|
From: Chris M. <cjm...@lb...> - 2014-08-27 21:38:55
|
There is a bit of redundancy between CHEBI and PR. For example
CHEBI:
is_a CHEBI:16670 ! peptide
is_a CHEBI:25905 ! peptide hormone
is_a CHEBI:65311 ! corticotropin-releasing hormone
is_a CHEBI:65312 ! corticotropin-releasing hormone
(human)
PR:
is_a CHEBI:23367 ! molecular entity
is_a PR:000018263 ! amino acid chain
is_a PR:000000001 ! protein
is_a PR:000005867 ! corticoliberin
is_a PR:P06850 ! corticoliberin (human)
These are equivalent (but note the classification difference, in that
CHEBI doesn't classify this as a protein).
Why put proteins in CHEBI? I mean do you intend to put TTN in there in
order to break everyone's autocomplete-by-synonym capabilities?
There are some benefits - e.g. from a GO perspective we like having the
classification under peptide hormone, which we don't get from PR. But
there may be better ways to achieve the same thing.
We are making use of both PR and CHEBI for axiomatization in GO. Ideally
there would be no overlap and an obvious choice of which class in which
ontology to use. Failing that, or as an interim step, it would be
extremely useful to have OWL axioms that connect the two ontologies,
e.g. in a bridging ontology. Does such an artefact exist already?
And a related minor issue that would make things a lot easier for our
automation processes: can CHEBI and PR agree on leading uppercase
conventions for proteins. CHEBI says "Elastin", PR says "elastin". I
prefer the lowercase, it's more in line with OBO conventions. |