Thursday, April 28, 2011
Importing RDF input in R for analysis
August last year I asked on BioStar about how to import RDF into the R statistical package and at the time nothing seemed existing. Over the past few weeks I ported code I wrote for Bioclipse to create the rrdf package for R, which is now available from CRAN. RDF can be loaded from RDF/XML, Notation3, and Turtle files, using Jena, and read data can be queried using SPARQL. I have not yet ported the remote SPARQL functionality we used in our recent paper, but that will make a nice alternative to using ChEMBL data in Pipeline Pilot or in Bioclipse :)
Friday, April 22, 2011
CDK 1.2.x hits Debian Unstable
Thanx to the work of Onkar Shinde, the Debian package for CDK has been updated from 1.0.2 (in Squeeze) to 1.2.7 in unstable. The popularity content shows some 140 installations:
BTW, Jmol 12 is also available as Debian package now!
Thursday, April 21, 2011
ChEMBL 09 as RDF
Update 2021-02: this post is still the more read post in my blog. Welcome! Some updates:
- Ammar Ammar in our BiGCaT group has set up a new SPARQL endpoint. Please use and tweet. blog, or otherwise let others now how you use the ChEMBL RDF.
- Since this post I have blogged a lot more about ChEMBL.
I'm having a really bad month, as you can see from the number of posts. Too much to do, too little time. One of the things I have been doing in the past weeks is update the RDF for ChEMBL, now up to version 09. The SPARQL end point has not been updated yet (which is still at ChEBML 04), but you can now download the triples for self-hosting here. Like the database itself, the RDF is available under the CC-SA-BY license, requiring attribution to both the ChEMBL team as well as our efforts to create the RDF (see this README).
Willighagen, E. (2011, April 21). ChEMBL 09 as RDF. Chem-bla-ics. https://doi.org/10.59350/9w84r-evn93
Thursday, April 14, 2011
How to create a proportional Venn diagram?
Above is a typical Venn diagram. But I like mine so that the area of the circles reflect the number of object in that circle (or logarithm thereof), and the areas of the overlaps to be proportional to the number of objects in that overlap. Is there an 'arp' for that? (I know that "there is an app for this" it trademarketed... 'arp', instead, is short for R package).