Data Class

[James Malone]

Hi All,

I added in the newly defined data class which is either derived data (output of a DT) or measured data (output of an assay) and also added these classes.  Please see updated file and let me know if this does not look right to you.  I would suggest over time we should obsolete data item and data set - although these of course come from IAO so I'm not sure what the procedure is there.

Alan, I've been doing a lot of work last couple of days and reasoning has  slowed slightly.  I haven't adding anything that is incorrect but feel free to have a prod with your new tool if you like :)  If you find something you want to take out can you let me know first please as I don't want to lose any of the new information I added.

Thanks!

James

[Chris Stoeckert]

Hi James,
Thanks for doing this. I'm looking at http://obi.svn.sourceforge.net/svnroot/obi/trunk/src/ontology/branches/DigitalEntityPlus.owl

data, measured data, derived data
I thought we were going to name these to be uncontroversially
singular: data item, measured data item, derived data item. So rather
than obsolete "data item", can we modify it? Also for data set - can
we just relabel as "data collection"?

def: "Measured data is an information content entity that is the
output of an assay."
Should also add. "Measured data can also be the output of some process
on information. Examples include the measured blood glucose level, the
count of matches to a information in a database."

measured data and derived data are both direct subclasses of
information content entity. Shouldn't these be under data (item) ?

Cheers,
Chris

[James Malone]

Hi Chris,

Thanks for looking at these.

data, measured data, derived data
I thought we were going to name these to be uncontroversially singular: data item, measured data item, derived data item.

Sounds good, I've changed the names now and checked them in.
 

So rather than obsolete "data item", can we modify it? Also for data set - can we just relabel as "data collection"?

'data item' is an IAO class which was why I added 'data' to show how I thought this should look (reflecting changes form the workshop).  I can change the IAO file, but not sure what the protocol is, Alan?  Happy to change it to data collection.
 

def: "Measured data is an information content entity that is the output of an assay."
Should also add. "Measured data can also be the output of some process on information. Examples include the measured blood glucose level, the count of matches to a information in a database."

Perfect, added that in now.
 

measured data and derived data are both direct subclasses of information content entity. Shouldn't these be under data (item) ?

They are asserted children if ICE but they do infer under data (the one I added not the IAO one) when the reasoner is ran.  Does that make sense?  For clarity to other people I added 'data' as a class just to show how it should look when changes are done (i.e. 'derived data' and 'measured data' are what we consider to be 'data').  We also have from IAO though 'data set'. I guess we need to harmonise these and remove the one I added and put the same restrictions on the IAO class. 

James
 

[Chris Stoeckert]

Hi James,

Sounds good.

I need to add more of the DENRIE core terms to the file. I've checked out from svn http://obi.svn.sourceforge.net/viewvc/obi/trunk/src/ontology/branches/

but don't recall how to merge these and make changes before recommitting. Are there instructions somewhere?

Thanks,

Chris

[James Malone]

I'm editing just using Protege but there is a script Jay produced to do this.  You just need to put the terms into a text file in the appropriate order.  I think Philippe send a link to it this morning on the list to Oliver (though for some reason I can not find it).  If you get stuck though give me a shout and I can add them.

James 

[Melanie Courtot]

Hi all,

On 12-Feb-09, at 8:33 AM, James Malone wrote:

Hi Chris,

Thanks for looking at these.

data, measured data, derived data
I thought we were going to name these to be uncontroversially singular: data item, measured data item, derived data item.

Sounds good, I've changed the names now and checked them in.
 

So rather than obsolete "data item", can we modify it? Also for data set - can we just relabel as "data collection"?

'data item' is an IAO class which was why I added 'data' to show how I thought this should look (reflecting changes form the workshop).  I can change the IAO file, but not sure what the protocol is, Alan?  Happy to change it to data collection.

I think that what we are calling data item in DENRIE is more restrictive than what is called data item in IAO: in DENRIE, we specifically restrict to output of processes (derived data from DT or measured data from assay).

In this case, I think the IAO class is to broad for what we are trying to represent and we should subclass it.

While there is nothing preventing us to keep the same label, data item, in both cases, I think that first this is confusing and second that the DENRIE term is more specific than that. On the other hand I don't have a good label to propose :) I am thinking something like "acquired data", "investigation output data": something that would reflect the fact that in this case we are talking about precisely measured or calculated data that are output of OBI processes.

