[Obo-relations] Incorporation of RO relations into BFO, and future development of RO
Chris Mungall
As many of you are aware, RO has been somewhat static for a while.
There have been a number of proposed extensions floating around, but
nothing official. Part of the problem has been the impedance mismatch
between the notion of type-level relations and the corresponding
treatment in OWL. In addition, many of the original instance level
relations are applicable outside biology and are more suited to an
upper level ontology such as BFO.
We propose that the next version of BFO will contain a minimal set of
instance level binary relations -- this set will overlap with the set
of instance level relations used in the original RO. These relations
will have URIs of the form http://purl.obolibrary.org/obo/BFO_nnnnnnn
(in the obo-format version, these will be rendered as BFO:nnnnnnn).
BFO will likely include the follow relations, with the following IDs/
URIs:
BFO_0000050 part of
BFO_0000051 has part
BFO_0000056 participates in
BFO_0000057 has participant
BFO_0000062 preceded by
BFO_0000063 precedes
BFO_0000060 immediately preceded by
BFO_0000061 immediately precedes
(and will most likely include additional ones too)
The corresponding older URIs with the prefix http://obofoundry.org/ro.owl#
would be mildly deprecated (i.e. they would not be rendered
obsolete, but ontologies would be encouraged to switch over). The new
BFO relations are unambiguously instance-level binary relations, and
the definitions and comments would reflect this. This also holds for
the obo-format version of these relations (previously the plan was to
have distinct type-level versions of these relations specified in obo-
format, but this has been abandoned in order to make the translation
between obo and owl simpler).
For example, the metadata for BFO_0000050 would be something along the
lines of:
definition: "part of is a primitive reflexive, transitive relation,
holding either between two processes or two continuants."
example: "every cell nucleus is part of some cell"
BFO will also likely include certain relations that were in various
proposed extensions to the RO, but not in the original version,
including, but no limited to:
BFO_0000052 inheres in
BFO_0000053 bearer of
BFO_0000054 realized by
BFO_0000055 realizes
The idea is that BFO includes the minimal set of relations required to
define BFO classes.
A draft is available here:
http://code.google.com/p/bfo/source/browse/#svn/trunk/src/ontology
The primary version is in OWL, with conversions to obo-format also
being made available. It will be possible to import the relations
without importing all the classes.
All other relations composed from BFO minimal relations will be in a
new ontology, with the namespace RO (the OBO_REL idspace/namespace
will be mildly deprecated). URIs will be of the form http://purl.obolibrary.org/obo/RO_nnnnnnn
(and in obo-format, RO:nnnnnnn). RO will include both generic
composed relations, as well as domain-specific relations. We can make
the domain specific sets available as subsets/slims. For example, RO-
neuron will contain a set of relations that have previously been under
discussion by the PONS task force.
Many of these relations can be treated as "macros", and expanded into
more complex expressions. For example:
RO_0002100 has soma part of
=> has_part some (GO:cell_body and part_of some ?Y)
RO_0002104
=> has plasma membrane part : has_part some (GO:plasma_membrane and
has_part some ?Y)
An early draft is available here:
http://code.google.com/p/obo-relations/source/browse/#svn/trunk/src/
ontology
We expect to be adding more soon. We've yet to specify the exact
policy and guidelines, but in general the idea is that it should be
fairly easy to get your relation added provided you can specify the
meaning of the relation using existing relations and OWL constructs.
The primary version is also in OWL, with conversions to obo-format
also being made available.
For both bfo-relations and RO, the expectation is that domain
ontologies will MIREOT in relations as required (although owl:imports
is also an option). The URIs would remain stable, and follow normal
OBO identifier lifecycle policy. If a definition changes, the the URI/
ID is obsoleted, and a new one is minted.
Some additional administrivia
There is now a mirror of this mail list here:
http://groups.google.com/group/obo-relations
(You can subscribe to either)
The RO existing tracker is deprecated and will be replaced by the one
here
http://code.google.com/p/obo-relations/
Comments and feedback welcome.
------------------------------------------------------------------------------
_______________________________________________
Obo-relations mailing list
Obo-re...@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/obo-relations
Larry Hunter
On May 26, 2010, at 1:35 PM, Chris Mungall wrote:
> We propose that the next version of BFO will contain a minimal set of
> instance level binary relations -- this set will overlap with the set
> of instance level relations used in the original RO.
Chris,
This is wonderful news. We will review this proposal and submit comments.
I was wondering if the Özgövde andl Grüninger paper from FOIS this year ("Foundational Process Relations in Bio-Ontologies") influenced the proposal? Their FOL characterization struck me as potentially very useful.
Larry
Stefan Schulz
I fully support your proposal and would like to propose to include a
general localization relation ("has-locus" or "has-location") as well.
Such a relation would be important to represent spatial inclusion of
independent continuants or occurrents within independent continuants,
e.g. a stone in a gallbladder, or a mitosis process in a cell.
Best regards
Stefan
2010/5/26 Larry Hunter <Larry....@ucdenver.edu>:
--
Stefan SCHULZ (apl. Prof. Dr. med.)
Institute of Medical Biometry and Medical Informatics
University Medical Center Freiburg
Stefan-Meier-Strasse 26 79104 Freiburg (Germany)
[home: Eschholzstr. 70, D-79115 Freiburg]
+49 (0)761 2036725, 2049089
http://purl.org/steschu
[stsc...@uni-freiburg.de], Skype: stschulz
Chris Mungall
Hi Stefan
Yes, I forgot to include located_in on the list - this was in the
original RO and would be in BFO too. There's been a bit of discussion
about this between Barry, Alan and others.
We've been unofficially using occurs_in for the process-continuant
relation, we don't have a satisfactory name for the inverse yet.
