.
Missing axioms
- TODO Upheno phenotypes should probably get the ‘has phenotype affecting` axioms as well.
- TODO DPO EQs are not preserved
- TODO PHIPO EQs are not preserved
REVIEW are only anatomy and biological process or all vars instantiated?
- if all, remove option from config
TODO There is an abnormal anatomical entity (II and IIPO) version!

matentzn commented 4 years ago

Ameded:

[X] Need to remove "has phenotype affecting OWL:thing" assertions (fix in relation module)
There are some unconnected classes in the final artefact, like
- http://purl.obolibrary.org/obo/CHEBI_1086440
- http://purl.obolibrary.org/obo/UBERON_0004355
the NBO behavior phenotype branch is pulled in because of HPO classes inhering in them.
All lesion terms are floating.. Need to be careful not to destroy the classification with the blacklist!
We need the phenotype root annotation for OLS
uPheno has now a whooping 8 cell classes:

There are still a significant number of references to obsolete classes

matentzn commented 4 years ago

Should there be reasoning after the SS components are merged in - this may confuse users of them.

matentzn commented 4 years ago

Need to be carefully with the assumption that root phenotypes are abnormal - in the case of MP, they are not!

matentzn commented 4 years ago

Why does http://purl.obolibrary.org/obo/PHIPO_0000274 have no valid EQ? -- because I did not yet dare to overwrite the matches! Should do that!

matentzn commented 4 years ago

upheno:phenotype and upheno:phenotype by species should be unqualified phenotypes. We only care about abnormal phenotypes

obophenotype / upheno-dev

uPheno 2: Alpha review #18