I attach a screenshot of what it may look like:

I declared measured data item and derived data item as subclasses of the IAO data item (on the left in the asserted hierarchy)

The DENRIE data stays as positioned by James, under ICE.

When running the reasonner, the correct hierarchy is inferred (see middle)

 

def: "Measured data is an information content entity that is the output of an assay."
Should also add. "Measured data can also be the output of some process on information. Examples include the measured blood glucose level, the count of matches to a information in a database."

Perfect, added that in now.
 

measured data and derived data are both direct subclasses of information content entity. Shouldn't these be under data (item) ?

They are asserted children if ICE but they do infer under data (the one I added not the IAO one) when the reasoner is ran.

So the change compared to that is that with the above the DENRIE data gets inferred under the IAO data item.

  Does that make sense?  For clarity to other people I added 'data' as a class just to show how it should look when changes are done (i.e. 'derived data' and 'measured data' are what we consider to be 'data').  We also have from IAO though 'data set'. I guess we need to harmonise these and remove the one I added and put the same restrictions on the IAO class. 

I'm not sure I understand this last bit and what needs to be modified. I think we agreed that data set should be asserted under data item (in IAO). If this is the case, then we should be ok with our definition stating measured data item + derived data item, as those would encompass the data sets too.

In any case, harmonizing with IAO makes sense :)

Thanks for the edits James!

Regarding the things that need to be added, I don't think there was a huge number of terms, so maybe the perl script is an overkill. If we can work out on a clear list of classes and their metadata + their position in the hierarchy I'm happy to do the edits (and I volunteer James too ;) )

Cheers,

Melanie

James
 

Cheers,
Chris

On Feb 11, 2009, at 1:01 PM, James Malone wrote:

Hi All,

I added in the newly defined data class which is either derived data (output of a DT) or measured data (output of an assay) and also added these classes.  Please see updated file and let me know if this does not look right to you.  I would suggest over time we should obsolete data item and data set - although these of course come from IAO so I'm not sure what the procedure is there.

Alan, I've been doing a lot of work last couple of days and reasoning has  slowed slightly.  I haven't adding anything that is incorrect but feel free to have a prod with your new tool if you like :)  If you find something you want to take out can you let me know first please as I don't want to lose any of the new information I added.

Thanks!

James

---

Mélanie Courtot

TFL- BCCRC

675 West 10th Avenue

Vancouver, BC

V5Z 1L3, Canada

[Bjoern Peters]

Two points:
- IAO should decide if they want to go with Melanie's option of having
OBI:'data item' as a more specific version of IAO:data item, or if both
can be considered equivalent. (which means IAO updates their definition,
and we don't need the new OBI one).
- IAO's existing 'measurement datum' is equivalent to the newly created
'measured data item'. There is no need for us to create the new class.

- Bjoern

[James Malone]

Ok.  So first point, I agree with.  We need to push this to IAO.
Second point, ok that's good although this is not the case axiomatically - we need to update the IAO measurement datum with the same restriction I placed on the newly created measured data item.  Also, I note there are several classes asserted under this IAO class (which are OBI so we have control over them) - we should probably place the restriction that these are output_information of assays or instruments on them so they continue to classify correctly (and are not asserted under a defined class) once we make 'measured datum' a defined class. 

In short, we need IAO to make the two edits Bjoern mentions.

James

[Chris Stoeckert]

Hi Melanie, James,

Regarding the things that need to be added, I don't think there was a huge number of terms, so maybe the perl script is an overkill. If we can work out on a clear list of classes and their metadata + their position in the hierarchy I'm happy to do the edits (and I volunteer James too ;) )

Thanks! Here's the list as I see it.

