Software Subclassing

[James Malone]

Hi All (specifically DENRIE branch),

I'd like to start adding some subclasses to software for our GenePattern use case.  I would only wish to add a few to start with but hope to add more later.  They will be in the form of some of the software modules used by GenePattern http://www.broad.mit.edu/cancer/software/genepattern/  I realise software is part of IAO but I presume that subclasses should go in the DENRIE branch file.  Does this sound ok?  I didn't want to touch anything without consulting you first, although I envisage their should be no impact on anything else in DENRIE.  If anyone has any thoughts on how software should appear I'd be happy to hear them, otherwise I presume they are just subclasses of the software class.

Thanks,

James

[James Malone]

That should be 'there should be' not 'their should be'.  :)

J

[Alan Ruttenberg]

On Tue, Jan 27, 2009 at 3:10 AM, James Malone <james....@gmail.com> wrote:
> Hi All (specifically DENRIE branch),
>
> I'd like to start adding some subclasses to software for our GenePattern use
> case. I would only wish to add a few to start with but hope to add more
> later. They will be in the form of some of the software modules used by
> GenePattern http://www.broad.mit.edu/cancer/software/genepattern/ I realise
> software is part of IAO but I presume that subclasses should go in the
> DENRIE branch file.

Yes, certainly to the extent that they are in the domain of biomedical
investigations.

If they are specific modules, then they would be instances, not
classes, though, right? What are some specific examples of things you
want to name.

-Alan

[Chris Stoeckert]

Hi James,
Looking at the link you gave:
http://www.broad.mit.edu/cgi-bin/cancer/software/genepattern/modules/gp_modules.cgi

I propose putting the categories in the drop down menu in OBI as
software types (e.g., clustering, imputation - note these need some
cleaning up to be consistent) but put the actual list as instances
(e.g., ARACNE. CaArray2.10Importer).

Cheers,
Chris

On Jan 28, 2009, at 10:18 AM, James Malone wrote:

> Alan,
>
> After our quick chat offline today, here are some examples of the
> software we'd like for now. They're all software modules from
> GenePattern. The ones I'm interested in are specifically those that
> perform DTs. Here is a list although this is a greater set than we
> would need
> http://www.broad.mit.edu/cgi-bin/cancer/software/genepattern/modules/gp_modules.cgi
> Whether things under software should be instances or classes, I'm
> still
> not clear about.
>
> James

>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by:
>> SourcForge Community
>> SourceForge wants to tell your story.
>> http://p.sf.net/sfu/sf-spreadtheword
>> _______________________________________________
>> Obi-datatrfm-branch mailing list
>> Obi-datat...@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/obi-datatrfm-branch
>>
>
>
> --
> European Bioinformatics Institute,
> Wellcome Trust Genome Campus,
> Hinxton,
> Cambridge, CB10 1SD,
> United Kingdom
> Tel: + 44 (0) 1223 494 676
> Fax: + 44 (0) 1223 492 468
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> SourcForge Community
> SourceForge wants to tell your story.
> http://p.sf.net/sfu/sf-spreadtheword
> _______________________________________________
> Obi-datatrfm-branch mailing list
> Obi-datat...@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/obi-datatrfm-branch

[James Malone]

Hi Chris,

I think this was more along the lines of what Alan was explainging to me too.  I haven't figured out the types yet but this seems like a sensible place to start.  Are we allowed to add instances to the file?

James

[Melanie Courtot]

Instances have been going in the file DataFormatSpecification.owl (named that way as data format specification instances were the only instances until now).

Exception to this is the Affymetrix instance of organization that has been added to the Biomaterial file, but when we did it we clearly said it was there just as illustration and we expect instances to go in their own file.

My preference would be having instances in a separate file. If we have lists of very specific instances I would prefer having them in different files, we could create a file SoftwareInstances.owl for example?

Melanie

---

Mélanie Courtot

TFL- BCCRC

675 West 10th Avenue

Vancouver, BC

V5Z 1L3, Canada

[Alan Ruttenberg]

Hi James,

Looking at the first couple, Aracne and CART, I would say they are
definitely instances. There is a bit more to say, not sure how much
you want to. First, each version would be a different instance.
Second, there is a difference between the Java implementation and the
R implementation. For the java one the nice thing is that since it is
cross platform we don't have to worry about the status of different
binaries on different platforms. Another difference is that in both
cases there is the software (the "executable") and the source code.
The source code is also a GDP, however in the case of R the executable
is a part of the source code, I think, at least at the level of GDCs.

-Alan

On Wed, Jan 28, 2009 at 10:18 AM, James Malone <mal...@ebi.ac.uk> wrote:
> Alan,
>
> After our quick chat offline today, here are some examples of the software
> we'd like for now. They're all software modules from GenePattern. The ones
> I'm interested in are specifically those that perform DTs. Here is a list
> although this is a greater set than we would need
> http://www.broad.mit.edu/cgi-bin/cancer/software/genepattern/modules/gp_modules.cgi
> Whether things under software should be instances or classes, I'm still not
> clear about.
>
> James
>
>
>
> Alan Ruttenberg wrote:
>>