Chris Mungall
On May 26, 2010, at 12:48 PM, Larry Hunter wrote:
>
> On May 26, 2010, at 1:35 PM, Chris Mungall wrote:
>
>> We propose that the next version of BFO will contain a minimal set of
>> instance level binary relations -- this set will overlap with the set
>> of instance level relations used in the original RO.
>
> Chris,
>
> This is wonderful news. We will review this proposal and submit
> comments.
>
> I was wondering if the Özgövde andl Grüninger paper from FOIS this
> year ("Foundational Process Relations in Bio-Ontologies") influenced
> the proposal? Their FOL characterization struck me as potentially
> very useful.
I haven't read this paper, but that's not to say that it didn't or
couldn't influence what we're doing - the belated progress we're
making is the result of numerous discussions with multiple people.
Robert Hoehndorf
Hi,
This is great news. Some remarks:
> Many of these relations can be treated as "macros", and expanded into
> more complex expressions. For example:
> RO_0002100 has soma part of
> => has_part some (GO:cell_body and part_of some ?Y)
> RO_0002104
> => has plasma membrane part : has_part some (GO:plasma_membrane and
> has_part some ?Y)
> An early draft is available here:
> http://code.google.com/p/obo-relations/source/browse/#svn/trunk/src/
> ontology
We have been working on something similar here:
http://bioonto.de/obo2owl, where we provide OWL pattern definitions
for all the RO relations and a few more.
Many are fairly straightforward, but some are not so easy, like
integral-part-of.
The patterns you use seem to contain only one variable, ?Y, while a
relation between two classes will generally contain both classes as
variables (?X and ?Y). So the implementation at your link does not
seem to be an implementation of the RO, but rather macros for creating
complex OWL class descriptions (instead of class axioms).
How would you, for example, express integral-part-of?
Rob.
Bill Hogan
part_of is also antisymmetric, although you cannot say that in OWL.
Since they're instance-instance relations, you can say that part_of and has_part are inverses.
Finally, what about instance_of and lacks (instance-type relations)? Will they remain in RO?
Bill
As many of you are aware, RO has been somewhat static for a while. There have been a number of proposed extensions floating around, but nothing official. Part of the problem has been the impedance mismatch between the notion of type-level relations and the corresponding treatment in OWL. In addition, many of the original instance level relations are applicable outside biology and are more suited to an upper level ontology such as BFO.
We propose that the next version of BFO will contain a minimal set of instance level binary relations -- this set will overlap with the set of instance level relations used in the original RO. These relations will have URIs of the form http://purl.obolibrary.org/obo/BFO_nnnnnnn (in the obo-format version, these will be rendered as BFO:nnnnnnn).
BFO will likely include the follow relations, with the following IDs/URIs:
BFO_0000050 part of
BFO_0000051 has part
BFO_0000056 participates in
BFO_0000057 has participant
BFO_0000062 preceded by
BFO_0000063 precedes
BFO_0000060 immediately preceded by
BFO_0000061 immediately precedes
(and will most likely include additional ones too)
The corresponding older URIs with the prefix http://obofoundry.org/ro.owl# would be mildly deprecated (i.e. they would not be rendered obsolete, but ontologies would be encouraged to switch over). The new BFO relations are unambiguously instance-level binary relations, and the definitions and comments would reflect this. This also holds for the obo-format version of these relations (previously the plan was to have distinct type-level versions of these relations specified in obo-format, but this has been abandoned in order to make the translation between obo and owl simpler).
For example, the metadata for BFO_0000050 would be something along the lines of:
definition: "part of is a primitive reflexive, transitive relation, holding either between two processes or two continuants."
example: "every cell nucleus is part of some cell"
BFO will also likely include certain relations that were in various proposed extensions to the RO, but not in the original version, including, but no limited to:
BFO_0000052 inheres in
BFO_0000053 bearer of
BFO_0000054 realized by
BFO_0000055 realizes
The idea is that BFO includes the minimal set of relations required to define BFO classes.http://code.google.com/p/bfo/source/browse/#svn/trunk/src/ontology
A draft is available here:
The primary version is in OWL, with conversions to obo-format also being made available. It will be possible to import the relations without importing all the classes.
All other relations composed from BFO minimal relations will be in a new ontology, with the namespace RO (the OBO_REL idspace/namespace will be mildly deprecated). URIs will be of the form http://purl.obolibrary.org/obo/RO_nnnnnnn (and in obo-format, RO:nnnnnnn). RO will include both generic composed relations, as well as domain-specific relations. We can make the domain specific sets available as subsets/slims. For example, RO-neuron will contain a set of relations that have previously been under discussion by the PONS task force.
Many of these relations can be treated as "macros", and expanded into more complex expressions. For example:
RO_0002100 has soma part of
=> has_part some (GO:cell_body and part_of some ?Y)
RO_0002104
=> has plasma membrane part : has_part some (GO:plasma_membrane and has_part some ?Y)
An early draft is available here:
http://code.google.com/p/obo-relations/source/browse/#svn/trunk/src/ontology
We expect to be adding more soon. We've yet to specify the exact policy and guidelines, but in general the idea is that it should be fairly easy to get your relation added provided you can specify the meaning of the relation using existing relations and OWL constructs.
The primary version is also in OWL, with conversions to obo-format also being made available.
For both bfo-relations and RO, the expectation is that domain ontologies will MIREOT in relations as required (although owl:imports is also an option). The URIs would remain stable, and follow normal OBO identifier lifecycle policy. If a definition changes, the the URI/ID is obsoleted, and a new one is minted.
Some additional administrivia
There is now a mirror of this mail list here:
http://groups.google.com/group/obo-relations
(You can subscribe to either)
The RO existing tracker is deprecated and will be replaced by the one
here
http://code.google.com/p/obo-relations/
Comments and feedback welcome.