Chris

data/ data item, measured data item, derived data item: already added by James and discussed

data format specification, objective format specification, study design: agree with Plan and Planned Process proposals

software: already in OBI. No changes based on OBI Workshop.

data collection  # already in OBI as data set

subclass of: information content entity

definition: data collection is an information content entity that has two or more data items as parts collected because they are related in some way. Examples of data collections include data items about the same thing,  data items from the same assay, study, or investigation, data items about the same quality.

synonym: data set

editor note: Originally included the bundling was for some purpose but that requires a matching logical definition. Removed “without implication of homogeneity or other structural constraints, at least at the broadest level” as it did not add anything.

definition source: OBI workshop

data representational model

subclass of: information entity about a realizable

definition: data representational model is an information content entity of the  relationships between data items. Examples include gene regulatory graph model, phylogenetic tree, protein interaction network. A data representational model is encoded in a data format specification such as for cytoscape or biopax.

synonym: data structure, data structure specification

definition source: OBI workshop

independent variable

subclass of: information entity about a realizable

definition: independent variable is an information content entity that is varied within an experiment (signifying, for example, the 'control' and 'experimental' groups).

editor note: Replace with definition from Feb 2, 2009 OBI Workshop

synonym: experimental factor

definition source: OBI workshop

hypothesis # currently in OBI as a Role. 

subclass of: information content entity

definition: hypothesis is an information content entity that is a reasoned proposal which is intended to be tested. Examples include: Tamoxifen cures cancer; Gene A regulates pathway B. Investigation has objective (hypothesis testing objective) which has part hypothesis

editor note: falsifiable versus testable

definition source: OBI workshop

study result

subclass of: information content entity

definition: Study Result is an information content entity  that is a specified data output of a study.

definition source: OBI workshop

study interpretation # replaces conclusion under narrative object?

subclass of: information content entity

definition:  Study interpretation is an information content entity interpreting a study result.  Examples include whether a hypothesis is false, the study failed to address the hypothesis, the study results have led to new hypotheses.

editor note: current definition is a narrative object which can be published in a paper summarizing and interpreting a protocol application

synonym: conclusion

definition source: OBI workshop

report # would replace narrative object

subclass of: information content entity

definition: A report is an information content entity assembled by an author for the purpose of providing information for the audience. 

A report is the output of a documenting process and has the objective to be consumed by a specific audience. Topic of the report is on something that has completed. A report is not a single figure. Examples of reports are journal article, patent application, grant progress report, case report (not patience record)

editor note: would replace narrative object

definition source: OBI workshop

report element # would replace report display element

subclass of: information content entity

definition: A report element is an information content entity in which information is presented and consumed by a human being. Examples of report elements are figure (dot plot), table, text portion … (may include a movie or audio clip on a web page).

editor note: needs some more work (clarify relations).

definition source: OBI workshop

measurement unit label # already in OBI

subclass of: label

definition: measurement unit label is as a label that is part of a scalar measurement data item and denotes a unit of measure. 

editor note: instances of unit label would be drawn from the PATO units of measurement ontology or equivalent. 

definition source: Alan Ruttenberg, OBI Workshop

!! Change location of label in hierarchy from subclass of data item to subclass of information content entity because label does not match logical constraints for data item

!! change scalar measurement datum to scalar measurement data item

numeral # already in OBI

!! OK as is however need to change location of symbol in hierarchy from subclass of data item to subclass of information content entity because label does not match logical constraints for data item

Hi all,

<data.png>

[James Malone]

Great thanks Chris.  I'll let Melanie have first pass at these and I can pick up the bits she doesn't have time to finish.  Melanie if you can delegate what you'd like me to do that's fine. 

JM

[Melanie Courtot]

Thanks :)

See my comments inline.

On 16-Feb-09, at 2:17 AM, James Malone wrote:

Great thanks Chris.  I'll let Melanie have first pass at these and I can pick up the bits she doesn't have time to finish.  Melanie if you can delegate what you'd like me to do that's fine. 

JM

On Sun, Feb 15, 2009 at 11:00 PM, Chris Stoeckert <stoe...@pcbi.upenn.edu> wrote:

Hi Melanie, James,

Regarding the things that need to be added, I don't think there was a huge number of terms, so maybe the perl script is an overkill. If we can work out on a clear list of classes and their metadata + their position in the hierarchy I'm happy to do the edits (and I volunteer James too ;) )

Thanks! Here's the list as I see it.

