[obs] proportional phenotypes

[Shawn Bowers]

Hi,

Representing the value (as in the 5 cm example below) would be trivial
to do in OBOE. However, representing a magnitude w.r.t. a different
property would be trickier. I haven't seen an example like this in the
datasets we've looked at in the KNB, and so haven't really thought
about this case.  But, I think it could be defined by introducing a
characteristic ...

In OBOE, a measurement is a combination of an entity (the petiole in
this case),  a characteristic (e.g., length, width, mass, etc.) of the
entity, a value (e.g., 5), , and a measurement standard (e.g., cm).

In this case, it seems like the easiest thing to do would be to have a
characteristic that captures "length compared to width" ... and then
use the default ratio measurement scale as the standard, or else
define a specialized unit for this. It seems like you'd want the
characteristic to be defined in such a way that it implies both a
length and a width were measured. This would allow for discovery
queries that search for petiole length and width to both return the
measurement (assuming this is the desired outcome).

Shawn


On Thu, Feb 17, 2011 at 4:13 PM, Hilmar Lapp <hlapp at nescent.org> wrote:
> I thought this would be an interesting question for the Observations list as
> well, and am hence forwarding it.
>
> How would "petiole 5x longer than wide" best be represented in an OBOE
> model?
>
> Jim -
>
> your revised suggestion below is exactly along my reaction to your original
> question. Often questions such as yours are framed as "what is the correct
> way to represent this using model Y". Many data models have indeed only one
> or very few correct way of representing some data, but I think this is much
> less true for the EQ model. Obviously something that is syntactically wrong
> or logically inconsistent in OWL would have to be considered "incorrect".
>
> The rationale I was trying to follow in thinking how best to represent this
> wasn't so much about what is the "correct" way, but rather what inferences
> do you want to support with it. Is it important to the ability to reason
> over the semantics that it is a proportionality that inheres in the petiole,
> or that it is a length, and the length in this case just happens to be
> stated relative to some other length, but that doesn't change that what is
> of interest here semantically is that it is a length characteristic that is
> being observed.
>
> Even so, that doesn't mean that in other cases it would really be the
> proportionality that's important.
>
>        -hilmar
>
> On Feb 17, 2011, at 2:36 PM, Jim Balhoff wrote:
>
>> I've been thinking a little more about this and wonder about modeling
>> proportions in a way that is more in line with how we are modeling relative
>> phenotypes and phenotypes with absolute values. At a PATO/Phenoscape/HAO
>> workshop in September, we decided on modeling relative phenotypes something
>> like this:
>>
>> length that inheres_in some petiole and increased_in_magnitude_relative_to
>> some (length that inheres_in some head)
>>
>> where the petiole is described as longer than the head.
>>
>> An absolute value for the length would be like this:
>>
>> length that inheres_in some petiole and has_value some (has_magnitude
>> value 5 and has_unit value cm)
>>
>> By the way I'm not sure what the standard properties are to be used to
>> relate the magnitude, unit, etc.
>>
>> So if we want to explicitly describe the length of one part as
>> proportional to some other part, perhaps we could use the comparison quality
>> as a unit:
>>
>> length that inheres_in some petiole and has_value some (has_magnitude
>> value 5 and has_unit some (length that inheres_in some head))
>>
>> Any opinions?
>>
>> Thanks,
>> Jim
>>
>> On Feb 16, 2011, at 4:00 PM, Jim Balhoff wrote:
>>
>>> Hi,
>>>
>>> Is there an accepted model for representing a proportional phenotype in
>>> OWL? I've taken a look at the properties in PATO such as
>>> has_dividend_quality and has_divisor_quality, but I'm not sure how to put it
>>> all together. Here is the phenotype description:
>>>
>>> character: petiole length
>>> state: petiole 5x longer than wide
>>>
>>> This is what I came up with:
>>>
>>> has_part some (petiole and bearer_of some ('proportionality to' and
>>> (has_dividend_quality some width) and (has_divisor_quality some length)))
>>>
>>> Firstly I'm not sure if I've constructed that correctly. Also, what data
>>> property can I use to connect 'proportionality to' with the value 5?
>>>
>>> Thanks,
>>> Jim
>>>
>>
>> ____________________________________________
>> James P. Balhoff, Ph.D.
>> National Evolutionary Synthesis Center
>> 2024 West Main St., Suite A200
>> Durham, NC 27705
>> USA
>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
>> Pinpoint memory and threading errors before they happen.
>> Find and fix more than 250 security defects in the development cycle.
>> Locate bottlenecks in serial and parallel code that limit performance.
>> http://p.sf.net/sfu/intel-dev2devfeb
>> _______________________________________________
>> Obo-phenotype mailing list
>> Obo-phenotype at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/obo-phenotype
>
> --
> ===========================================================
> : Hilmar Lapp  -:- Durham, NC -:- informatics.nescent.org :
> ===========================================================
>
>
>
> _______________________________________________
> obs mailing list
> obs at ecoinformatics.org
> http://lists.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/obs
>