W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > January 2010

Re: magetab2magerdf

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Wed, 06 Jan 2010 11:07:04 -0500
Message-ID: <4B44B528.4070907@openlinksw.com>
To: mdmiller <mdmiller53@comcast.net>
CC: Jim McCusker <james.mccusker@yale.edu>, Tony Burdett <tburdett@ebi.ac.uk>, w3c semweb HCLS <public-semweb-lifesci@w3.org>
mdmiller wrote:
> hi jim,
> great progress!
> and i believe you are right about how limpopo parser stores the 
> protocols, i've included tony, who's been the lead developer of the 
> parser on the email and he might be able to shed some light.
> if tony can outline a plan, i might have the time to correct this.
> cheers,
> michael
>     ----- Original Message -----
>     *From:* Jim McCusker <mailto:james.mccusker@yale.edu>
>     *To:* w3c semweb HCLS <mailto:public-semweb-lifesci@w3.org>
>     *Sent:* Tuesday, January 05, 2010 10:43 PM
>     *Subject:* Re: magetab2magerdf
>     I have checked in a first (complete) cut for magetab2magerdf 0.9a.
>     Everything should be there, both IDF and SDRF, but I haven't
>     tested against a wide range of files yet. This is alpha software.
>     *Features:
>     *
>         * Conversion of nearly all (see known issues) data from IDF
>           and SDRF MAGE-TAB files into RDF.
>         * RDF is compliant with the MGED Ontology, with minimal
>           additional classes and properties.
>         * RDF conforms to DAG structure with direct links between
>           nodes (has_derivative and has_derivation_source) as well as
>           through ProtocolApplication instances.
>         * Terms and Term Sources are automatically mapped onto the OWL
>           Classes and Individuals that are available, and will "guess"
>           at URIs for terms that do not map onto an actual OWL ontology.
>         * Validation of Terms and Term Sources through the above
>           mapping process. If an ontology is loaded and there is no
>           term available in it, the converter errors out.
>     *Known issues:*
>         * Units for characteristics and factor values are ignored. I
>           assume that if there is a unit type then these are
>           non-terms, and should be handled differently. Ideas?
>         * All protocols listed for a given SDRF node are used to
>           produce all child nodes of that node. As I mentioned before,
>           I'm not sure how I'm supposed to distinguish these in the
>           Limpopo API.
>     *Examples:*
>         * Example RDF is at
>           http://magetab2rdf.googlecode.com/svn/trunk/magetab2magerdf/examples/E-MEXP-986/E-MEXP-986.rdf
>         * Add-on MAGE-OM ontology is available at
>           http://magetab2rdf.googlecode.com/svn/trunk/ontologies/mage-om.owl
>     *Availability:* http://code.google.com/p/magetab2rdf/
>     Jim McCusker

Quick request re. the RDF documents holding the RDF triples, you need to 
add a relation that connects the RDF Documents to its contents e.g., 
using one of the Topic oriented properties from Dublic Core, FOAF, 
Bibliographic Ontology etc.

The current effect of not doing the above is demonstrated here (sadly 
this problem is way too common across the RDF data publishing realm):

1. http://bit.ly/5I62KG -- Nothing shown here since the Page Description 
seeks to describe the RDF doc, and then from the doc description expose 
RDF graph its hosts
2. http://bit.ly/7v11ym  -- Data is Explorable here because the Browers 
is processing the RDF doc contents.

Everyone: Happy New Year!



Kingsley Idehen	      Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO 
OpenLink Software     Web: http://www.openlinksw.com
Received on Wednesday, 6 January 2010 16:07:38 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:52:42 UTC