Chris

data/ data item, measured data item, derived data item: already added by James and discussed

data format specification, objective format specification, study design: agree with Plan and Planned Process proposals

software: already in OBI. No changes based on OBI Workshop.

data collection  # already in OBI as data set

subclass of: information content entity

definition: data collection is an information content entity that has two or more data items as parts collected because they are related in some way. Examples of data collections include data items about the same thing,  data items from the same assay, study, or investigation, data items about the same quality.

synonym: data set

editor note: Originally included the bundling was for some purpose but that requires a matching logical definition. Removed "without implication of homogeneity or other structural constraints, at least at the broadest level" as it did not add anything.

definition source: OBI workshop

data set is actually imported from IAO: IAO_0000100

The data item discussion needs to be pushed to IAO (as mentioned by Bjoern). Either the OBI data item is equivalent to the IAO one and then IAO updates their definition, or they are not the same and OBI subclasses the IAO data item. I would suggest to decide on that one, and then address the data set case?

data representational model

subclass of: information entity about a realizable

definition: data representational model is an information content entity of the  relationships between data items. Examples include gene regulatory graph model, phylogenetic tree, protein interaction network. A data representational model is encoded in a data format specification such as for cytoscape or biopax.

synonym: data structure, data structure specification

definition source: OBI workshop

I could add into OBI, but maybe we want to submit 

independent variable

subclass of: information entity about a realizable

definition: independent variable is an information content entity that is varied within an experiment (signifying, for example, the 'control' and 'experimental' groups).

editor note: Replace with definition from Feb 2, 2009 OBI Workshop

synonym: experimental factor

definition source: OBI workshop

hypothesis # currently in OBI as a Role. 

subclass of: information content entity

definition: hypothesis is an information content entity that is a reasoned proposal which is intended to be tested. Examples include: Tamoxifen cures cancer; Gene A regulates pathway B. Investigation has objective (hypothesis testing objective) which has part hypothesis

editor note: falsifiable versus testable

definition source: OBI workshop

study result

subclass of: information content entity

definition: Study Result is an information content entity  that is a specified data output of a study.

definition source: OBI workshop

what about a defined class for that one? for example, conclusion is a narrative object, but also a study result. The logical definition would be the same, except that it wouldn't be asserted under ICE.

study interpretation # replaces conclusion under narrative object?

subclass of: information content entity

definition:  Study interpretation is an information content entity interpreting a study result.  Examples include whether a hypothesis is false, the study failed to address the hypothesis, the study results have led to new hypotheses.

editor note: current definition is a narrative object which can be published in a paper summarizing and interpreting a protocol application

synonym: conclusion

definition source: OBI workshop

Study interpretation a specified data output of a study, and as such would get inferred by the reasoner under the defined class study result, even though it would replace conclusion in the asserted hierarchy.

I am not sure if the 2 are actually synonyms: could I have several study interpretations (e.g. James thinks A, I think B) and then one conclusion (Bjoern publishes the paper on that study and decides that C, his interpretation, was the final conclusion of the study)?

report # would replace narrative object

subclass of: information content entity

definition: A report is an information content entity assembled by an author for the purpose of providing information for the audience. 

A report is the output of a documenting process and has the objective to be consumed by a specific audience. Topic of the report is on something that has completed. A report is not a single figure. Examples of reports are journal article, patent application, grant progress report, case report (not patient record)

editor note: would replace narrative object

definition source: OBI workshop

to be pushed to IAO.

report element # would replace report display element

subclass of: information content entity

definition: A report element is an information content entity in which information is presented and consumed by a human being. Examples of report elements are figure (dot plot), table, text portion … (may include a movie or audio clip on a web page).

editor note: needs some more work (clarify relations).

definition source: OBI workshop

to be pushed to IAO as well.

measurement unit label # already in OBI

subclass of: label

definition: measurement unit label is as a label that is part of a scalar measurement data item and denotes a unit of measure. 

editor note: instances of unit label would be drawn from the PATO units of measurement ontology or equivalent. 

definition source: Alan Ruttenberg, OBI Workshop

in IAO too.

!! Change location of label in hierarchy from subclass of data item to subclass of information content entity because label does not match logical constraints for data item

Sorry if I missed something, but I don't understand that one.

!! change scalar measurement datum to scalar measurement data item

in IAO

numeral # already in OBI

!! OK as is however need to change location of symbol in hierarchy from subclass of data item to subclass of information content entity because label does not match logical constraints for data item

same as above - I am not following. Would you mind telling me a bit more?

