aortic (valve) stenosis

[GoogleCodeExporter]

MP has:

[Term]
id: MP:0010463
name: aorta stenosis
def: "diffuse constriction or narrowing of the aorta" [MGI:csmith]
****comment: Do not confuse with aortic valve stenosis.
synonym: "aortic stenosis" RELATED []
is_a: MP:0000272  ! abnormal aorta morphology
is_a: MP:0006135  ! artery stenosis

HP has:

[Term]
id: HP:0001650
name: Aortic valve stenosis
def: "The presence of a stenosis (narrowing) of the aortic valve." 
[HPO:probinson]
comment: Aortic stenosis can lead to a pressure gradient between the left 
ventricle and the aorta and may result in left ventricular hypertrophy and 
decreased left ventricular compliance.
*****synonym: "Aortic stenosis" EXACT []
synonym: "Valvular aortic stenosis" EXACT []
is_a: HP:0001646  ! Abnormality of the aortic valve

The HP exact syn contradicts the MP comment.

Note there is a separate term: MP:0006117 ! aortic valve stenosis

Recommendations:

1. HP demotes the syn from EXACT OR adds aortic stenosis as a separate term
2. (optional) the MP comment is represented explicitly as an OWL disjointness 
axiom

Original issue reported on code.google.com by cmung...@gmail.com on 31 Oct 2013 at 12:38

[GoogleCodeExporter]

This usage is standard in the medical community, and thus I think it is 
appropriate to have this as an exact synonym. Nobody should be using text 
mining to match between HP and MP (hmm....., at least not after we finish all 
of the logical definitions :-0).

Original comment by peter...@gmail.com on 16 Nov 2013 at 9:51

[GoogleCodeExporter]

OK, that suggests the MP should have a different primary label for the grouping 
class.

I agree text mining is best avoided. But unless we are striving for 100% 
sensitivity then it's the best strategy to complement the logical defs. There 
are 766 pairs of matching labels (many of which are problematic of course) for 
which there is no logical def for at least one of the partners.

Many of these are simply too hard to give a meaningful OWL logical definition 
to, so the best we can do is a verified direct equivalence axiom between the 
two ontologies.

Our strategy is much better than most other text mining efforts. We have a 
strategy for overriding spurious matches (the disjoints file), and of course 
having the logical definitions override.

Original comment by cmung...@gmail.com on 18 Nov 2013 at 1:39

[GoogleCodeExporter]

An example of a difficult class to define using an OWL pattern using PATO:

[Term]
id: MP:0004110
name: transposition of great arteries
namespace: MPheno.ontology
alt_id: MP:0006127
def: "cardiovascular structural anomaly in which the aorta arises from the 
right ventricle while the pulmonary artery arises from the left ventricle" 
[MGI:monikat]
subset: CvDC_Terms  ! NHLBI Cardiovascular Development Consortium (CvDC) Terms
synonym: "TGA" EXACT []
synonym: "discordant VA connections" EXACT []
synonym: "transposition of great vessels" RELATED []
is_a: MP:0010426  ! abnormal heart and great artery attachment

But I suspect the equivalence to the identically labeled HP:0001669 ! 
Transposition of the great arteries is correct.

In fact I think it *is* possible to do good OWL definitions for defects such as 
this, but it involves creating an abnormal anatomy ontology -- 
https://github.com/cmungall/vmao -- but is it worth this extra effort if we can 
make a direct link and infer the basic anatomical axioms from the parent....

Original comment by cmung...@gmail.com on 18 Nov 2013 at 1:46