--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To post to this group, send email to bfo-d...@googlegroups.com.
To unsubscribe from this group, send email to bfo-discuss...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/bfo-discuss?hl=en.
Chris Mungall
[removed bfo-discuss, since this pertains specifically to the
relations in RO]
Essentially there are two types of macros. The first type expands an
annotation property used in a triple. The second type expands an
expression.
An example of the first (expanding an axiom):
integral_part_of(A,B) => A subclass of part_of some B, B subclass of
has_part some B
An example of the second (expanding an expression):
soma_located_in_some Y => has_part some (soma and part_of some Y)
The first requires two variables in the template, the second only one.
There's a bit more to it than that, more details to follow
> Rob.
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "BFO Discuss" group.
> To post to this group, send email to bfo-d...@googlegroups.com.
> To unsubscribe from this group, send email to bfo-discuss...@googlegroups.com
> .
> For more options, visit this group at http://groups.google.com/group/bfo-discuss?hl=en
> .
>
>
Chris Mungall
On May 26, 2010, at 3:45 PM, Bill Hogan wrote:
> I can't open the bfo2-relations.owl file, in either Protege 4.1
> alpha or Protege 3.4.4.
Hmm, seems fine in my P4.1
> part_of is also antisymmetric, although you cannot say that in OWL.
we should have a standard annotation property to indicate this, such
that it will be available in FOL translations
> Since they're instance-instance relations, you can say that part_of
> and has_part are inverses.
already declared
> Finally, what about instance_of and lacks (instance-type
> relations)? Will they remain in RO?
The instance_of relation corresponds to a class assertion in OWL.
There's a case for retaining it in a FOL version of the ontology though.
RO will most likely contain a lacks_part relation, defined in terms of
zero cardinality. This could be considered either an instance-class
relation, or a syntactic macro. Note that this doesn't do justice to
Werner's original treatment in x lacks_part P iff x has part exactly 0
P and a typical x has part some P. I suggest that if this is required
we treat this as two relations, one which is normality-neutral, and a
stronger one that has implications of abnormality. I suggest retaining
the label lacks_part for the simple zero-parts case and something like
'abnormally_lacks_part' for the stronger case.
At one time there was a proposal for a 3-ary lacks relation in which
the first argument was a relation. For now it seems simplest just to
explicitly enumerate the different binary relations.
For the cell ontology we will be using a relation
lacks_plasma_membrane_part X => has_part some (GO:plasma_membrane and
has_part 0 X)
As defined in http://www.biomedcentral.com/1471-2105/10/70
Chris Mungall
On May 26, 2010, at 3:45 PM, Bill Hogan wrote:
> I can't open the bfo2-relations.owl file, in either Protege 4.1
> alpha or Protege 3.4.4.
Try downloading the file first - or this url
http://purl.obolibrary.org/obo/ro.owl
Bill Hogan
When I try to open the file bfo-relation.owl, I get the following error: Server returned HTTP response code: 503 for URL: http://www.w3.org/TR/html4/loose.dtd
with the following stack trace:
org.semanticweb.owl.io.OWLOntologyCreationIOException: Server returned HTTP response code: 503 for URL: http://www.w3.org/TR/html4/loose.dtd
at uk.ac.manchester.cs.owl.ParsableOWLOntologyFactory.loadOWLOntology(ParsableOWLOntologyFactory.java:174)
at uk.ac.manchester.cs.owl.OWLOntologyManagerImpl.loadOntology(OWLOntologyManagerImpl.java:461)
at uk.ac.manchester.cs.owl.OWLOntologyManagerImpl.loadOntology(OWLOntologyManagerImpl.java:424)
at org.protege.editor.owl.model.OWLModelManagerImpl.loadOntology(OWLModelManagerImpl.java:328)
at org.protege.editor.owl.model.OWLModelManagerImpl.loadOntologyFromPhysicalURI(OWLModelManagerImpl.java:390)
at org.protege.editor.owl.OWLEditorKit.handleLoadFrom(OWLEditorKit.java:147)
at org.protege.editor.owl.OWLEditorKit.handleLoadRequest(OWLEditorKit.java:141)
at org.protege.editor.core.ProtegeManager.openAndSetupEditorKit(ProtegeManager.java:146)
at org.protege.editor.core.ProtegeWelcomeFrame$ProtegeWelcomePanel$2.actionPerformed(ProtegeWelcomeFrame.java:113)
at org.protege.editor.core.ui.util.LinkLabel.activateLink(LinkLabel.java:97)
at org.protege.editor.core.ui.util.LinkLabel.access$100(LinkLabel.java:30)
at org.protege.editor.core.ui.util.LinkLabel$1.mouseReleased(LinkLabel.java:63)
at java.awt.Component.processMouseEvent(Unknown Source)
at javax.swing.JComponent.processMouseEvent(Unknown Source)
at java.awt.Component.processEvent(Unknown Source)
at java.awt.Container.processEvent(Unknown Source)
at java.awt.Component.dispatchEventImpl(Unknown Source)
at java.awt.Container.dispatchEventImpl(Unknown Source)
at java.awt.Component.dispatchEvent(Unknown Source)
at java.awt.LightweightDispatcher.retargetMouseEvent(Unknown Source)
at java.awt.LightweightDispatcher.processMouseEvent(Unknown Source)
at java.awt.LightweightDispatcher.dispatchEvent(Unknown Source)
at java.awt.Container.dispatchEventImpl(Unknown Source)
at java.awt.Window.dispatchEventImpl(Unknown Source)
at java.awt.Component.dispatchEvent(Unknown Source)
at java.awt.EventQueue.dispatchEvent(Unknown Source)
at java.awt.EventDispatchThread.pumpOneEventForHierarchy(Unknown Source)
at java.awt.EventDispatchThread.pumpEventsForHierarchy(Unknown Source)