In summary, most of those are in IAO, so I need to submit them under the tracker there.

If we can just clarify the few points above, I'll add them under the IAO tracker, and will leave 2 weeks for other IAO members to object if needed. At which point I'll implement the modifications in the IAO file.

An other side note regarding the definition source: we are in the process of trying to homogenize them. Most of the above read "definition source: OBI workshop", would everybody be happy to consider that equivalent to "definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"? 

Melanie

[Bjoern Peters]

More comments inline:

I would like to take this term, and place it under the Plan and Planned
process branch. I can take care of the definitions etc. from the workshop.

>> hypothesis # currently in OBI as a Role.
>> subclass of: information content entity
>> definition: hypothesis is an information content entity that is a
>> reasoned proposal which is intended to be tested. Examples
>> include: Tamoxifen cures cancer; Gene A regulates pathway
>> B. Investigation has objective (hypothesis testing objective)
>> which has part hypothesis
>> editor note: falsifiable versus testable
>> definition source: OBI workshop
>>
>> study result
>> subclass of: information content entity
>> definition: Study Result is an information content entity that
>> is a specified data output of a study.
>> definition source: OBI workshop
>>
>
> what about a defined class for that one? for example, conclusion is a
> narrative object, but also a study result. The logical definition
> would be the same, except that it wouldn't be asserted under ICE.
>

Yes, it can be a defined class. It will be a subclass of 'data item'
though, which 'conclusion' is not (see below)

>
>>
>> study interpretation # replaces conclusion under narrative object?
>> subclass of: information content entity
>> definition: Study interpretation is an information content
>> entity interpreting a study result. Examples include whether a
>> hypothesis is false, the study failed to address the
>> hypothesis, the study results have led to new hypotheses.
>> editor note: current definition is a narrative object which can
>> be published in a paper summarizing and interpreting a protocol
>> application
>> synonym: conclusion
>> definition source: OBI workshop
>>
> Study interpretation a specified data output of a study, and as such
> would get inferred by the reasoner under the defined class study
> result, even though it would replace conclusion in the asserted hierarchy.
>
> I am not sure if the 2 are actually synonyms: could I have several
> study interpretations (e.g. James thinks A, I think B) and then one
> conclusion (Bjoern publishes the paper on that study and decides that
> C, his interpretation, was the final conclusion of the study)?
>

They are not synonyms. 'Results' are like 'data items', e.g. 'the
average weight of mice in group A was lower than in group B with a
p-value of p=0.02', is a result. A conclusion or now 'study
interpretation' is going beyond stating what the data is and
generalizing / interpreting it, e.g. 'This confirms our hypothesis that
feeding mice sugar makes them gain weight'.

I agree with Chris, Currently: 'measurement unit label' is_a 'data
item', which is not true.

> In summary, most of those are in IAO, so I need to submit them under
> the tracker there.
> If we can just clarify the few points above, I'll add them under the
> IAO tracker, and will leave 2 weeks for other IAO members to object if
> needed. At which point I'll implement the modifications in the IAO file.
>
> An other side note regarding the definition source: we are in the
> process of trying to homogenize them. Most of the above read
> "definition source: OBI workshop", would everybody be happy to
> consider that equivalent to "definition source :GROUP:OBI" with an
> editor note "2009-xx-xx: work on this term has been finalized during
> the OBI workshop winter 2009"?

Fine with me
Bjoern

[Chris Stoeckert]

And here are my comments also in line:

I think I go along with the proposal to create an OBI data item as
defined at the OBI workshop that is currently more restricted than the
IAO datum. This is related to comments below on subclasses.

>>
>>
>>>
>>> data representational model
>>> subclass of: information entity about a realizable
>>> definition: data representational model is an information content
>>> entity of the relationships between data items. Examples include
>>> gene regulatory graph model, phylogenetic tree, protein
>>> interaction network. A data representational model is encoded in
>>> a data format specification such as for cytoscape or biopax.
>>> synonym: data structure, data structure specification
>>> definition source: OBI workshop
>>>
>>
>> I could add into OBI, but maybe we want to submit

The key part of the definition for me is relationships between data
items where data items are measured data and derived data. I vote for
OBI.

