Menu
▾
▴
Re: [Brickit] [BBF Standards] data exchange: new Pobol draft on Google Code
|
From: Raik G. <rai...@cr...> - 2008-06-01 20:35:56
|
Hi Jason, Jason Morrison wrote: > Raik, > > This is fantastic news - I was actually just looking at the google code > project yesterday (hanging out with some semantic web hackers at > RailsConf) to see if the ontology description was there. I've been > doing some thinking and light hacking on Brickit to expose RDF/XML and > n3 versions of data contained inside. I have next week off, and look > forward to getting some solid time to dig in then. That sounds great! Do you want to have a developer account on the sourceforge project? You could open your own branch or directly check things into the devel branch. My big Biobrick order is going to arrive tomorrow so I am going to be pretty occupied by wet lab work. The minimal brickit plans are to just merge the current devel version (with its much nicer Biobrick view) into the trunk and release it as soon as possible. After that, there won't be much activity from my side for a while. There are anyway some more strategic decisions needed before going on. It's not such a good strategy to expose the Admin interface to normal users. Besides, Brickit is currently using the classic django admin which will be superseeded by the newforms admin branch in the upcoming Django 1.0 . Writing a public interface from scratch (rather than adapting admin) would probably make more sense... if there is the time for that. None of this really affects the RDF/XML export though. Greetings, Raik > > Cheers! > Jason > > On Sun, Jun 1, 2008 at 12:05 PM, Raik Gruenberg <rai...@cr... > <mailto:rai...@cr...>> wrote: > > Hi everyone, > > I've shifted the Pobol ontology files to a new Google Code project > so that we > can track changes and issues: > > http://code.google.com/p/pobol/ > > The initial release contains the ontology in OWL/XML and Turtle > format, plus an > example Biobrick XML file. It has changed with respect to the > previous version 0.1: > > * restrictions clarify which properties are obligatory for certain > classes. For > example, a BiobrickFormat needs exactly one 'prefix', 'suffix', and > one 'self_scar' > > * all sequence-related properties (partSequence, prefix, suffix, > scar) are now > derived from a common dnaSequence ancestor -- this should allow for > some more > general queries and data handling in the future. > > I may not have much time to work on this during the next weeks. > Please review > the wiki and see where you can contribute. It would be great to get more > involvement from people working on the upcoming registries or on > concrete > web/software projects that import / export Biobrick data. Early > coordination can > spare us all a lot of later work and frustration. > > Open issues: > -------------- > (see also http://code.google.com/p/pobol/issues/list) > > * earlier, Jason had suggested to tie the Biobrick Format to the > PhysicalDNA > rather than to a given Biobrick record -- surely worth considering. > Though, it > would be good to know how the MIT registry handles this before > making a decision. > > * The sequenceOntology project provides OWL terms for sequence > features and > locations -- I think it would make sense to already include such feature > properties in the initial pobol draft. > > * Looking at the evolving structure, I would suggest to modify the > BiobrickFamily concept. I'll describe this in a separate mail. > > Greetings, > Raik > > Raik Gruenberg wrote: > > CC to the list for the other Pobol fans ;-) > > Note, this is still very much work in progress. > > /Raik > > > > -------- Original Message -------- > > Subject: pobol first draft > > Date: Sun, 04 May 2008 02:20:30 +0200 > > From: Raik Gruenberg <rai...@cr... > <mailto:rai...@cr...>> > > To: mgaldzic@u.washington.edu <mailto:mgaldzic@u.washington.edu> > > CC: Mackenzie Cowell <mac...@gm... > <mailto:mac...@gm...>>, Jason Morrison > > <jas...@gm... <mailto:jas...@gm...>> > > > > Hi Michael, Mac, Jason, > > > > attached is a first OWL-formatted draft of our minimal ontology: see > > partsrdf.owl. I've written this from scratch as N3 file (see > > partsrdf.n3) and > > converted it to the xml format using the RDF validator at > > http://www.rdfabout.com/demo/validator/ > > > > The owl file loads into Protege and Swoop for perhaps easier > editing. I > > would > > suggest that we continue to ship N3 versions though for better > > readability (they > > say XML was invented as a human readable format, yeah right...). > > > > What is still missing are restrictions, like that a Biobrick must > have > > exactly > > one sequence and format and so on. So far I have just fixed the > range of > > values > > of each property (owl:range) and what kind of properties a class > can have > > (owl:domain, note this goes into the property not the class > definition). > > > > More explanations: "a owl:FunctionalProperty" means this property can > > only occur > > ones per object. And "a owl:InverseFunctionalProperty" means this > > property has > > to be unique for each object. > > > > Apart from the restrictions, I didn't find out yet how to define > a datatype > > "list of Biobrick instances" for the subParts property (aka > > composition). N3 > > allows to formulate lists of values like this: > > > > BBa0102 bbf:subParts (BBa0001 BBa0002). > > > > But I didn't find any documentation how to enforce that in an > ontology. > > Hints > > anyone? > > > > Last note, I changed some property names (dnaSequence rather than > sequence, > > isCircular rather than Circular, subFamilyOf rather than parent) > to better > > conform with informal owl conventions. These are open to debate > of course. > > Ah, and I forgot the selfScar and constraint in Format... > > > > Anyway, work in progress obviously. I've some last traveling to > do starting > > tomorrow and, probably, won't touch these files for the week to come. > > Feel free > > to fill in the missing bits. Could anyone please set up a google code > > project > > for this to get version control? Or this site may be an interesting > > alternative: > > http://www.knoodl.com > > Other ideas? > > > > Greetings, > > Raik > > > > PS: also attached is an incomplete example Biobrick file using this > > ontology in > > N3 and XML format. > > > > > > Michal Galdzicki wrote: > >> i am very impressed as to how well this workshop went. and happy to > >> participate, i will make a better diagram of the pobol class > >> structures and try to explain some of the theory that goes into it. > >> mac: yell at me if i dont post anything by sunday :) > >> > >> It's a pleasure to work with you guys, > >> mike > >> > >> On Wed, Apr 30, 2008 at 9:50 PM, Mackenzie Cowell > <mac...@gm... <mailto:mac...@gm...>> > >> wrote: > >>> Hi Raik, > >>> > >>> I had a great time at the SSSB workshop. I think we accomplished a > >>> lot more > >>> than anyone expected. Michal and I have put up a little bit of info > >>> on the > >>> pobol page on OWW. We still need to fill in the rationale for > all the > >>> classes (my job) and perhaps make a better diagram of the data > model. I > >>> would also like to put together an example xml document of what a > >>> real part > >>> (F2620 or JCA's promoter family?) would look like in pobol format. > >>> > >>> Besides that, what's the next step in moving forward with > >>> implementing pobol > >>> in freebase and in brickit? > >>> > >>> I'll see you on the technical standards list! > >>> > >>> > >>> Mac > >> > > > > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Standards mailing list > > Sta...@bi... <mailto:Sta...@bi...> > > http://biobricks.org/mailman/listinfo/standards_biobricks.org > > -- > ________________________________ > > Dr. Raik Gruenberg > http://www.raiks.de/contact.html > ________________________________ > > _______________________________________________ > Standards mailing list > Sta...@bi... <mailto:Sta...@bi...> > http://biobricks.org/mailman/listinfo/standards_biobricks.org > > > > > -- > Jason Morrison > jas...@gm... <mailto:jas...@gm...> > http://jayunit.net > (585) 216-5657 -- ________________________________ Dr. Raik Gruenberg http://www.raiks.de/contact.html ________________________________ |