at java.awt.EventDispatchThread.pumpEvents(Unknown Source)
at java.awt.EventDispatchThread.pumpEvents(Unknown Source)
at java.awt.EventDispatchThread.run(Unknown Source)
Caused by: java.io.IOException: Server returned HTTP response code: 503 for URL: http://www.w3.org/TR/html4/loose.dtd
at sun.net.www.protocol.http.HttpURLConnection.getInputStream(Unknown Source)
at com.sun.org.apache.xerces.internal.impl.XMLEntityManager.setupCurrentEntity(Unknown Source)
at com.sun.org.apache.xerces.internal.impl.XMLEntityManager.startEntity(Unknown Source)
at com.sun.org.apache.xerces.internal.impl.XMLEntityManager.startDTDEntity(Unknown Source)
at com.sun.org.apache.xerces.internal.impl.XMLDTDScannerImpl.setInputSource(Unknown Source)
at com.sun.org.apache.xerces.internal.impl.XMLDocumentScannerImpl$DTDDispatcher.dispatch(Unknown Source)
at com.sun.org.apache.xerces.internal.impl.XMLDocumentFragmentScannerImpl.scanDocument(Unknown Source)
at com.sun.org.apache.xerces.internal.parsers.XML11Configuration.parse(Unknown Source)
at com.sun.org.apache.xerces.internal.parsers.XML11Configuration.parse(Unknown Source)
at com.sun.org.apache.xerces.internal.parsers.XMLParser.parse(Unknown Source)
at com.sun.org.apache.xerces.internal.parsers.AbstractSAXParser.parse(Unknown Source)
at javax.xml.parsers.SAXParser.parse(Unknown Source)
at edu.unika.aifb.rdf.api.syntax.RDFParser.parse(RDFParser.java:111)
at org.coode.owl.rdfxml.parser.RDFXMLParser.parse(RDFXMLParser.java:86)
at uk.ac.manchester.cs.owl.ParsableOWLOntologyFactory.loadOWLOntology(ParsableOWLOntologyFactory.java:159)
... 30 more
Terry Meehan
You need to go to the checkout tab on the code.google. page. There it
will tell you to use the following line command
svn checkout http://bfo.googlecode.com/svn/trunk/ bfo-read-only
this will create a folder in your directory containing owl files
readable by protege.
Terry m.
> <mailto:bfo-d...@googlegroups.com>.
> To unsubscribe from this group, send email to
> bfo-discuss...@googlegroups.com
> <mailto:bfo-discuss%2Bunsu...@googlegroups.com>.
> For more options, visit this group at
> http://groups.google.com/group/bfo-discuss?hl=en.
>
>
>
> --
> You received this message because you are subscribed to the
> Google Groups "BFO Discuss" group.
> To post to this group, send email to
> bfo-d...@googlegroups.com
> <mailto:bfo-d...@googlegroups.com>.
> To unsubscribe from this group, send email to
> bfo-discuss...@googlegroups.com
> <mailto:bfo-discuss%2Bunsu...@googlegroups.com>.
> For more options, visit this group at
> http://groups.google.com/group/bfo-discuss?hl=en.
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "BFO Discuss" group.
> To post to this group, send email to bfo-d...@googlegroups.com
> <mailto:bfo-d...@googlegroups.com>.
> To unsubscribe from this group, send email to
> bfo-discuss...@googlegroups.com
> <mailto:bfo-discuss%2Bunsu...@googlegroups.com>.
> For more options, visit this group at
> http://groups.google.com/group/bfo-discuss?hl=en.
>
>
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
Alan Ruttenberg
If you want to view RO, please use
http://purl.obolibrary.org/obo/ro.owl
If you want to view BFO 2, such as it is at this early stage, use
http://purl.obolibrary.org/obo/bfo.owl
Note that this version is a minimalist BFO, without regions and
without granularity, but with relations.
http://purl.obolibrary.org/obo/bfo/all.owl
Includes regions and granularity.
New classes discussed haven't been added yet, nor have the definitions
of the relations been add, nor is the axiomatization complete - this
is early work.
-Alan
------------------------------------------------------------------------------
Alan Ruttenberg
> Dear Chris,
>
> I fully support your proposal and would like to propose to include a
> general localization relation ("has-locus" or "has-location") as well.
> Such a relation would be important to represent spatial inclusion of
> independent continuants or occurrents within independent continuants,
> e.g. a stone in a gallbladder, or a mitosis process in a cell.
We need a good definition for this.
The reason I haven't yet added located in and contained in is that
they confuse me.
(Attached is a picture I drew for discussion some time ago.)
The issue is that first, located in was defined for two cases
1) Continuant->Region
2) Continuant->Continuant
Contained in was similarly defined, but said to relate immaterial
continuants to continuants.
The second relation was defined in terms of the first. The problem is
that the first isn't defined. In order to probe this question, I drew
the attached diagram of a bottle (wall), bottle (interior),
(bottle=bottle wall+ bottle interior), some water in the bottle, and a
coin in the water, and asked around what the location and containment
relations were among things.
The ensuing discussion did not yield a satisfactory conclusion.
Moreover there has been a recent challenge, spearheaded by Bjoern, to
say that we should not define location by resort to regions (hence
regions being currently optional in my draft bfo2).
Barry points to the paper "Containment Relations in Anatomical
Ontologies" by Maureen Donnelly, which I like, which presents 5
candidate relations of the sort we might use. Note, however, that they
still use a region function in their definitions, and afaik, this
region function is not documented, and while it might seem intuitive
for simple solids, it is not, IMO, in a situation in which we can have
immaterial continuants (aka sites, holes, cavities).