>>
>>>
>>>
>>> independent variable
>>> subclass of: information entity about a realizable
>>> definition: independent variable is an information content entity
>>> that is varied within an experiment (signifying, for example, the
>>> 'control' and 'experimental' groups).
>>> editor note: Replace with definition from Feb 2, 2009 OBI
>>> Workshop
>>> synonym: experimental factor
>>> definition source: OBI workshop
>>>
> I would like to take this term, and place it under the Plan and
> Planned process branch. I can take care of the definitions etc. from
> the workshop.

OK. Thanks.

Agree with Bjoern. I don't think that study interpretation is a
specified data output.

label is a subclass of IAO data item (datum). It is not a subclass of
OBI-defined data item (union of measured data item and derived data
item).

>>
>>
>>
>>> !! change scalar measurement datum to scalar measurement data
>>> item
>>>
>>
>> in IAO
>>
>>>
>>> numeral # already in OBI
>>> !! OK as is however need to change location of symbol in
>>> hierarchy from subclass of data item to subclass of information
>>> content entity because label does not match logical constraints
>>> for data item
>>>
>>
>> same as above - I am not following. Would you mind telling me a bit
>> more?
>>
> I agree with Chris, Currently: 'measurement unit label' is_a 'data
> item', which is not true.

As above.

>
>> In summary, most of those are in IAO, so I need to submit them
>> under the tracker there.
>> If we can just clarify the few points above, I'll add them under
>> the IAO tracker, and will leave 2 weeks for other IAO members to
>> object if needed. At which point I'll implement the modifications
>> in the IAO file.
>>
>> An other side note regarding the definition source: we are in the
>> process of trying to homogenize them. Most of the above read
>> "definition source: OBI workshop", would everybody be happy to
>> consider that equivalent to "definition source :GROUP:OBI" with an
>> editor note "2009-xx-xx: work on this term has been finalized
>> during the OBI workshop winter 2009"?
> Fine with me
> Bjoern

Me too.
Chris

[Melanie Courtot]

Chris - thanks for the explanations - I had trouble following between the various data items :)

A bit more in line.

  data/ data item, measured data item, derived data item: already

  added by James and discussed

  data format specification, objective format specification, study

  design: agree with Plan and Planned Process proposals

  software: already in OBI. No changes based on OBI Workshop.

  data collection  # already in OBI as data set

  subclass of: information content entity

  definition: data collection is an information content entity that

  has two or more data items as parts collected because they are

  related in some way. Examples of data collections include data

  items about the same thing,  data items from the same assay,

  study, or investigation, data items about the same quality.

  synonym: data set

  editor note: Originally included the bundling was for some

  purpose but that requires a matching logical definition. Removed

  "without implication of homogeneity or other structural

  constraints, at least at the broadest level" as it did not add

  anything.

  definition source: OBI workshop

data set is actually imported from IAO: IAO_0000100

The data item discussion needs to be pushed to IAO (as mentioned by Bjoern). Either the OBI data item is equivalent to the IAO one and then IAO updates their definition, or they are not the same and OBI subclasses the IAO data item. I would suggest to decide on that one, and then address the data set case?

I think I go along with the proposal to create an OBI data item as defined at the OBI workshop that is currently more restricted than the IAO datum. This is related to comments below on subclasses.

I am happy with that too. As proposed in a previous mail I will declare measured data item and derived data item as subclasses of the IAO data item, thus making the OBI data infer under the IAO data item.

This change has been made and committed (see attached pic for the inferred hierarchy).

Let me know if this is not what was expected - I'll wait for feedback before making the other updates.

  data representational model

  subclass of: information entity about a realizable

  definition: data representational model is an information content

  entity of the  relationships between data items. Examples include

  gene regulatory graph model, phylogenetic tree, protein

  interaction network. A data representational model is encoded in

  a data format specification such as for cytoscape or biopax.

  synonym: data structure, data structure specification

  definition source: OBI workshop

I could add into OBI, but maybe we want to submit

The key part of the definition for me is relationships between data items where data items are measured data and derived data. I vote for OBI.

OK.

  independent variable

  subclass of: information entity about a realizable

  definition: independent variable is an information content entity

  that is varied within an experiment (signifying, for example, the

  'control' and 'experimental' groups).

  editor note: Replace with definition from Feb 2, 2009 OBI Workshop

  synonym: experimental factor

  definition source: OBI workshop

I would like to take this term, and place it under the Plan and Planned process branch. I can take care of the definitions etc. from the workshop.

