Faldo:OneOfPosition add possiblePositions

[Jerven Tjalling Bolleman]

Message Body
Hi All,

There is a git pull request for FALDO at
https://github.com/OBF/FALDO/pull/38

The pull request covers two parts.
* OneOfPosition not nicely modelled
* Probable Positions

Regarding the OneOfPositions the proposal is to add a new predicate

:possiblePosition
rdf:type owl:ObjectProperty ;
rdfs:comment "One of the possible positions listed for a
OneOfPosition element."^^xsd:string ;

Very embarrassingly this is something I have added to the UniProt RDF
from before FALDO was standardized and never realized it was missing
from the manuscript. Luckely my good colleague at EBI, Leyla Garcia,
spotted this. And she was kind enough to open issue
https://github.com/OBF/FALDO/issues/37

I think this part of the pull request can be accepted as is after a
small typo fix.



The seconds part for Probable positions needs a little bit more work. It
is roughly what we had in our Readme for an extension regarding
probable positions. But I think it misses a much need separation between
the prediction/probability and the position itself.

Leyla proposes to my understanding :
example:r a faldo:OneOfPosition ;
faldo:possiblePosition example:1, example:2;

example:1 a faldo:FuzzyPosition ;
faldo:probability 0.1;
faldo:location _:3bb1 .
example:2 a faldo:FuzzyPosition ;
faldo:probability 0.9;
faldo:location _:3bb2 .

While we proposed

example:r a faldo:ProbablePosition ;
probablisticfaldo:posibilities ( _:3bp1 _:3bp2 ) .

example:1 a faldo:ProbablePosition ;
probablisticfaldo:probability 0.1e0 ;
faldo:location _:3bb1 .

example:2 a faldo:ProbablePosition ;
probablisticfaldo:probability 0.9e0 ;
faldo:location _:3bb2 .

Which is similar enough that either one is fine. The set in
possibilities is nice in Turtle but probably a pain in SPARQL practice.

In any case I don't think the probability as a float is _not_ good
enough.
And was wondering what everyone else thinks of something like this

example:r a faldo:ProbablePosition ;
probablisticfaldo:posibilities ( _:3bp1 _:3bp2 ) .

example:1 a faldo:ProbablePosition ;
prob:hasProbability [ a prob:probality ; rdf:value 0.1 ;
prop:caculatedBy program:a ] ;
faldo:location _:3bb1 .

example:2 a faldo:ProbablePosition ;
prob:hasProbability [ a prob:Probality ; rdf:value 0.9 ;
prop:caculatedBy program:a ] ;
faldo:location _:3bb2 .

program:a rdfs:comment "Some function " .



Or a slightly different variety


example:r a faldo:ProbablePosition ;
probablisticfaldo:probalityFunction [ rdfs:comment "Some
function" ] ;
probablisticfaldo:posibilities ( _:3bp1 _:3bp2 ) .

example:1 a faldo:ProbablePosition ;
prob:hasProbability 0.1 ;
faldo:location _:3bb1 .

example:2 a faldo:ProbablePosition ;
prob:hasProbability 0.9 ;
faldo:location _:3bb2 .

In both cases prob is a place holder for an existing probability
vocabulary.
With faldo:ProbablePosition being a rdfs:subClassOf faldo:OneOfPosition
.

Let me know what you all think. Either here or on the issue tracker.

Regards,
Jerven

--
Jerven Tjalling Bolleman
SIB | Swiss Institute of Bioinformatics
CMU - 1, rue Michel Servet - 1211 Geneva 4
t: +41 22 379 58 85 - f: +41 22 379 58 58
Jerven....@sib.swiss - http://www.sib.swiss

[jerven....@sib.swiss]

The first part was accepted today