[BioC] DEGraph graph format?

[laurent jacob]

Hi Oliver,

2012/6/10 Oliver Ruebenacker <curoli at gmail.com>:

>  What kind of graph you are constructing? Is it a bi-partite graph
> where every physical entity is a node and every reaction is a node,
> and you connect every reaction with its reactants, catalysts and
> products?

Ultimately, the graph I construct has nodes corresponding exclusively
to genes, with only one node by gene, and edges corresponding to
expected correlations between gene expressions. For exemple if the
protein encoded by gene A activates a complex which promotes the
expression of gene B, I draw a positive edge between gene A and gene
B. But as a first step I load the graph that cytoscape builds from the
BioPAX files. An exemple of such a graph is given in the vignette
http://bioconductor.org/packages/2.8/bioc/html/NCIgraph.html.

> In BioPAX Level 2, getting that graph was quite tricky, but
> Level 3 is much easier (although catalysts ad modulators are still a
> bit awkward).

The NCI PID people sent me BioPAX Level 2 data, I don't know if Level
3 is available for all their networks.

>> Are you planning to develop a bioconductor package or an independent
>> Java parser? If you plan on using Java, you may want to look at what
>> the mskcc people did for their Cytoscape plugin, which I used for my
>> own package: http://cbio.mskcc.org/cytoscape/plugins/biopax/
>
>  I'd love to submit to Bioconductor, if that is not too difficult.

Great, good luck with the development.

Best,

Laurent

-- 
Laurent Jacob
Department of Statistics
UC Berkeley
http://cbio.ensmp.fr/~ljacob