[BioC] GOENTREZID2GO hard-coded in GOstats:makeGOGraph

Seth Falcon sfalcon at fhcrc.org
Tue Dec 19 20:06:26 CET 2006


Hi Paul,

Paul Shannon <pshannon at systemsbiology.org> writes:
> I use the 'makeGOGraph' function in the GOstats package 
> to obtain the induced graph for supplied gene IDs.  I typically
> follow with GOHyperG (now just a wrapper around hyperGTest) to
> calculate statistics on the nodes of the induced graph, and
> then render everything as a Cytoscape network.  

I'm glad to hear that you and others are finding the software in the
Category and GOstats packages useful.  I will (shamelessly) take this
opportunity to plug new developments as recently published here:

  http://bioinformatics.oxfordjournals.org/cgi/content/abstract/btl567v1

and encourage you to look at the updated interfaces which provide
among other things, more complete summarization of results.  The
GOHyperG function is deprecated and will, eventually, go away.  I've
recently added a lot of detail to the vignette in the devel version so
that would be a good place to look for an intro to the new functions
and classes.

> Occasionally I use the yeast genome and the YEAST 
> annotation package... which  works fine with makeGOGraph and
> GOHyperG after I made one tiny change to the former function:
> I parameterized the environment consulted for GO annotation.
>
>   makeGOGraph <- function (x, Ontology = "MF", 
>                            geneName2GO.env = GOENTREZID2GO, removeRoot = TRUE) 

[snip]

> If I am not completely missing the point here, may I ask if it might
> be worthwhile to change the signature of this method slightly, to
> something like this?
>
>  makeGOGraph <- function (x, Ontology = "MF", 
>                           geneName2GO.env = GOENTREZID2GO,
>                           removeRoot = TRUE) 

This is a clear improvement and I will patch the devel version of
GOstats so that this will be available in the upcoming release.  The
environment parameter needs to come last so as to minimize the impact
on existing code.

Thanks for a nice suggestion :-)

Best Wishes,

+ seth



More information about the Bioconductor mailing list