So, do you have a favorite definition for has-locus that works for
material and immaterial continuants, which relies on intuitive (or at
least workable) primitives, and does not resort to regions as its
basis? Or an argument as to which of these requirements should give?
The attached bottle2.graffle.pdf shows the last state of the
discussion and the proposed relations among parts. A point of mystery
for me is the claim that the coin not located nor contained in the
water.
Regards,
Alan
Dr. Mathias Brochhausen
this is indeed very good news and important work to be carried out. Just
some small rmearks of issues of which I presume you might already be aware:
It should be ensured that all implicite relations in BFO are made
explicite (inheres_in of some kind, etc.).
On the other hand it needs to be clear that no domain specific relation
(or even relation named in a highly domain-contaminated way) is included
in BFO since this might compromise BFO as a general Upper Ontology. One
of the more theoretical issues with using RO as BFO-relations in the
past (which was done inofficial sometimes) was that RO was be definition
a domain specific artefact whereas BFO is an Upper Ontology in the sense
of the word as used by IEEE.
Therefore, I fully support getting these things straightend out. If
there is some worh that need to be done, besides discussions on the
list, I would be happy to contribute, especially since I will be on a
research leave in Buffalo between Septmeber 2010 and February 2011,
doing some theoretical work on relations and occurents.
Best,
Mathias
Colin Batchelor
> RO will most likely contain a lacks_part relation, defined in terms of
> zero cardinality. This could be considered either an instance-class
> relation, or a syntactic macro. Note that this doesn't do justice to
> Werner's original treatment in x lacks_part P iff x has part exactly 0
> P and a typical x has part some P. I suggest that if this is required
> we treat this as two relations, one which is normality-neutral, and a
> stronger one that has implications of abnormality. I suggest retaining
> the label lacks_part for the simple zero-parts case and something like
> 'abnormally_lacks_part' for the stronger case.
What is the zero-parts case for?
A Manx cat lacks a tail, a battleship, and the first half of *Like a Rolling Stone*. But only one of these is present in the typical cat.
I've seen PRO using the lacks relation for the peptide chain that comes fresh off the ribosome---it lacks phosphoryl groups, it lacks a ubiquitinyl group on lysine number 57, it lacks sharks---but isn't it better to say that the peptide chain has_part only canonical amino acid residues?
Best wishes,
Colin.
DISCLAIMER:
This communication (including any attachments) is intended for the use of the addressee only and may contain confidential, privileged or copyright material. It may not be relied upon or disclosed to any other person without the consent of the RSC. If you have received it in error, please contact us immediately. Any advice given by the RSC has been carefully formulated but is necessarily based on the information available, and the RSC cannot be held responsible for accuracy or completeness. In this respect, the RSC owes no duty of care and shall not be liable for any resulting damage or loss. The RSC acknowledges that a disclaimer cannot restrict liability at law for personal injury or death arising through a finding of negligence. The RSC does not warrant that its emails or attachments are Virus-free: Please rely on your own screening.
Ward Blondé
Chris Mungall wrote:
> BFO will likely include the follow relations, with the following IDs/
> URIs:
>
> BFO_0000050 part of
> BFO_0000051 has part
> BFO_0000056 participates in
> BFO_0000057 has participant
> BFO_0000062 preceded by
> BFO_0000063 precedes
> BFO_0000060 immediately preceded by
> BFO_0000061 immediately precedes
>
> (and will most likely include additional ones too
>
> BFO will also likely include certain relations that were in various
> proposed extensions to the RO, but not in the original version,
> including, but no limited to:
>
> BFO_0000052 inheres in
> BFO_0000053 bearer of
> BFO_0000054 realized by
> BFO_0000055 realizes
>
Hello,
I am glad to see a follow-up on relations after all this time! But I
still wonder why some of these labels contain a verb and others not? Is
there an unofficial policy that if the label doesn't contain a verb in
the present tense, that it is preceded by 'is'? Why should 'part of'
be read as 'is part of' instead of for instance 'has part of'?
I think such labels may make it difficult for natural language
processing systems that want to exploit these labels, because it is only
clear for humans what to make of them.
Regards,
Ward
Robert Hoehndorf
Hi,
> What is the zero-parts case for?
> A Manx cat lacks a tail, a battleship, and the first half of *Like a Rolling Stone*. But only one of these is present in the typical cat.
> I've seen PRO using the lacks relation for the peptide chain that comes fresh off the ribosome---it lacks phosphoryl groups, it lacks a ubiquitinyl group on lysine number 57, it lacks sharks---but isn't it better to say that the peptide chain has_part only canonical amino acid residues?
The last statement would be incorrect, assuming that has_part is
transitive and amino acid residues have parts which are not themselves
amino acid residues (atoms and electrons, for example).
In OWL, it is sometimes useful to infer explicitly things which are
not there, mostly when they are expected to be there. You could not do
this otherwise because OWL uses open world reasoning, so you have to
explicitly assert things which are not there, using negation.
Btw, what is the benefit of writing "has-part exactly 0" instead of
"not has-part"? Both are equivalent, and the latter seems more
intuitive to me.
Rob.
Colin Batchelor
> The attached bottle2.graffle.pdf shows the last state of the
> discussion and the proposed relations among parts. A point of mystery
> for me is the claim that the coin not located nor contained in the
> water.
Since this is for scientific ontologies, we should be asking what the causal consequences of the coin being underwater are. I think what's important here is that the water has the disposition to resist the movement of the coin. Equally the coin has the disposition to leach out metal atoms into the water. But these could be handled by something like a meets or an externally_connected_to relation.
