Re: [Obi-instrument-branch] instrument branch report

[Melanie C. <co...@eb...>]

>
> 1) description of platform: they are composed of multiple instruments
> (instruments aggregates), do we name them with all the individual
> names?
>
> A platform, as the OWL file definition suggests can be a composition
> of divices and software. The first thing we would do would be name the
> platform class. Then we would use a relations, most likely has_part
> and state the parts (particular devices/software) - if a particular
> community has requested the inclusion of a platform and also require
> the parts to be stated

We are on the same page.
Daniel just mentioned the naming issue.E.g.,
liquid_chromatography_mass_spectrometry_platform, do we name the platform
with all the names of the instrument composing it?

As said in the previous mail, I feel this is part of naming conventions
and probably doesn't need to be addressed in our report, because naming
conventions impact other branches first (a homogeneous approach would be
nice) and also because this is not one of our top issue at the moment.
>
>
> 2) proposal to use _ as prefix for helper classes: immediately
> identifiable by peopple
>
> I just dont get this - why would a so called helper class be in the
> ontology? My opinion is they should not. Place terms in the ontology
> based on the knowledge you know at the time of inclusion

Ideally we would also have only curation_complete :-)
As said, I believe this doesn't impact highly now on the status. Keeping
in mind that a draft release should be out in a month, I think we have
more urgent matters to deal with, and I hope the progression toward the
release will make these disappear.
As Daniel mentioned it I will put it under the agenda for general
discussion though.

>
> 3) overlap between biomaterial and instrument (e.g. cDNA vector would
> fit in current definition of artefact object -> proposal from
> biomaterial to use material with role instrument, whereas biomaterial
> would be material with role sample or reagent)
>
> This is the asserted structure which has existed in sepCV for the last
> two years (without the roles and using substance instead of
> material).........., However, This addresses a long standing issue
> within the object hierarchy Matt and I provided an alternative using
> "countables" and "portions".

a pointer to a thread (or better a small summary) would have been nice
here...I'll try to see if I can dig it up on sourceforge (assuming they
didn't loose these emails)

 What you are proposing here does not
> solve the issue. Some things to consider
>
> 3.1 what is the difference between object and material?
I have to go back to the proposal, by I believe it will be object with
role sample or reagent.

> 3.2 The roles would only be realised as the result of a process, not
> in the asserted hierarchy under object

Ok. Role are realized in a process but the bearer of the role exists
independently. Protein can bear the role "sample" will never bear the role
"instrument"

> 3.3 I support the removal of the label "biomaterial" from the object
> hierarchy - it cant happen fast enough

I don't think we are going that way at all in the biomaterial branch. One
option would be for you to post there
http://groups.google.com/group/obi-biomaterial?hl=en (this is outside of
the scope of the instrument branch)

> 3.4 An instrument is not a material - it is composed from materials

cf up. An instrument is an object. so:
--object
------biomaterial (role sample and reagent)
------artefact_object (role instrument)

> 3.5 My propposal would be that directly under object we have
> subclasses of both "material" and "device"

so that goes along what I describe up (assuming (bio)material)) except
that you removed artefact_object. Why?

>
>
> Melanie>>
> I think this means we need to clarify our definition of artefact object. I
> should have explicited more the proposal from biomaterial, they propose to
> define biomaterial as object with role sample or reagent, and to define
> artefact object as object with role Instrument.
> This would allow us to disambiguate, a cDNA vector will never bear the
> role Instrument, and would be classified as biomaterial.
> Melanie>>
>
> You have got it back to front. artefact_object is there as a disjoint
> to biomaterial - remove biomaterial there is no need for artifact
> object.

I don't follow you.
Currently:
--object
-----biomaterial
-----artefact_object
------------device
--------------------Instrument
------------physical document

Ok for the disjoint with biomaterial and artefact_object, that is exactly
what is my issue, the current definition of artefact object doesn't
support this, allowing cDNA vector to be considered as artefact object,
whereas they should go under biomaterial. Having the disjoint shouldn't
prevent us from having a correct definition.


>
> 6)defined classes
>
> Yes we need defined classes in the intrument branch, we need the
> relations that we submitted before Christmas to do this. However we
> can include the defined class
>
> Mass spectrometer, has_part=ion_source, has_part=mass_analyser,
> has_part=ion_detector.
>
> I can put this in the OWL file if it helps the discussion

ok. I could use that with the list of relations we submitted in december
to show that we have a way to sort this out.

>
> 7)commercial terms: where do we put them (branch file, external file..)
>
> My opionion would be to have OBI-core exist as a separate file and to
> address the competancy questions listed here
> https://wiki.cbil.upenn.edu/obiwiki/index.php/Informal_Compentency_Questions
>
>
> 8) sepCV inclusion
>
> No timeline fro inclusion has been discussed yet.
Ok.
Am I correct in assuming that these will most probably not be part of the
draft release then, and would go under our long term plans?


Thanks for the help, I will start preparing the report in 3 hours
(meetings before that), so any last suggestion in that time frame is ok
:-)

Melanie