OK. Thanks.

independent variable specification has already been added in IAO, see http://purl.obofoundry.org/obo/iao.owl, as IAO_0000011.

Definition source: Bjoern Peters.

They are not synonyms. 'Results' are like 'data items', e.g. 'the average weight of mice in group A was lower than in group B with a p-value of p=0.02', is a result. A conclusion or now 'study interpretation' is going beyond stating what the data is and generalizing / interpreting it, e.g. 'This confirms our hypothesis that feeding mice sugar makes them gain weight'.

Agree with Bjoern. I don't think that study interpretation is a specified data output.

Ok. I'll add the example from Bjoern.

Ok, makes more sense now - so nothing to be done for this, it stays as subclass of the IAO data item. It won't get inferred under the OBI data though.

  !! change scalar measurement datum to scalar measurement data item

in IAO

  numeral # already in OBI

  !! OK as is however need to change location of symbol in

  hierarchy from subclass of data item to subclass of information

  content entity because label does not match logical constraints

  for data item

same as above - I am not following. Would you mind telling me a bit more?

I agree with Chris, Currently: 'measurement unit label' is_a 'data item', which is not true.

As above.

as above too :)

In summary, most of those are in IAO, so I need to submit them under the tracker there.

If we can just clarify the few points above, I'll add them under the IAO tracker, and will leave 2 weeks for other IAO members to object if needed. At which point I'll implement the modifications in the IAO file.

An other side note regarding the definition source: we are in the process of trying to homogenize them. Most of the above read "definition source: OBI workshop", would everybody be happy to consider that equivalent to "definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"?

Fine with me

Bjoern

Me too.
Chris

Perfect.

Melanie

[James Malone]

So presumably 'data set' as a class is now redundant?  If derived data are data items, then things like 'classified data set' should also be data items so I should remove them from under data set it seems.  (This is reason data sets are not classifying under 'derived data' presently)

James

[Bjoern Peters]

I don't think it is redundant, and think it will classify under
'processed data item', once it is defined as the output of a 'data
collecting' process.

[Larry Hunter]

Sorry I couldn't make it to the workshop. I do have one comment,
though.

On Feb 15, 4:00 pm, Chris Stoeckert <stoec...@pcbi.upenn.edu> wrote:

> hypothesis # currently in OBI as a Role.
> subclass of: information content entity
> definition: hypothesis is an information content entity that is a  
> reasoned proposal which is intended to be tested. Examples include:  
> Tamoxifen cures cancer; Gene A regulates pathway B. Investigation has  
> objective (hypothesis testing objective) which has part hypothesis
> editor note: falsifiable versus testable
> definition source: OBI workshop

A related definition (also using "reasoned proposal") was discussed
here last year:
https://sourceforge.net/tracker/?func=detail&atid=886178&aid=1887478&group_id=177891

And in light of that discussion, perhaps something more like

definition: hypothesis is an information content entity about an
assertion that can be tested. Examples include..."

[Chris Stoeckert]

Hi Larry,
I think changing "reasoned proposal" to "assertion" could work. The
sense I got for "reasoned proposal" was that it was something
reasonable to test. Simply stating that an assertion that can be
tested is a hypothesis loses that sense but that may not matter.

Do others have objections to the change?

Thanks,
Chris

[Melanie Courtot]

Hi all,

Regarding Larry's email: I am happy which the change to assertion too. (Replying here to get thread with all classes below)

As I didn't hear any objection to the way data (OBI data, i.e. measured and derived data) has been added to the file, I will continue with remaining terms. (apologies for the delay, not much time atm)

Trying to summarize what remains:

1. hypothesis # currently in OBI as a Role (to be removed from Role)

subclass of: information content entity

definition: hypothesis is an information content entity that is an assertion which is intended to be tested. 

Example of usage:Tamoxifen cures cancer; Gene A regulates pathway B. Investigation has objective (hypothesis testing objective) which has part hypothesis

editor note: falsifiable versus testable

definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"

2. data representational model

subclass of: information entity about a realizable

definition: data representational model is an information content entity of the  relationships between data items. 

