[Bioperl-l] Offering to help

Chris Mungall cjm at berkeleybop.org
Thu Mar 5 18:48:07 EST 2009


On Mar 4, 2009, at 9:49 PM, Kevin Clancy wrote:

>
> Hi Guys
> You might also want to check out Ontologies  for Biomedical  
> Investigations - OBI
> http://obi-ontology.org/page/Main_Page
> I've been following this thread and am intreagued by the use of  
> roles in Moose. So if I have a defined ontology with classes, each  
> of which has roles, can that be dropped into the Moose framework?

Hi Kevin

Can you explain what you have in mind here? Particularly with respect  
to classes in ontologies having roles.

The idea of having the ontology drive the object model is appealing,  
but there are a number of gotchas here. There's a significant  
impedance mismatch between the semantics of classes, roles and so in  
between ontology formalisms and object-oriented concepts.

> The main reason for asking is that a lot of work has gone into OBI  
> including normalizing the ontology to others in the national center  
> for biomedical ontologies http://bioontology.org/ and obo foundry http://www.obofoundry.org/ 
> . having some of this ontological work extended into the behaviour  
> of the Bio* languages would be an interesting extension of these  
> projects. My main concern with FuGE is that I don't know how  
> compatable it is with these communities.

I don't really know much about FuGE.

But speaking of ontologies and Moose, I have been thinking it's about  
time to update the creaky old go-perl framework, fixing it to support  
some of the more modern features of OBO and OWL. I'd like this to be  
done in such a way that bioperl2 can use this. If Moose is the way to  
go we'll move in this direction, though probably not until after I'm  
convinced that Moose and DBIx::Class can play well together.

>
> kevin
>
>> Date: Tue, 3 Mar 2009 14:05:22 -0600
>> From: kanzure at gmail.com
>> To: awitney at sgul.ac.uk
>> CC: cjfields at illinois.edu; bioperl-l at lists.open-bio.org; kanzure at gmail.com
>> Subject: Re: [Bioperl-l] Offering to help
>>
>> On Tue, Mar 3, 2009 at 12:49 PM, Adam Witney wrote:
>>> On 3 Mar 2009, at 17:21, Bryan Bishop wrote:
>>>> On Tue, Mar 3, 2009 at 11:14 AM, Chris Fields wrote:
>>>>> For a Moose-based bioperl implementation I suggest a separate repo
>>>>> completely.  We're using svn currently on dev.open-bio.org,  
>>>>> though I and
>>>>> a
>>>>> few others are also using git.  I'm neutral on the matter but it's
>>>>> possible
>>>>> the consensus may be to keep everything in the open-bio svn repo  
>>>>> (not
>>>>> everyone has git or uses it).
>>>>
>>>> Yes, I've been figuring that the protocols work should be done in a
>>>> separate repository for now anyway, I just haven't got around to
>>>> starting the code, although with XML::Simple it should be easier or
>>>> quicker to do than I'm making it out to be. Part of the issue is  
>>>> that
>>>> there needs to be some good thought put into what exactly the  
>>>> schema
>>>> should be or how exactly to represent the information, which I've
>>>> always had a little bit of anxiety over. It would be fantastic if I
>>>> could somehow come up with an easy format and a wizard creation  
>>>> tool
>>>> front-end for something like protocol-online.org, since all of  
>>>> those
>>>> protocols truly deserve to be in some sort of microformat, whether
>>>> CLP-ML or something based off of YAML. Maybe I'll post a thread  
>>>> soon
>>>> outlining what a discussion would have to go over to make sure  
>>>> there's
>>>> no shooting of self in the foot, to come up with that relatively
>>>> mature API or microformat, and see what others on the list have  
>>>> to say
>>>> about it? My background is more computer science than  
>>>> bioinformatics
>>>> and biology (even though I've been in biology labs longer  
>>>> (strange)),
>>>> so it would be great to get some support on that front.
>>>
>>> The FuGE folks have modeled Protocols as part of their object  
>>> model. I
>>> haven't looked at it recently but it is supposed to be fairly  
>>> generic for
>>> any kind of protocol. We have used the MAGE (FuGE's almost  
>>> predecessor)
>>> version of Protocol within our database for several years and it  
>>> covers most
>>> cases.
>>>
>>> The FuGE project pages can be found here:
>>>
>>> http://fuge.sourceforge.net/
>>>
>>> Maybe that would be a good place to start?
>>
>> Thanks Adam, I hadn't seen that before. I glanced over their API  
>> documentation.
>>
>> http://fuge.sourceforge.net/dev/V1Final/FuGEv1-refManual.html#Action
>>
>> So, it looks like they are defining individual steps with an ordinal
>> number (which is good), also equipment via make and model tags. One  
>> of
>> the ideas that I have been working on is the concept of instructions
>> being "solved" by equipment: so you can either execute a PCR protocol
>> by manually dunking test tubes into warm and cold baths, or you can
>> use a thermocycler or microfluidic chip using some standardized
>> interface. This is also the same problem that "semantic web"  
>> advocates
>> have to put up with: UDDI, WSDL, B2B, "service discovery", etc.,
>> without overly restricting the standards to the point that nobody can
>> do anything new. But anyway, getting off track.
>>
>> There is a link to "example files"- which is broken-
>> http://fuge.sourceforge.net/dev/V1Final/Instances/
>>
>> Although this one does work:
>> http://fuge.sourceforge.net/presentation/master_example.xml
>> though it might be too meta for me :-).
>>
>> Ultimately I'd like to come up with a system that I can dump into  
>> EXPO
>> (like The Robot Scientist) but also into a human readable format
>> (comparably, instructables?).
>>
>> http://expo.sf.net/
>> videos: http://www.aber.ac.uk/compsci/Research/bio/robotsci/video/
>>
>> - Bryan
>> http://heybryan.org/
>> 1 512 203 0507
>>
>> _______________________________________________
>> Bioperl-l mailing list
>> Bioperl-l at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
>
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioperl-l
>



More information about the Bioperl-l mailing list