Investigate differences between old and new uberpheno production strategies

sivarajankumar / phenotype-ontologies

Automatically exported from code.google.com/p/phenotype-ontologies

0 stars 1 forks source link

Investigate differences between old and new uberpheno production strategies #39

Closed GoogleCodeExporter closed 8 years ago

GoogleCodeExporter commented 8 years ago

See attached DIFF file.

Methods:

1. downloaded build 182 of obo-pipeline uberpheno 
http://compbio.charite.de/hudson/job/hpo.ontology.uberpheno/182/
2. Removed redundant links:
{{{
owltools crossSpeciesPheno.obo --reasoner elk --assert-inferred-subclass-axioms 
--removeRedundant -o -f obo crossSpeciesPheno-nr.obo
}}}
3. Removed grouping classes from owl build:
{{{
obo-grep.pl --neg -r 'id:.*PHENOTYPE' vertebrate-merged-simple.obo  > z.obo
}}}
4. Compared using a simple logical diff (isa tags only)
{{{
obo-simple-diff.pl -l crossSpeciesPheno-nr.obo z.obo > DIFF
}}}

Original issue reported on code.google.com by cmung...@gmail.com on 9 Jul 2014 at 11:48

Attachments:

DIFF

GoogleCodeExporter commented 8 years ago

Note:

a "-" shows something "lost" from the old obo pipeline
a "+" shows something gained in the new

(we expect to see gains of grouping classes, but these can be largely ignored)

Original comment by cmung...@gmail.com on 9 Jul 2014 at 11:49

GoogleCodeExporter commented 8 years ago

Here is something lost:

{{{
[Term]
id: HP:0100011 ! Scleral schwannoma
-is_a: HP:0000591 ! Abnormality of the sclera
+is_a: UBERON:0001773PHENOTYPE ! sclera phenotype
}}}

It's not totally clear what the difference in axioms are that allows this to be 
inferred in the old pipeline.

Either way, this is not asserted in HP. Should it be?

Original comment by cmung...@gmail.com on 9 Jul 2014 at 11:51

GoogleCodeExporter commented 8 years ago

For a fair comparison, could you re-generate the "DIFF"-file using the latest 
Uberpheno generated with the old pipeline? 
(http://compbio.charite.de/hudson/job/hpo.ontology.uberpheno/190/)

Original comment by dr.sebas...@gmail.com on 23 Jul 2014 at 9:08

GoogleCodeExporter commented 8 years ago

Versions still not quite aligned but closer

Original comment by cmung...@gmail.com on 25 Jul 2014 at 6:51

Attachments:

DIFF

GoogleCodeExporter commented 8 years ago

OK, I can see where we're missing some inferences with the new pipeline, and 
it's todo with entailment of abnormality, as you pointed out earlier Sebastian.

Original comment by cmung...@gmail.com on 25 Jul 2014 at 6:56

GoogleCodeExporter commented 8 years ago

OK, just to be extra-confusing, I accidentally entered comments and commit 
messages to this issue:
https://code.google.com/p/phenotype-ontologies/issues/detail?id=27

What were intended to go here.

I added an axiom to extra that PATO:deviation from normal is a subclass of 
has-abnormal. This restored the desired inferences. See the diff in issue 27.

Original comment by cmung...@gmail.com on 28 Jul 2014 at 9:00

GoogleCodeExporter commented 8 years ago

This screenshot illustrated how the additional axiom works

Original comment by cmung...@gmail.com on 8 Aug 2014 at 3:38

Attachments:

[Screen Shot 2014-08-08 at 8.35.51 AM.png](https://storage.googleapis.com/google-code-attachments/phenotype-ontologies/issue-39/comment-7/Screen Shot 2014-08-08 at 8.35.51 AM.png)

GoogleCodeExporter commented 8 years ago

Need to do the same for process qualities

Original comment by cmung...@gmail.com on 15 Aug 2014 at 11:28

GoogleCodeExporter commented 8 years ago

This issue was closed by revision r826.

Original comment by cmung...@gmail.com on 15 Aug 2014 at 11:38

Changed state: Fixed