If I throw the bottle into the air at night, then from where I stand that bottle might be briefly in Cassiopeia or Ursa Major. But those positions have no causal consequences. And there is no sense in which the bottle meets Alcor or Mizar.
If we were manufacturing coins-in-bottles then I think the questions would be different.
Best wishes,
Colin.
DISCLAIMER:
This communication (including any attachments) is intended for the use of the addressee only and may contain confidential, privileged or copyright material. It may not be relied upon or disclosed to any other person without the consent of the RSC. If you have received it in error, please contact us immediately. Any advice given by the RSC has been carefully formulated but is necessarily based on the information available, and the RSC cannot be held responsible for accuracy or completeness. In this respect, the RSC owes no duty of care and shall not be liable for any resulting damage or loss. The RSC acknowledges that a disclaimer cannot restrict liability at law for personal injury or death arising through a finding of negligence. The RSC does not warrant that its emails or attachments are Virus-free: Please rely on your own screening.
------------------------------------------------------------------------------
David Osumi-Sutherland
On 27 May 2010, at 10:45, Robert Hoehndorf wrote:
> On Thu, 2010-05-27 at 10:19 +0100, Colin Batchelor wrote:
>
> Hi,
>
>> What is the zero-parts case for?
>> A Manx cat lacks a tail, a battleship, and the first half of *Like
>> a Rolling Stone*. But only one of these is present in the typical
>> cat.
>> I've seen PRO using the lacks relation for the peptide chain that
>> comes fresh off the ribosome---it lacks phosphoryl groups, it lacks
>> a ubiquitinyl group on lysine number 57, it lacks sharks---but
>> isn't it better to say that the peptide chain has_part only
>> canonical amino acid residues?
> The last statement would be incorrect, assuming that has_part is
> transitive and amino acid residues have parts which are not themselves
> amino acid residues (atoms and electrons, for example).
>
> In OWL, it is sometimes useful to infer explicitly things which are
> not there, mostly when they are expected to be there. You could not do
> this otherwise because OWL uses open world reasoning, so you have to
> explicitly assert things which are not there, using negation.
In my experience, we occasionally need to use negation if we are to
capture the classifications that scientists find useful. Terry has
especially good examples of this for classification of cell types in
the immune system.
>
> Btw, what is the benefit of writing "has-part exactly 0" instead of
> "not has-part"? Both are equivalent, and the latter seems more
> intuitive to me.
As I understand it, we can't use cardinality restrictions with
transitive relations in OWL. As has_part is clearly transitive, to
express cardinality we would need a non-transitive child relation for
has_part - perhaps 'has_component'? In this case I think I'd prefer
to use explicit negation - although I worry about the hit on reasoner
efficiency that might result.
- David
>
> Rob.
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Obo-relations mailing list
> Obo-re...@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/obo-relations
David Osumi-Sutherland, PhD
Ontologist / Curator
Virtual Fly Brain / FlyBase
Department of Genetics
University of Cambridge
Downing Street
Cambridge, CB2 3EH
UK
+44 (0)1223 333 963
Alan Ruttenberg
> Alan Ruttenberg writes:
>
>> The attached bottle2.graffle.pdf shows the last state of the
>> discussion and the proposed relations among parts. A point of mystery
>> for me is the claim that the coin not located nor contained in the
>> water.
>
> Since this is for scientific ontologies, we should be asking what the causal consequences of the coin being underwater are. I think what's important here is that the water has the disposition to resist the movement of the coin. Equally the coin has the disposition to leach out metal atoms into the water. But these could be handled by something like a meets or an externally_connected_to relation.
>
> If I throw the bottle into the air at night, then from where I stand that bottle might be briefly in Cassiopeia or Ursa Major. But those positions have no causal consequences. And there is no sense in which the bottle meets Alcor or Mizar.
>
> If we were manufacturing coins-in-bottles then I think the questions would be different.
The coin in bottle is not unlike the stone in kidney, or the protein
in nucleus. There are plenty of biological cases where this
configuration is relevant.
Yours firmly on earth,
Alan
Alan Ruttenberg
I agree with you Ward. Now that we have ids for relations we can have
a reasonable discussion of what the labels should be, without having
there be the same kind of consequences as there were before. I'm
inclined to change part of to is part of, and then add, as alternative
terms, "part of" and "part_of", to satisfy different constituencies.
If anyone has objections to this course of action, please shout.
-Alan
Michel Dumontier
--For more options, visit this group at http://groups.google.com/group/bfo-discuss?hl=en.
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To post to this group, send email to bfo-d...@googlegroups.com.
To unsubscribe from this group, send email to bfo-discuss...@googlegroups.com.
--
Michel Dumontier
Associate Professor of Bioinformatics
Carleton University
http://dumontierlab.com
Michel Dumontier
As I understand it, we can't use cardinality restrictions with
transitive relations in OWL. As has_part is clearly transitive, to
express cardinality we would need a non-transitive child relation for
has_part - perhaps 'has_component'? In this case I think I'd prefer
to use explicit negation - although I worry about the hit on reasoner
efficiency that might result.
- David
>
> Rob.
>David Osumi-Sutherland, PhD
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Obo-relations mailing list
> Obo-re...@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/obo-relations
Ontologist / Curator
Virtual Fly Brain / FlyBase
Department of Genetics
University of Cambridge
Downing Street
Cambridge, CB2 3EH
UK
+44 (0)1223 333 963
------------------------------------------------------------------------------
_______________________________________________
Obo-relations mailing list
Obo-re...@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/obo-relations
David Osumi-Sutherland
and incidently, for composing queries using manchester syntax.
Michel Dumontier
and incidently, for composing queries using manchester syntax.
I don't find writing an underscore any more onerous than quoting a name with spaces in it.