 A data representational model is encoded in a data format specification such as for cytoscape or biopax.

Examples of usage: gene regulatory graph model, phylogenetic tree, protein interaction network. 

synonym: data structure, data structure specification

definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"

3.   study result

subclass of: information content entity

 definition: Study Result is an information content entity  that is a specified data output of a study.

definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"

I will add these 3 classes before Monday if no objection.

4.   study interpretation # replaces conclusion under narrative object?

subclass of: information content entity

definition:  Study interpretation is an information content entity interpreting a study result.  Examples include whether a hypothesis is false, the study failed to address the  hypothesis, the study results have led to new hypotheses. editor note: current definition is a narrative object which can be published in a paper summarizing and interpreting a protocol application 

 synonym: conclusion

definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"

I am unsure how to deal with that one: currently conclusion is imported from IAO, but refers specifically to  paper publication. We can either add an OBI conclusion (i.e. study interpretation), or liaise with IAO to modify the IAO conclusion.

To be submitted to IAO (volunteers?)

report # would replace narrative object

subclass of: information content entity

definition: A report is an information content entity assembled by an author for the purpose of providing information for the audience. 

A report is the output of a documenting process and has the objective to be consumed by a specific audience. Topic of the report is on something that has completed. A report is not a single figure. Examples of reports are journal article, patent application, grant progress report, case report (not patience record)

editor note: would replace narrative object

definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"

report element # would replace report display element

subclass of: information content entity

definition: A report element is an information content entity in which information is presented and consumed by a human being. Examples of report elements are figure (dot plot), table, text portion … (may include a movie or audio clip on a web page).

editor note: needs some more work (clarify relations).

definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"

measurement unit label # already in OBI

subclass of: label

definition: measurement unit label is as a label that is part of a scalar measurement data item and denotes a unit of measure. 

editor note: instances of unit label would be drawn from the PATO units of measurement ontology or equivalent. 

definition source: Alan Ruttenberg

definition source :GROUP:OBI" with an editor note "2009-xx-xx: work on this term has been finalized during the OBI workshop winter 2009"

+  change scalar measurement datum to scalar measurement data item

Thanks,

Melanie

They are not synonyms. 'Results' are like 'data items', e.g. 'the average weight of mice in group A was lower than in group B with a p-value of p=0.02', is a result. A conclusion or now 'study interpretation' is going beyond stating what the data is and generalizing / interpreting it, e.g. 'This confirms our hypothesis that feeding mice sugar makes them gain weight'.

[Chris Stoeckert]

Melanie,

No objections here. Thanks!

Chris

[Melanie Courtot]

Hi all,

The 3 classes have been added.

I will submit the other items on the IAO tracker and will keep the DENRIE list posted with relevant URLs.

Side note: In the current files, all information about a realizable entity are in the PAPP branch file, I guess historically as they were originally in the Plan file. I added data representational model in the DENRIE file.

In order to ease future editing, I would like to homogenize this, either having data representational model in the PAPP branch file with other information about a realizable entity, or we could group all "information" items (i.e., all subclasses of ICE) in the DENRIE file.

Even though the names of the files would be less relevant, my preference goes to the latter - having all information in one place seems clearer to me. On the other hand, all study design terms and objective are currently under information entity about a realizable, and they are clearly linked to processes.

What do others think?

Cheers,

Melanie

[James Malone]

Cheesr MC.  I think they should go in DENRIE and I also think objectives should go in DENRIE (there being linked with processes is irrelevant in some ways, as most things in OBI are.  I think it's probably more important to keep children of ICE in the DENRIE file).

Jm

[Chris Stoeckert]

I agree. Would be great to have essentially an ICE branch file.

Thanks Melanie for your efforts on this.

Chris

[Bjoern Peters]

I would want to keep the realizables in our branch, as it is the 'plan +
planned process branch'. If not, we will have to constantly switch files
when editing, or send large numbers of requests to Denrie. Also, I want
to point out that with this logic, all 'data transformations' + the rest
of our branch would need to be merged into a 'process branch', and all
instruments + biomaterials should be merged into the 'material' branch.

[James Malone]

Fine, the branch structure should primarily aid development so if this is detrimental I would support leaving where it is then.