Robert Hoehndorf
Hi David,
> > Btw, what is the benefit of writing "has-part exactly 0" instead of
> > "not has-part"? Both are equivalent, and the latter seems more
> > intuitive to me.
> As I understand it, we can't use cardinality restrictions with
> transitive relations in OWL. As has_part is clearly transitive, to
> express cardinality we would need a non-transitive child relation for
> has_part - perhaps 'has_component'? In this case I think I'd prefer
> to use explicit negation - although I worry about the hit on reasoner
> efficiency that might result.
I agree, negation would certainly be clearer. "has_part exactly 0" and
"not has_part some" are exactly equivalent, and at least Hermit seems
to convert "has_part exactly 0" to the negated form, because it does
not complain about cardinality and transitivity for "has_part exactly
0" (as it would for "has_part exactly 1").
So I guess it is just two different ways of writing the same thing
("has_part only not" would be a third way), I just wondered why this
one was prefered over the negated one. But it won't make any
difference formally.
Ward Blondé
>>
>> I'm absolutely in favor of having 'is part of' as a preferred
>> name. It mostly certainly makes it easier for language generation,
>
> It does? What about pluralization?
>
> what about it?
>
> the relation is between two entities.
>
> but still, for the purpose of form generation over a set of such
> relations, you may need to specify a more sophisticated syntax.
The labels are given to relation types at the instance level. A natural
language processing system could use the semantics that are given to the
class level variant to conjugate the sentence. For instance 'nucleus'
'is part of' 'cell', could be conjugated to 'Every nucleus is part of
some cell.', which is a sound sentence. If a verb is missing in the
labels, there is no basis to do this.
Indeed, quoting relation types is nicer with short labels. Probably
there should be made a separate quoting label? The consequence of sound
labelling might even lead to calling the subsumption relation just
'is'... instead of 'is a'. This 'a' could fit after any of the labels.
It doesn't fit in the scheme either.
regards,
Ward
>> <mailto:Obo-re...@lists.sourceforge.net>
>> >> https://lists.sourceforge.net/lists/listinfo/obo-relations
>> >>
>> >
>> >
>> ------------------------------------------------------------------------------
>> >
>> > _______________________________________________
>> > Obo-relations mailing list
>> > Obo-re...@lists.sourceforge.net
>> <mailto:Obo-re...@lists.sourceforge.net>
>> > https://lists.sourceforge.net/lists/listinfo/obo-relations
>> >
>>
>> --
>> You received this message because you are subscribed to the
>> Google Groups "BFO Discuss" group.
>> To post to this group, send email to
>> bfo-d...@googlegroups.com
>> <mailto:bfo-d...@googlegroups.com>.
>> To unsubscribe from this group, send email to
>> bfo-discuss...@googlegroups.com
>> <mailto:bfo-discuss%2Bunsu...@googlegroups.com>.
>> For more options, visit this group at
>> http://groups.google.com/group/bfo-discuss?hl=en.
>>
>>
>>
>>
>> --
>> Michel Dumontier
>> Associate Professor of Bioinformatics
>> Carleton University
>> http://dumontierlab.com
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> Obo-relations mailing list
>> Obo-re...@lists.sourceforge.net
>> <mailto:Obo-re...@lists.sourceforge.net>
>> https://lists.sourceforge.net/lists/listinfo/obo-relations
>
> David Osumi-Sutherland, PhD
> Curator/ Ontologist
> FlyBase / Virtual Fly Brain
> Department of Genetics,
> University of Cambridge,
> Downing Street,
> Cambridge, CB2 3EH, UK
> Tel: +44 (0)1223 333 963
> Fax: +44 (0)1223 766 732
>
>
>
>
> --
> Michel Dumontier
> Associate Professor of Bioinformatics
> Carleton University
> http://dumontierlab.com
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
Colin Batchelor
> Moreover there has been a recent challenge, spearheaded by Bjoern, to
> say that we should not define location by resort to regions (hence
> regions being currently optional in my draft bfo2).
Could we have a recap of the key arguments in favour of this, please?
Best wishes,
Colin.
DISCLAIMER:
This communication (including any attachments) is intended for the use of the addressee only and may contain confidential, privileged or copyright material. It may not be relied upon or disclosed to any other person without the consent of the RSC. If you have received it in error, please contact us immediately. Any advice given by the RSC has been carefully formulated but is necessarily based on the information available, and the RSC cannot be held responsible for accuracy or completeness. In this respect, the RSC owes no duty of care and shall not be liable for any resulting damage or loss. The RSC acknowledges that a disclaimer cannot restrict liability at law for personal injury or death arising through a finding of negligence. The RSC does not warrant that its emails or attachments are Virus-free: Please rely on your own screening.
------------------------------------------------------------------------------
Kerry Trentelman
the idea of projection functions which map independent continuants
(ICs) and processes to regions.
Then have located_in holding between ICs.
The relation contained_in holds between material continuants (MCs) and
sites.
If we ruled that located_in held between MCs (rather than ICs) then
all part_of, located_in and contained_in relations between the ICs of
Alan's bottle/water/coin picture are:
I part_of B
W contained_in I
C located_in W
W located_in B
A located_in B
And that's it. (Note parthood and location coincides for material
continuants W, C, A and B.) Location is transitive, so also have:
C located_in B
Wondering however if we need an axiom:
X located_in Y & Y contained_in Z => X contained_in Z
Then we'd also get:
C contained_in I
A contained_in I
Welcome more discussion.
Cheers,
Kerry
> Bottle2
> .graffle
> .pdf
> >
> ------------------------------------------------------------------------------
James Malone
ontologies) and part-time text miner I would like to encourage the
consideration of relation labels that either i) are amenable to
natural language transformation (as Michel had previously highlighted)
or ii) contain some form of synonym/alternative label which can be
used instead. I think text mining is an important use case for a lot
of ontology consumption so would like to wave the flag now while I
have the opportunity.
Cheers,
James
On Thu, May 27, 2010 at 2:13 PM, Colin Batchelor <Batch...@rsc.org> wrote:
> Alan Ruttenberg writes:
>
>> Moreover there has been a recent challenge, spearheaded by Bjoern, to
>> say that we should not define location by resort to regions (hence
>> regions being currently optional in my draft bfo2).
>
> Could we have a recap of the key arguments in favour of this, please?
>
> Best wishes,
> Colin.
>
> DISCLAIMER:
>
> This communication (including any attachments) is intended for the use of the addressee only and may contain confidential, privileged or copyright material. It may not be relied upon or disclosed to any other person without the consent of the RSC. If you have received it in error, please contact us immediately. Any advice given by the RSC has been carefully formulated but is necessarily based on the information available, and the RSC cannot be held responsible for accuracy or completeness. In this respect, the RSC owes no duty of care and shall not be liable for any resulting damage or loss. The RSC acknowledges that a disclaimer cannot restrict liability at law for personal injury or death arising through a finding of negligence. The RSC does not warrant that its emails or attachments are Virus-free: Please rely on your own screening.
>
> --
> You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
> To post to this group, send email to bfo-d...@googlegroups.com.
> To unsubscribe from this group, send email to bfo-discuss...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/bfo-discuss?hl=en.
>
>
------------------------------------------------------------------------------
Alan Rector
On 27 May 2010, at 00:19, Chris Mungall wrote:
>>
>> part_of is also antisymmetric, although you cannot say that in OWL.
>
> we should have a standard annotation property to indicate this, such
> that it will be available in FOL translations
There should also probably be a "lint" or "OPPL" script to check for loops
in real ontologies.
Alan
-----------------------
Alan Rector
Professor of Medical Informatics
School of Computer Science
University of Manchester
Manchester M13 9PL, UK
TEL +44 (0) 161 275 6149/6188
FAX +44 (0) 161 275 6204
www.cs.man.ac.uk/~rector
www.co-ode.org
http://clahrc-gm.nihr.ac.uk/
Alan Ruttenberg
> All
>
> On 27 May 2010, at 00:19, Chris Mungall wrote:
>
>>>
>>> part_of is also antisymmetric, although you cannot say that in OWL.
>>
>> we should have a standard annotation property to indicate this, such
>> that it will be available in FOL translations
>
>
> There should also probably be a "lint" or "OPPL" script to check for loops
> in real ontologies.
It would be great to be able to implement some of this stuff in OPPL
I'm unaware of how to run it outside protege (and had some issues with
it behaving flakey in that context) - is there actually some way to
run it as a script?
-(the other)Alan
>
> Alan
>
> -----------------------
> Alan Rector
> Professor of Medical Informatics
> School of Computer Science
> University of Manchester
> Manchester M13 9PL, UK
> TEL +44 (0) 161 275 6149/6188
> FAX +44 (0) 161 275 6204
> www.cs.man.ac.uk/~rector
> www.co-ode.org
> http://clahrc-gm.nihr.ac.uk/
>
>
>
>
>
> --
> You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
> To post to this group, send email to bfo-d...@googlegroups.com.
> To unsubscribe from this group, send email to bfo-discuss...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/bfo-discuss?hl=en.
>
>
------------------------------------------------------------------------------
Ward Blondé
On Thu, May 27, 2010 at 5:27 AM, Ward Blondé <ward....@ugent.be> wrote:
>
>
> Chris Mungall wrote:
>> BFO will likely include the follow relations, with the following IDs/
>> URIs:
>>
>> BFO_0000050 part of
>> BFO_0000051 has part
>> BFO_0000056 participates in
>> BFO_0000057 has participant
>> BFO_0000062 preceded by
>> BFO_0000063 precedes
>> BFO_0000060 immediately preceded by
>> BFO_0000061 immediately precedes
>>
>>
> Hello,
>
> I am glad to see a follow-up on relations after all this time! But I
> still wonder why some of these labels contain a verb and others not? Is
> there an unofficial policy that if the label doesn't contain a verb in
> the present tense, that it is preceded by 'is'? Why should 'part of'
> be read as 'is part of' instead of for instance 'has part of'?
> I think such labels may make it difficult for natural language
> processing systems that want to exploit these labels, because it is only
> clear for humans what to make of them.
I agree with you Ward. Now that we have ids for relations we can have
a reasonable discussion of what the labels should be, without having
there be the same kind of consequences as there were before. I'm
inclined to change part of to is part of, and then add, as alternative
terms, "part of" and "part_of", to satisfy different constituencies.
If anyone has objections to this course of action, please shout.
-Alan
Did anybody ever consider to have human readable identifiers for BFO? Like for instance BFO_precedes. I think it could help a lot to promote BFO for practical usage. Querying on URIs with numbers is very cumbersome, because the programmer cannot see what the query means. A software-project that uses BFO as an upper-level ontology may need many queries that include BFO terms, and this applies even more to the relation types.
The fixation of a human readable URI may seem like a big decision, because terminology may change over time. If we look at RDF, there are some ugly URIs that stem from the past like rdf:type (meaning in fact rdf:isInstanceOf), but this is still nicer for usage than only numbers.
I think this reasoning applies to the whole OBO Foundry, but for an upper level ontology and for relation types, the matter is most important. Of course I realize that identifiers should not be seen by end users and I look at this as a programmer and a query-person. But it may be an issue for the uptake of BFO. OWL and RDF do not use numbers to identify their constructs at least. Numbers seem to appeal only to classical databases.
Regards,
Ward
>